Full text of "Reports of the Committee on Electrical Standards appointed by the British Association for the Advancement of Science"

See other formats

Google

This is a digital copy of a book that was preserved for generations on Hbrary shelves before it was carefully scanned by Google as part of a project

to make the world's books discoverable online.

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject

to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books

are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover.

Marks, notations and other maiginalia present in the original volume will appear in this file - a reminder of this book's long journey from the

publisher to a library and finally to you.

Usage guidelines

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we liave taken steps to
prevent abuse by commercial parties, including placing technical restrictions on automated querying.
We also ask that you:

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for
personal, non-commercial purposes.

+ Refrain fivm automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the
use of public domain materials for these purposes and may be able to help.

+ Maintain attributionTht GoogXt "watermark" you see on each file is essential for informing people about this project and helping them find
additional materials through Google Book Search. Please do not remove it.

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner
anywhere in the world. Copyright infringement liabili^ can be quite severe.

About Google Book Search

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web

at|http : //books . google . com/|

physic

REPORTS

OF THE

ELECTRICAL STANDARDS COMMITTEE

OF THE

BRITISH ASSOCIATION

CAMBRIDGE UNIVERSITY PRESS

l^OlUlon: FETTER LANE, EC.

C. F. CLAY, Manager

\tj\

fftiinbursi): xoo, PRINCES STREET
lontion: WILLIAM WESLEY & SON. a8, ESSEX STREET, STRAND

»«tlin: A. ASHER AND CO.

l.rip>{fi: F. A. BROCKHAUS

^<to %otk: G. P. PUTNAM'S SONS

Vornbag antl ealnttta: MACMILLAN AND CO., Ltd.

A// rights reserved

REPORTS

OF THE COMMITTEE

ELECTRICAL STANDARDS

APPOINTED BY

THE BRITISH ASSOCIATION FOR THE

ADVANCEMENT OF SCIENCE ' ■'■'■''■

;

j( -u^U.. '' '" -'

Reprinted by Permission of the Council

A RECORD OF THE HISTORY OF "ABSOLUTE UNITS"
AND OF LORD KELVIN'S WORK IN CONNEXION

WITH THESE

Cambridge :

at the University Press

•• ' • I

PRINTED BY JOHN CLAY, M.A. |

AT THE UNIVERSITY PRESS

PREFACE

riiHE early Reports of the Committee on Electrical Standards
-^ were for many years the highest authority on their subject.
From 1880 to 1884 they were never long out of my hands. As
they are now rather difficult of access, it has been decided to
republish them not only on account of their historical interest,
but as providing an excellent statement of principles by such
masters as Kelvin, Maxwell, and Jenkin. This has been made
possible by a generous donation of £100 from Mr R. K. Gray
and by a similar grant from the Oeneral Committee of the
Association.

Mr F. E. Smith has acted as Editor. His intimate acquaint-
ance with recent developments of the subject, in which he has
taken a leading part, eminently qualify him for the task.

RAYLEIGH.

December 1912.

CONTENTS

of
meeting pagbs

1862 FifiST Report — Cambridge,

Unit and standard of resistance— magnitude of unit —
relation to other units — unit should be definite and re-
producible— Weber's and Thomson's system — Siemens'
mercury unit — absolute determinations of resistance by
Weber—notification of Committee's appointment . . 1 — 16

Appendix A. Variation of resistance of alloys due to
change of temperature. By Dr Matthiessen . . . 16 — 21

Appendix B. Electrical permanency of metals and
alloys. By Dr Matthiessen 21—23

Appendix C. Reproducibility of Electrical Standards
by chemical means. By Prof. Williamson and Dr
Matthiessen 24—37

Appendices 2), E^ F, Letters from Prof. Eirchoff, Dr
-Siemens, and Dr Esselbach, on the general introduction
of one unit of electrical resistance 37^46

Appendix O, Circular addressed to foreign men of
science, stating reasons for and against various units and
standards of resistance 46 — 60

Appendix ff. Description of the apparatus arranged
by Fleeming Jenkin for copying standards of resistance . 50 — 57

1863 Second Report— Newcastle-on-Ttne.

Practical adoption of "absolute" system — meaning
of "absolute" — electrostatic and electromagnetic units
— definitions of magnetic field and unit magnetic pole —
intensity and direction of field — uniform ma^etic field
— ^illustrations of measurements of resistance m absolute
measure — Prof. Thomson's method as adopted at King's
College — difficulties experienced in the experiments —
relation of the B. A. unit to that derived by Weber and
Siemens — constancy of resistance alloys — reproduction
of a resistance standard by means of mercury —
Thomson's electrometer— objects to be pursued by the
Committee 58—78

Appendix A- Electrical permanency of metals and
alloys. By Dr Matthiessen. — Experiments on wires of
8il?er, oopjper, gold, platinum, gold-silver, and Qerman-
ailver-— diffiarences between hard drawn and annealed
wires — change due to annealing 78 — 86

CONTENTS

Tear of
meeting

1863

PAGBB

Appendix B, On the elementary relations between
electncal measurements. By Prof. J. Clerk Maxwell and
Mr Fleeming Jenkin.

1. Objects of treatise.— 2. Derivation of units from
fundamental standards. — 3. Standard mechanical \inits. —
4. Dimensions of derived units. — 6. Magnets and mag-
netic poles. — 6. Magnetic field. — 7. Magnetic moment. —
8. Intensity of magnetization. — 9. Coefficient of mag-
netic induction. — 10. Magnetic potentials and equi-
potential surfaces. — 11. Lines of magnetic force. — 12.
Relation between lines of force and equipotential sur-
faces.— 13. Meaning of expressions such as conductor. —
14. Electric quantity. — 15. Current. — 16. Electromotive
force. — 17. Resistance. — 18. Measurement of electric
currents by their action on a magnetic needle. — 19.
Measurement of electric currents by their mutual action
on one another. — 20. Weber's electrodynamometer. — 21.
Comparison of the electromagnetic and electrochemical
action of currents. — 22. Magnetic field near a current. —
23. Mechanical action of a magnetic field on a closed
conductor conveying a current. — 24. Law of the mechanical
action between electric currents and other electric
currents or magnets. — 25. Electromagnetic measurement
of electric quantity. — 26. Electric capacity of a con-
ductor.— 27. Direct measurement of electromotive force. —
28. Indirect measurements of electromotive force. — 29.
Measurement of resistance. — 30. Resistance in electro-
magnetic units is measured by an absolute velocity. —
31. Magneto-electric induction.->32. Material standards
for the measurement of electric magnitudes. — 33.
Electrostatic measure of electric quantity. — 34. Elec-
trostatic system of units. — 35. Ratio between electro-
static and electromagnetic measures of quantity.— 36.
Electrostatic measure of currents. — 37. Electrostatic
measure of electromotive force. — 38. Electrostatic measure
of resistance. — 39. Electric resistance in electrostatic
units is measured by the reciprocal of an absolute
velocity. — 40. Electrostatic measure of the capacity of
a conductor.— 41. Absolute condenser. Practical measure-
ment of quantity. — 42. Practical measurement of currenta
— 43. Practical measurement of electromotive force. — 44.
Comparison of electromotive forces by their statical effects.
— 45. Practical measurement of electric resistance. — 46.
Experimental determination of the ratio "v'' between
electromagnetic and electrostatic measures of quantity. —
47. Electnc potential. — 48. Density, resultant electric
force, electric pressure. — 49. Tension. — 60. Conducting
power, specific resistance, and specific conducting power.
— 51. Specific inductive capacity. — 52. Heat produced
in a conductor by a current. — 53. Electrochemical equi-
valents.—54 Electromotive force of chemical affinity. —
55. Tables of dimensions and other constants. — 56.
Magnitude of units and nomenclature ....

Ajppendix D, Description of an experimental measure-
ment of electrical resistance at King's College. By Prof.
J. Clerk Maxwell and Messrs Balfour Stewart and

86—140

CONTENTS IX

Year of

meeting faqeb

FleemiDg Jenkin. Part I — Qeneral description of the

method employed. Part II — Description of tne apparatus.

Part III — Mathematical theory of the experiment.

Part IV — Details of the experiments — results . . 140 — 158

1864 Third Report — Bath.

Issuing of coils for public use — second determination of
a resistance in absolute measure — probable error— repro-
duction of standard coils — table giving the resistance of
mercury unit in terms of B.A. unit — adoption of B.A.
system of units in colonies and foreign countries . 169 — 166

Appendix A, Description of a further experimental
measurement of electrical resistance made at King's
College. By Prof. J. C. Maxwell, Mr Fleeming Jenkm,
and Mr Charles Hockin 166—167

^ Appendix B, On the electrical permanency of metals

and alloys. By A. Matthiessen 167—169

Appendix C, On the reproduction of electrical
standards by chemical means. By Dr Matthiessen
and C. Hockin

Results on wires of silver, copper, gold, lead, gold-
silver alloy, and columns of mercury .... 170 — 18^

1865 Fourth Report — Birmingham.

Construction of wire standards of resistance — tem-
perature coefficient of platinum-silver — distribution of
coils — unit used in tests of the Atlantic cable . 190 — 195

Appendix A. On the construction of the copies of
the B. A. unit. By Dr Matthiessen and Mr C. Hockin . 196—197

1867 Fifth Report — Dundee.

Experiments by Dr Joule — Siemens' resistance
measurer — comparison of the resistance units — diffi-
culties encountered in determination of imit of capacity
— B.H.F. of a Daniell^s cell — absolute electrometer —
determination of "i;'* 198 — 206

Appendix I. On a ** Resistance-Measurer." By C. W.
Siemens 206—208

Appendix IL On a modification of Siemens' re-
sistance-measurer. By Fleeming Jenkin 209 — 210

Appendix III, Comparison of B. A. units to be
deposited at Kew Observatory. By C. Hockin . 211

Appendix IV. Experiments on capacity. By Fleeming
Jenkm 212 — 219

Appendix V. Report on electrometers and electro-
static measurements. By Sir Wm Thomson.

Electrometer and electroscope — repulsion electro-
meters — symmetrical electrometers — attracted-disk
electrometm — construction of quadrant electrometer —
constancy of charge — replenisher— gauge— estimation of
position of index — sensitiveness and constancy — absolute
electrometer — equation for potential difference— port-
able electrometer — preparation and use of instrument —
Bouroes of error — stanaard and long range electrometers . 219 — 266

& A. h

X CONTENTS

Tear of

meeting pages

Appendix VL DjDamical equivalent of heat from the
thermal effects of electric curreuts. By Dr Joule.

Current balance — calorimeter and thermometers used
— first, second, and third series of thermal and radiation
experiments — sources of error — results .... 256 — 270

1869 Sixth Report — Exeter.

Description of Sir Wm Thomson's experiments for the
determination of "t;." By W. F. King .... 271—273

Experiments on the value of " v" By J. C. Maxwell . 274 — 276

Report on the new unit of electrical resistance. By
Fleeming Jenkin.

Historical introduction — the experiments of the B. A.
Committee— reproduction of standards .... 277 — 290

1870 Seventh Report — Liverpool.

Suggestions for determination of units of capacity,
E.H.F., and current 291 — 292

1881 Eighth Report— York.

Testing of coils — determination of absolute capacity
by Dr Muirhead and Mr Hockin — Latimer Clark cell-
fundamental standards 293—296

Appendix L Preliminary experiments on determina-
tion or electrical resistance in absolute measure. By Prof.
Carey Foster 296—304

Appendix II. Causes of the variation in the tempera-
ture coefficient of the alloys of platinum and silver. By
H. Taylor 306—316

1882 Ninth Report— Southampton.

Results of experiments made on the temperature-
coefficient of resistance of metals and alloys . . 317 — 316

1883 Tenth Report — Southport.

Testing of standard coils at Cavendish Laboratory —
comparison of ten units with single unit .... 319 — 324

1884 Eleventh Report — Montreal.

Paris Congress — adoption of " legal ohm '' as standard
— ^relation between ** legal ohm " and B. A. unit . . 325 — 326

Appendix. On the values of the B.A. standards of
resistance greater than one B.A. unit. By R. T. Glaze-
brook and H. M. Elder 327—328

1885 Twelfth Report — Aberdeen.

Legal ohm standards — comparison of coils with French
mercury standards of resistance 329 — 331

1886 Thirteenth Report— Birmingham.

Insulation resistance of coils — ^faulty paraffin wax 332 — 333

Appendix. On the values of some standard resistance
coils. By R. T. Glazebrook and T. C. Fitzpatrick . . 333—338

CONTENTS

of
meeting

1887

1888

1889

1890

1891

1892

PAOEB

Fourteenth Report — Manchester.

Consideration of " legal '» ohm, ampere, volt, coulomb,
and farad— recommendation of the "Watt" as the unit
^^?oweT 339-340

Fifteenth Report— Bath.

standard air condensers— experiments on specific re-
sistance of copper— adoption of name "Therm" for the
unit of heat, and "Joule " for the unit of work— specific
resistance of mercury in B.A. units .... 341—342

Appendix. On the permanence of the original B. A.
standards of resistance and of other standard coils. By
R. T. Glazebrook and T. C. Fitzpatrick .... 343—369

Sixteenth Report — Newcastle-upon-Tyne.

Standard air condensers — specific resistance of copper
— resolutions of Electrical Congress in Paris . . . 360—363

Seventeenth Report — Leeds.

Resolutions relating to mercury imit and absolute
ohm, and B.A. imit and ohm 3^4 3^7

Appendix L On the values of certain standard re-
sistance coila By R. T. Glazebrook .... 367—373

Appendix IL On the air condensers of the British Asso-
ciation. By R. T. Glazebrook with note by Dr Muirhead.

Construction of condensers — tests on leakage vi-
brating and rotating commutator — variation of capacity
with finequency — absorption and instantaneous capacity
— ^results with mica ana with air condensers . . 373 397

Appendix HI, On the specific resistance of copper.
By T. C. Fitzpatrick.

iZtffttm^of Matthiessen's results — resistance of various
specimens of wire — difference between hard drawn and
annealed wires 397 410

Appendix IV. A comparison of a platinum thermo-
meter with some mercury thermometers at low tempera-
tures. By E. H. Griffiths 411 419

Appendix V. On the absolute resistance of mercury.
By R T. Glazebrook.

Table giving value of ohm expressed as the resistance
of a column of mercury 41&— 421

Eighteenth Report — Cardiff.

Comparison of RA. resistance coils with M. Benoit's
mercury tubes 42SJ— 424

Appendix I. Report of the Electrical Standards Com-
mittee appointed by the Board of Trade. Specifications
for silver voltamet^ and for the Clark cell . . . 424 432

Nineteenth Report — Edinburgh.

Board of Trade Standards — electromotive force of
Clark cell— resolutions relating to mercury tmit of re-
sistance 433—436

CONTENTS

Year of
meeting

1892

1893

1894

PAGES

Appendix L Information circulated by Secretary for
the August Meeting of the Committee — values for
electrochemical equivalent of silver — values found for
E.M.F. of Clark cell. 435 — 438

Appendix IL On the temperature coefficient of re-
sistance of mercury. By M. G. Quillaume . . . 438

Appendix III, On a special form of Clark cell. By
H. J. Carhart 439

Appendix IV, On wire standards of electrical re-
sistance. By Dr St Lindeck.

Experiments on wires of german-silver and of
manganin. Mercury standards of resistance . 440 — 450

Appendix V, On the Clark cell. By Dr Kahle . 460—456

Appendix VI. On the values of certain standard
resistance coils. By R. T. Glazebrook .... 465 — 467

Appendix VII, On the standard condensers of the
Association and on certain resistance coils. By R. T.
Glazebrook 468 — 461

Appendix VIIL On the values of certain standards
of resistance and electromotive force sent from Berlin.
By R. T. Glazebrook 461—464

Twentieth Report — Nottingham.

Relation between B.A. unit and the ohm — Chicago
Congress — name "Henry" for unit of self-induction —
"International" Ohm 465—466

Appendix I. Supplementary Report of the Electrical
Standards Committee of the Board of Trade — specification
for silver voltameter — specification for Clark cell — reso-
lutions of B. A. Electricial Standards Committee relating
to the mercury unit of resistance 467 — 476

Appendix II, Heating effect produced in coils by the
currents used in testing. By R. T. Glazebrook . . 476 — 477

Appefidix III, On standards of low electrical re-
sistance. By J. Viriamu Jones.

Description of a Lorenz apparatus — results 478 — 482

Twenty-first Report — Oxford.

Traces of acid found in paraffin- wax .... 483 — 484

Appendix I. Report of International Electrical
Congress in Chicago on electrical units — List of Delegates
— resolutions relating to international ohm, ampere,
volt, coulomb, farad, joule, watt, and henry . 486—489

Appendix II, Determination of the International
Ohm m absolute measure. By J, Viriamu Jones . 489 — 496

Appendix III. Comparison of the standard coils used
by rrof. Jones with the standards of the Association.
By R. T. Glazebrook 497—499

Appendix IV, Comparisons of certain Ohm-standards
of the Board of Trade. By J. Rennie .... 499—500

CONTENTS

XUI

of
maeting

1895

1896

1897

1898

PAGES

Appendix V, Values of five standard coils B. A. units
belonging to the Indian Qovemment as compared with
Dr Muirhead's standard 501

Appendix VI. On the specific resistance of copper
and of silver. By T. C. Fitzpatrick .... 502— 5Q8

Appendix VII. Final Report of the Electrical
Standards Committee of the Board of Trade — Order in
Council r^;arding standard measurements — standards
of resistance, current, and electromotive force— speci-
fication of silver voltameter— preparation of the Qark
cell 509—619

TWENTY-SECX)ND REPORT — IPSWICH.

Appendix on Magnetic Units by Dr 0. Lodge, with
remarks by F. G. Baily, Profs. Everett and G. C. Foster,
and Dr G. J. Stoney 520 — 538

Twenty-third Report — Liverpool.

Mr E. H. Griffiths' letter to a large number of physicists
on a standard thermal unit 539 — 543

Appendix L Extracts from letters received, dealing
with tne unit of heat 544 — 554

Appendix 11. Capacity for heat of water from 10** to
ard 554

Appendix III. Recalculation of total heat of water
from Regnault's and Rowland*s experiments. By W. N.
Shaw 555—559

Twenty-fourth Report — Toronto.

The calorie — comparison of platinum thermometers
with hydrogen thermometer — the mechanical equivalent
of heat — Schuster and Gannon's, Griffiths', and Row-
land's experiments — values of the specific heat of water
at 15" C. — variation of the specific heat of water . . ^60 — 564

Appendix I. Note on the constant- volume gas ther-
mometer. By G. Carey Foster 564 — 567

Appendix II. Determination of the Ohm made in
testing the MKj^ill University Lorenz apparatus. By
W. K Ayrton and J. Viriamu Jones .... 567 — 675

Twenty-fifth Report — Bristol.

Appendix I. Comparison of standard coils used by
Prof. W. E. Ayrton and J. V. Jones in their measure-
ments of the specific resistance of mercury. By R. T.
Glaaebrook 577—581

Appendix 11. Determination of temperature co-
efficients of two 10 ohm coils used in the 1897 deter-
mination of the ohm. By M. Solomon . . 581 — 589

Appendix III. An ampere balance. By W. £. Ayrton
and J. Viriamu Jones 589 — 591

XIV CONTENTS

Year of

meeting paoes

1899 Twenty-sixth Report— Dover.

Qrant for construction of current balance — standard
scale of temperature and platinum resistance thermo-
meter 592—693

Appendix L The mutual induction of coaxial helices.
By Lord Rayleigh 693—696

Appendix II. Proposals for a standard scale of tem-

Esrature based on the platinum resistance thermometer,
y H. L. Callendar 695—697

Appendix III A comparison of platinum and gas
thermometers made at the International Bureau of
Weights and Measures at Sevres. By P. Chappuis and
J. A. Harker 597—600

Appendix IV, On the expansion of porcelain with
rise of temperatura By T. G. Bedford .... 600—601

1900 Twenty-seventh Report — Bradford.

Report of Sub-committee on platinum themometers
— mercury resistances to be set up — progress with
ampere balance — resolutions of Elecmcal Congress at
Paris relating to unit of magnetic field and unit of
magnetic flux 602—604

Appendix. Note on an improved standard resistance
coiL By R S. Whipple 604—606

1901 Twenty-eighth Report— Glasgow.

Appendix. Note on a comparison of the silver de-
posited in voltameters containing different solvents. By
S. Skinner 607—611

1902 Twenty-ninth Report— Belfast.

Progress with mercury units of resistance and Clark
and Weston cells — the B. A. air condensers — con-
struction of platinum thermometers — a new Lorenz
apparatus 612 — 615

Appendix. On the definition of the unit of heat 615 — 619

1903 Thirtieth Report — Southport.

Appendix I. On the values of the resistance of
certain standard coils of the British Association. By
F. E. Smith 627—636

Appendix II. The relation between the International
Ohm and the unit of resistance employed at the National
Physical Laboratory. By F. E. Smith .... 636—637

Appendix III. On the platinum thermometers of the
British- Association. By J. A. Harker .... 638 — 646

A ppendix I V. Resistance of metre-gramme of annea led
copper 646

CONTENTS XV

meeliiig pages

1904 Thirty-first Report — Cambridge.

Progress with current balance —proposal to substitute
saturated Weeton cell for the Clark cell — nomenclature
for magnetic units 647 — 650

Appendix L On anomalies of standard cells. By
F. if. Smith 661—661

Appendix 11, On the electromotive force of Clark's
cell. By A. P. Trotter 661

1905 Thirty-second Report — South Africa.

Progress with current balance — St Louis Electrical
Congress — International Standardisation — consideration
of International Congress — Conference of Representatives
at the Reichsanstalt 662—666

Appendix. On the preparation of a cadmium cell.
By F. E. Smith 666—673

1906 Thirty-third Report — York.

Ohm and Ampere to be defined independently —
opinions of Reichsanstalt Conference .... 674 — 676

Appendix, On methods of high precision for the
comparison of resistances. By F. £. Smith . . 676 — 696

1907 Thirty-fourth Report— Leicester.

Results obtained with Ayrton-Joues current balance
— electromotive force of Weston and Clark cells . 697 — 700

Appendix L On the present condition of the work
on electric units at the National Physical Laboratory.
By F. E. Smith 700—702

Appendix II. Specification for the practical applica-
tion of the definition of the International Ampere . 703 — 707

Appendix III, Preparation of the Weston (cadmium)
standaidoell 707—711

1908 Thirty-fifth Report — Dublin.

Comparison of Board of Trade ampere standard with
the Ayrton-Jones current weigher — progress with new
mercury standards 712 — 716

Appendix I, On the secular changes of the standards
of resistance at the National Physical Laboratory. By
F. E. Smith 716—738

Appendix II. Specifications for the practical realisa-
tion of the definitions of the International Ohm and
International Ampere, and instructions for the prepara-
tion of the Weston cadmium cell 738—743

1909 Thirty-sixth Report — Winnipeg.

Comparison between standards of resistance of the
National Physical Laboratory, the Bureau of Standards,
and the Reichsanstalt .... 744 — 747

XVI

CONTENTS

Tear of
meeting

Appendix'. Report of International Conference on
Electrical Units and Standards, 1908— List of Countries
and Delegates — resolutions relating to the fundamental
units — specifications relating to mercury standards of
resistance, the silver voltameter, and the Weston normal
cell — recommendation for establishment of a permanent
International Commission for Electrical Standai*ds

1910 Thirty-seventh Report—Sheffield.

International co-operative work at Washington —
anomalies of cadmium amalgams — Order in Council
relating to Electrical Standees, dated January 10,

piass

748—758

1910

759—764

1911 Thirty-eighth Report — Portsmouth.

Lorenz apparatus— effect of changing humidity on
standard coils of manganin — progress with research on
silver voltameter and standanl cells ....

1912 Thirty-ninth Report — Dundee.

765—767

Absolute measurements of current and resistances-
current balances of National Physical Ijaboratory and
Bureau of Standards — International experiments with
silver voltameter — coniparison of resistance standards
and Weston cells in England, America, France, and
Gtermany 768 — 772

Approximate Relative Values of various Units of
Electrical Resistance

Appendix A. — Relative Values of various Units of
Electrical Resistance

Plate 1

)»

7o face p. 165

288

56
142
172
196
224
240
272
334
336
346

INTKODUCTION

The original British Association Committee on Standards of
Electrical Resistance was appointed at the suggestion of Professor
William Thomson (later Lord Kelvin) in 1861 and consisted of
Professor A. Williamson, F.R.S., Professor C. Wheatstone, F.R.S.,
Professor W. Thomson, F.R.S., Professor W. H. Miller, F.RS.,
Dr A. Matthiessen, F.R.S., and Mr Fleeming Jenkin. The principal
object of the Committee was, first, to determine what would be
the most convenient unit of resistance, and second, what would
be the best form and material for the standard representing that
unit.

When the Committee was first appointed no coherent system
of units for the measurement of electric resistance, current,
electromotive force, quantity, or capacity, had met with general
approval. It was true that Professor W. Weber's absolute system
existed on paper, but it was not understood or used by practical
men.

From 1862 to 1870 much valuable work was done by the
Committee which in the interim had been strengthened by
the addition of Sir Charles Bright, Professor J. Clerk Maxwell,
Mr C. W. Siemens, Mr Balfour Stewart, Mr C. F. Varley, Professor
G. Carey Foster, Mr Latimer Clark, Mr D. Forbes, Mr Charles
Hockin, Dr Joule, and Dr Esselbach. During this period Professor
Thomson was a particularly active member. He not only devised
the well-known revolving coil method for the absolute measure-
ment of a resistance (independently we believe of the prior
suggestion of the method by W. Weber), but he also designed
apparatus for the absolute measurement of electric current, and
electrometers for the measurement of electromotive force. He
made a long report on Electrometers and Electrostatic Measure-
ments in 1867, and in 1869 he made a determination of the
number of electrostatic units in the electromagnetic unit. He

XVlll INTRODUCTION

remained a member of the Standards' Committee until its dis-
solution in 1870 ; in 1881 he was appointed a member of the new
Committee, and continued to take an active interest in the work
until his death in 1908.

During the period 1862 — 1870 the Committee reported
(1) The measurement of a resistance in terms of the centimetre
and second, by Professor J. Clerk Maxwell and Messrs Balfour
Stewart and Fleeming Jenkin. (2) A determination of the
dynamical equivalent of heat from the thermal effects of electric
currents, by Dr Joule. (3) An investigation of resistance alloys,
by Dr Matthiessen. (4) The determination of a unit of capacity,
by Dr Matthiessen, Mr Hockin, Professor Carey Foster, and
Mr F. Jenkin. (5) The determination of " v " the ratio between
the electrostatic and electromagnetic units, by Sir William
Thomson, and also by Professor J. Clerk Maxwell. In addition
the Committee caused to be printed (6) A Report of Electro-
meters and Electrostatic Measurements, by Sir William Thomson,
and (7) A Treatise on The Elementary Relations between Electrical
Measurements, by Professor J. Clerk Maxwell and Mr Fleeming
Jenkin.

The choice of a unit of resistance was, at the time of the
appointment of the Committee, a matter of considerable import-
ance. Until about 1850 all units of resistance were based on the
more or less arbitrary size and weight of some conductor in the
form of a wire. In England, one such unit (proposed by Professor
Wheatstone in 1843) was that of a foot of copper wire weighing
100 grains, and a second unit was equal to 1 mile of copper wire
of ^ inch in diameter. In 1851 W. Weber proposed a system
of electrical and magnetic measurement in which an electrical
resistance would be expressed as a velocity. Subsequently,
Professor W. Thomson defined a unit of work in Weber's system
and thus allowed of all physical measurements being connected
together.

The immense value of such a coherent system as outlined by
W. Weber and Professor Thomson was fully appreciated by the
Committee on Electrical Standards; there was, however, one
difficulty. If the unit of resistance were defined in terms of
Length and Time without qualification the material standard
practically representing it would require continual correction
from time to time as successive determinations were made with

INTRODUCTION XIX

increasing accuracy. It became then a matter for consideration
whether the advantages of the arbitrary material standard and
those of the absolute system could not be combined, and ulti-
mately the Committee decided that a material standard should be
prepared in such a form as should ensure practical permanency.
Farther that this should be equal to ten millions metres/second
and be known as the unit of 1862. The magnitude of one
metre/second was &r too small to be convenient in practice, and
the decimal multiple chosen was thought to be a most convenient
one. For this Latimer Clark suggested the name ''Ohmad/'
which in the abbreviated fonn " Ohm " was finally adopted. It
is of more than passing interest to note that these proposals of the
1862 Committee are practically identical with the resolutions of
the last Conference on Electrical Units and Standards which met
in London in 1908. In the interim of 46 years the form and
nature of the material standard has been much discussed and
altered, but the unit has remained fixed and is now international.

For the purposes of the construction of a material standard
Dr Matthiessen undertook a special investigation on the electrical
properties of alloys and pure metals in the solid and liquid
states, and much of oar knowledge on the change of resistance
with the physical changes produced by annealing, hardening,
bending, etc. was then obtained. A number of resistance coils
of special form, known as the B.A. type, were constructed by
Dr Matthiessen and Dr Muirhead, and the majority of these are
at present lodged at the National Physical Laboratory. It is
practically certain that no other resistance coils in existence are
of such great historical interest. They have been compared
together very many times during the past 50 years, and the
secular changes of resistance have been traced with some success.
We now know that the various resistance alloys experimented
with have not kept constant in resistance, but there is strong
evidence that the platinum resistance coils have kept constant,
and if so, these serve to connect the early work of the Committee
with some of the researches of Lord Rayleigh, of Dr Glaze-
brook, of Professors Rowland and Mascart, and with modem
investigations.

For the purposes of the 1862 Committee Professor W. Thomson
designed the revolving coil apparatus by which the resistance of
a coil could be determined in electromagnetic measure. Such

XX INTRODUCTION

a method had been previously proposed by W. Weber. The unit
derived from the experiments made in 1863 and 1864 was about
8^ per cent, larger than the unit derived from a German-silver
coil previously measured by Professor Weber in terms of the
metre and second. It was 6^ per cent, larger than the unit
derived from a value published by Professor Weber of Dr Siemens'
mercury unit. On the other hand it was about 5 per cent, smaller
than the unit derived from other coils also based on a determina-
tion by Professor Weber. These discrepancies did not, however,
cause considerable surprise.

In 1870 the Committee was dissolved, but on the suggestion
of Professor Ayrton it was reappointed in 1881. The reappoint-
ment was largely due to the discrepant results which had been
obtained by experimenters who had reexamined the absolute
resistance of the British Association unit. The Committee were
of opinion that further experiments should be made, and they
also thought that it would be well to reconsider the question
whether the ohm should be defined by reference to a concrete
standard, or whether the term "ohm" should be understood to
mean a resistance of 10* cm./sec units. At that time, according
to Professor Kohlrausch, the British Association unit was nearly
2 per cent, too great, and according to Professor Rowland it was
nearly 1 per cent, too small. On the other hand, H. Weber had
obtained by more than one method results very nearly in harmony
with those of the Committee. In 1881-3, Lord Rayleigh, Professor
Schuster, and Dr Glazebrook, all of whom were members of the
Standards Committee, made absolute measurements of the British
Association unit, and the results were most satisfisM^tory. In 1881,
Lord Rayleigh and Professor Schuster, using the revolving coil
method, found that one British Association unit was equal to
09893 10' C.G.S. units, and with a new apparatus constructed in
1882 it was found that one B.A. unit was equal to 0*9866 10* C.G.S.
units. In 1883, using the method of Lorenz, Lord Rayleigh and
Mrs Sidgwick found the ratio to be 098677, and in 1882 Dr Glaze-
brook found the ratio to be 0-9867. There was, therefore, little
doubt but that the first determination of the Committee was in
error by at least 1 per cent., and the question naturally arose as
to the reason of this. Lord Rayleigh probed the matter very
carefully, and pointed out the possibility of a considerable error
due to an under-estimation of the self-induction of the coil.

INTRODUCTION XXI

The value of the B.A. unit in absolute measure was involved
in the experiments by Dr Joule on the Dynamical Equivalent
of Heat. These are described in the Report for 1867. The result
fiom the agitation of water is 24868 (= 41*586 x !()• c.g.s.), while
that derived from the passage of a known absolute current
through a resistance compared with the B.A. unit was 25187
(= 42"119 X 10" C.G.S.). The latter result is on the supposition
that the B.A. unit is really 10* C.G.S. units. If the unit of work
had been used as a means of deriving the unit of resistance, it
follows that the B.A. unit was 24868/25187 «= 0-9873 c.g.s. units,
which is in very close agreement with the values obtained by
Lord Rayleigh, Professor Schuster, and by Dr Glazebrook.

It will be seen that the work of the members of the 1881
Committee practically established the ratio of the B.A. unit to
the ohm within a few parts in ten thousand, and some years
elapsed before further measurements were made of a resistance
in terms of the centimetre and second. In 1888 Dr Glazebrook
made another determination, and in 1891 Professor Viriamu Jones
measured a resistance by the method of Lorenz. A new Lorenz
apparatus, ordered by Professor Callendar for the McGill University,
was used by Professors Ayrton and Viriamu Jones in 1897. In
more recent years National Standardising Laboratories have arisen
in Germany, Great Britain and America, and at all of these institu-
tions apparatus has been constructed for the determination of a
resistance in absolute measure. The apparatus at the National
Physical Laboratory is a modified form of Lorenz apparatus, and
was presented to the Laboratory as a memorial of Professor
Viriamu Jones, a former member of the Committee. The con-
struction of the appanitus was rendered possible by a generous
gift of £700 from the Drapers' Company of London, and by the
kindness of Sir Andrew Noble, Bart., K.C.B., F.R.S.

The first member of the Committee to measure a current in
absolute measure was Dr Joule, who in 1864 employed a current
weigher in his determination of the d}mamical equivalent of heat.
The current balance used by Joule had three circular flat coils
wound with copper strip, one being suspended from the beam
of a balance so that its mean plane, which was horizontal, was
midway between those of the other two coils which were fixed.
In 1882 Lord Rayleigh showed that by suitable design the
constant of such a current balance could be reduced to a numeric

XXll INTRODUCTION

depending on the mean radii of the coils as a ratio, which could
be determined electrically with high precision without any linear
measurements whatever having to be made. In 1883 Lord Bay-
leigh published the results he had obtained with such a current
weigher, and in recent years balances on the same principle have
been constructed at the Laboratoire Central d'Electricit^, Paris,
and at the Bureau of Standards, Washington. In 1898 the
Standards Committee appointed Professors Ayrton and Viriamu
Jones to construct a new current weigher, and this, now known as
the British Association Ayrton-Jones Current Weigher, is at the
National Physical Laboratory and can be used at any time. It is
of interest to record that measurements of current in CG.s. units,
made at the National Physical Laboratory and the American Bureau
of Standards agree within 4 parts in 100,000.

As early as 1881 the Committee arranged for the systematic
testing of resistance coils, and about the same time Dr Muir-
head undertook to make and issue standard condensers. Since
1883 the resistance standards of the Association have been in
charge, first of Lord Rayleigh and afterwards of Dr Glazebrook,
both of whom have also investigated standards of electromotive
force and mercury standards of resistance.

In addition to dealing with the primary electrical standards
the Committee have also considered the subjects of platinum-
thermometry, thermal and magnetic units, and physical constants
in general. During the latter years of the Committee's existence
it was active in its efforts to promote international uniformity in
standards, and for this purpose many experiments were under-
taken at the National Physical Laboratory, on behalf of the
Committee.

The appointment by the London Conference of 1908 of an
International Scientific Committee of fifteen to direct work in
connection with the maintenance of electrical standards relieved
the Committee of much of its responsibility. The main objects
for which it had been appointed had been achieved; in all the
principal countries of the world the same units of resistance, of
current, and of electromotive force had been adopted and the
standards in use were practically identical.

Very few notes have been added to the Reports and but little
matter has been omitted. In general where notes have been
introduced references to other portions of the Collected Reports

INTRODUCrriON

XXIU

are made. The parts omitted referred to tests on a number of
resistance coils intended for commercial use, and were not of
general interest. In reprinting the papers any errors of inadvert-
ence which were discovered are of course corrected. Corrections
of this kind are not indicated.

R. T. G.
F. E. S.

Past and Present Members of the Committee of the British
Association for Improving the Construction of Practical Standards
for Electrical Measurements.

[* Members of the Committee in 1912.]

1862-70. 1881-1907.

Lord Kelvin.

1862-70.

Professor A. Williamson.

1862-70.

Sir Charles Wheatstonb.

1862-70.

Profeflsor W. H. Miller.

1862-70.

Dr A. Matthiessen.

1862-70. 1881-1884.

Professor Flkeming Jenkin.

1863-70.

Mr C. F. Varley.

1863-70.

Professor Balfour Stewart.

1863-70.

Mr C. W. Siemens.

1863-70.

Professor J. Clerk Maxwell.

1863-70.

Dr Joule.

1863-70.

Sir Charles Bright.

1863.

Dr Esselbagh.

1867-70. 1881-*.

Professor Q. C. Foster.

1867-70.

Mr Latimer Clark.

1867-70.

Mr D. Forbes.

1867-70. 1881.

Mr Charles Hogkin.

1881-1908.

Profwwor W. E. Ayrton.

1881-*.

Professor J. Perry.

1881-*.

Professor W. Q. Adams.

1881-*

Lord Rayleioh.

1881-*.

Sir Oliver J. Lodge.

1881-97.

Dr John Hopkinson.

1881-*.

Dr A. Muirhbad.

1881-*.

Sir W. H. Prbbcb.

1881-1897.

Mr Herbert Taylor.

1882-1904.

Professor J. D. Everett.

1882r*.

Professor A. Schuster.

1883.

Sir W. Siemens.

1883-*.

Dr J. A. Fleming.

]

XXIV

INTEODUCTION

1883-1900.

Professor G. F. Fitzgerald.

188a-*

Dr R. T. Qlazebrook.

1883-1897.

Professor G. Chrybtal.

1884-1891.

Mr H. ToMLiNSON.

1884-1891.

Professor W. Garnbtt.

1886-*

Sir J. J. Thomson.

1886-*

Dr W. N. Shaw.

1887-*

Dr J. T. BOTTOMLBY.

1888-1892.

Mr T. Gray.

1892-1900.

Professor J. Viriamu Jones.

1892-1910.

Dr G. Johnstone Stonby.

1892-*

Proffwsor S. P. Thompson.

1893-*

Rev. T. C. FiTZPATRICK.

1893-1897.

Mr G. Forbes.

1895-*

Mr J. Rennie.

1895-*

Principal E. H. Griffiths.

1896-*.

Sir A. W. RucKBR.

1898.

Professor A. G. Webster.

1899 *.

Professor H. L. Callendar.

1900-*

Mr Gborge Matthey.

1900-1902.

Sir W. C. Roberts- Austen.

1908-1909.

Mr A. P. Trotter.

1908-*.

Profes-sor T. Mather.

1908-*.

Mr F. E. Smith.

FIRST REPORT— CAMBRIDGE, 1862.

The Committee regret that they are unable this year to
submit a final Report to the Association, but they hope that the
inherent difficulty and importance of the subject they have to
deal with will sufficiently account for the delay.

The Committee considered that two distinct questions were
before them, admitting of entirely independent solutiona They
had first to determine what would be the most convenient unit of
resistance, and second what would be the best form and material
for the standard representing that unit. The meaning of this
distinction will be apparent when it is observed that, if the first
point were decided by a resolution in favour of a unit based on
Ptofessor Weber's or Sir Charles Bright and Mr Latimer Clark's
system, this decision would not affect the question of construction;
while, on the other hand, if the second question were decided in
&vour of any particular arrangement of mercury or gold wire as
the best form of standard, this choice would not affect the
question of what the absolute magnitude of the unit was to be.

The Committee have arrived at a provisional conclusion as to
the first question; and the arguments by which they have been
guided in coming to this decision will form the chief subject of
the present Report.

They have formed no opinion as to the second question, viz.
the best form and material for the standard.

In determining what would be the most convenient unit for
aU purposes, both practical and purely scientific, the Committee
were of opinion that the unit chosen should combine, as far as was
possible, the five following qualities.

1. The magnitude of the unit should be such as would lend
itself to the more usual electrical measurements, without requiring
the use of extravagantly high numbers of cyphers or of a long
series of decimals.

ax. 1

2 PRACTICAL STANDARDS

2. The unit should bear a definite relation to units which
may be adopted for the measurement of electrical quantity>
current, and electromotive force, or, in other words, it should
form part of a complete system for electrical measurements.

3. The unit of resistance, in common with the other units of
the system, should, so far as is possible, bear a definite relation to
the unit of work, the great connecting link between all physical
measurements.

4. The unit should be perfectly definite, and should not be
liable to require correction or alteration from time to time.

5. The unit should be reproducible with exactitude, in order
that, if the original standard were injured, it might be replaced^
and also in order that observers who may be unable to obtain
copies of the standard may be able to manu&cture them without
serious error.

The Committee were also of opinion that the unit should be
based on the French metrical system, rather than on that now
used in this country.

Fortunately no very long use can be pleaded in fiivour of any
of the units of electrical resistance hitherto proposed, and the
Committee were therefore at liberty to judge of each proposal by
its inherent merits only; and they believe that, by the plan which
they propose for adoption, a unit will be obtained combining to a
great extent the five qualities enumerated as desirable, although
they cannot yet say with certainty how far the fourth quality^
that of absolute permanency, can be ensured.

The question of the most convenient magnitude was decided by
reference to those units which have already found some acceptance*

These, omitting for the moment Weber's • 3 , were found to

range between one foot of copper wire weighing one hundred
grains (a unit proposed by Professor Wheatstone in 1843) and one
mile of copper wire of ^ inch diameter, and weighing consequently
about 84^ grains per foot. The smaller units had generally been
used by purely scientific observers, and the larger by engineers or
practical electricians.

Intermediate between the two lay Dr Werner Siemens's mercury
unit, and the unit adopted by Professor W. Thomson as approxi-^

mately equal to one hundred millions of absolute ^ . The

seconds

FOB ELECTRICAL MEASUREMENTS 3

former is approximately equal to 371 feet, and the latter to 1217
feet, of pure copper wire ^ inch diameter at 15° C. Both of these
units have been adopted in scientific experiments and in practical
tests ; and it was thought that the absolute magnitude of the unit
to be adopted should not diflFer widely firom these resistances.

The importance of the second quality required in the unit,
that of forming part of a coherent system of electrical measure-
ments, is felt not only by purely scientific investigators, but also
by practical electricians, and was indeed ably pointed out in a
paper read before this Association in Manchester by Sir Charles
Bright and Mr Latimer Clark.

The Committee has thus found itself in the position of deter-
mining not only the unit of resistance, but also the units of
current, quantity, and electromotive force. The natural relations
between these units are, clearly, that a unit electromotive force
maintained between two points of a conductor separated by the
unit of resistance shall produce the unit current, and that this
current shall in the unit of time convey the unit quantity of
electricity.

The first relation is a direct consequence of Ohm's law; and
the second was independently chosen by Weber and by the two
electricians above-named.

Two only of the above units can be arbitrarily chosen ; when
these are fixed, the others follow firom the relations just stated.

Sir Charles Bright and Mr Latimer Clark propose the electro-
motive force of a Daniell's cell as one unit, and choose a unit of
quantity depending on this electromotive force. Their resistance-
unit, although possessing what we have called the second requisite
quality, and superior consequently to many that have been pro-
posed, does not in any way possess the third quality of bearing
with its co-units a definite relation to the unit of work, and has
therefore been considered inferior to the equally coherent system
proposed by Weber many years since, but until lately com-
paratively little known in this country.

Professor Weber chose arbitrarily the unit of current and the
unit of electromotive force, each depending solely on the units of
mass, time, and length, and consequently independent of the
physical properties of any arbitrary material.

Professor W. Thomson has subsequently pointed out that this
system possesses what we have called the third necessary quality,

1—2

4 PRACTICAL STANDARDS

since, when defined in this measure, the unit current of electricity,
in pissing through a conductor of unit resistance, does a unit of
work or its equivalent in a unit of time*.

The entire connexion between the various units of measure-
ment in this system may be summed up as follows.

A battery or rheomotor of unit electromotive force will generate
a current of unit strength in a circuit of unit resistance, and in
the unit of time will convey a unit quantity of electricity through
this circuit, and do a unit of work or its equivalent.

An infinite number of systems might fulfil the above conditions,
which leave the absolute magnitude of the units undetermined.

Weber has proposed to fix the series in various ways, of which
two only need be mentioned here — ^first by reference to the force
exerted by the current on the pole of a magnet, and secondly by
the attraction which equal quantities of electricity exert on one
Another when placed at the unit distance.

In the first or electro-magnetic system, the unit current is
that of which the unit length at a unit distance exerts a unit of
force on the unit magnetic pole, the definition of which is depen-
dent on the units of mass, time, and length alone. In the second
or electrostatic system, the series of units is fixed by the unit of
quantity, which Weber defines as that quantity which attracts
another equal quantity at the unit distance with the unit force.

Starting from these two distinct definitions, Weber, by the
relations defined above, has framed two distinct systems of
electrical measurement, and has determined the ratio between
the units of the two systems — a matter of great importance in
many researches; but the electro-magnetic system is more con-
venient than the other for dynamic measurements, in which
currents, resistances, etc., are chiefly determined from observations
conducted with the aid of magnets.

As an illustration of this convenience, we may mention that
the common tangent galvanometer afibrds a ready means of
determining the value in electro-magnetic units of any current 7
as a function of the horizontal component of the earth's magnetism
Hy the radius of the coil 22, its length Z, and the deflection S.

* Vide *' Applioation of Electrical Effect to the Measurement of Electromotive
Force," PML Mag. 1861.

FOR ELECTRICAL MEASUREMENTS 5

In this Report, wherever Professor Weber's, or Thomson's, or
the absolute system is spoken of, the electro-magnetic system
only is to be understood as referred to. The immense value of a
coherent system, such as is here described, can only be appreciated
by those who seek after quantitative as distinguished from merely
qualitative results. The following elementary examples will
illustrate the practical application of the system.

It is well known that the passage of a cuiTent through a metal
conductor heats that conductor; and if we wish to know how
much a given conductor will be heated by a given current in a
given time, we have only to multiply the time into the resistance
and the square of the current, and divide the product by the
mechanical equivalent of the thermal unit. The quotient will
express the quantity of heat developed, from which the rise of
temperature can be determined with a knowledge of the mass
and specific heat of the conductor.

Again, let it be required to find how much zinc must be
consumed in a Daniells cell or battery to maintain a given
current through a given resistance. The heat developed by the
consumption of a unit of zinc in a Daniell's battery has been
determined by Dr Joule, as also the mechanical equivalent of that
heat; and we have only to multiply the square of the current into
the resistance, and divide by the mechanical equivalent of that
heat, to obtain the quantity of zinc consumed per unit of time.

Again, do we wish to calculate the power which must be used
to generate by a magneto-electric machine a given current of
(say) the strength known to be required for a given electric light ?

Let the resistance of the circuit be determined, and the power
required will be simply obtained by multiplying the resistance
into the square of the current.

Again, the formula for deducing the quantity of electricity
contained in the charge of a Ley den jar or submarine cable from
the throw of a galvanometer-needle depends on the relation
between the unit expressing the strength of current, the unit of
force, and the unit magnet-pole. When these are expressed in
the above system, the quantity in electro-magnetic measure is
immediately obtained from the ballistic formula. In estimating
the value of the various insulators proposed for submarine cables,
this measure is of at least equal importance with the measure of
the resistance of the conductor and of the insulating sheath ; and

6 PRACTICAL STANDARDS

the unit in which it is to be expressed would be at once settled
by the adoption of the general system described.

These five very simple examples of the use of Weber's and
Thomson's system might be multiplied without end; but it is
hoped that they will suffice to give some idea of the range and
importance of the relations on which it depends to those who may
hitherto not have had their attention directed to the dynamical
theory.

No doubt, if every unit were arbitrarily chosen, the relations
would still exist in nature, and by a liberal use of coefficients
experimentally determined the answer to all the problems
depending on these relations might still be calculated; but the
number of these coefficients and the complication resulting from
their use would render such an arbitrary choice inexcusable.

A large number of units of resistance have firom time to time
been proposed, founded simply on some arbitrary length and
section or weight of some given material more or less suited for
the purpose ; but none of these units in any way possessed what
we have called the second and third requisite qualities, and could
only have been accepted if the unit of resistance had been entirely
isolated from all other measurements. We have already shown
how far this is &om being the case ; and the Committee consider
that, however suitable mercury or any other material may be for
the construction or reproduction of a standard, this furnishes no
reason for adopting a foot or a metre length of some arbitrary
section or weight of that material.

Nevertheless it was apparent that, although a foot of copper
or a metre of mercury might not be very scientific standards, they
produced a perfectly definite idea in the minds of even ignorant
men, and might possibly, with certain precautions, be both
permanent and reproducible, whereas Weber's unit has no
material existence, but is rather an abstraction than an entity.
In other words, a metre of mercury or some other arbitrary
material might possess what we have called the first, fourth, and
fifth requisite qualities, to a high degree, although entirely
wanting in the second and third. Weber's system, on the
contrary, is found to fulfil the second and third conditions, but
is defective in the fourth and fifth ; for if the absolute or Weber's
unit were adopted without qualification, the material standard by
which a decimal multiple of convenient magnitude might be

FOR ELECTRICAL MEASUREMENTS 7

practically represented would require continual correction as
aaccessive determinations made with more and more skill deter*
mined the real value of the absolute unit with greater and
greater accuracy. Few defects could be more prejudicial than
this continual shifting of the standard. This objection would not
be avoided even by a determination made with greater accuracy
than is expected at present, and was considered fatal to the
unqu€dified adoption of the absolute unit as the standard of
resistance.

It then became matter for consideration whether the advan-
tages of the arbitrary material standard and those of the absolute
system could not be combined; and the following proposal was
made and adopted as the most likely to meet every requirement.
It was proposed that a material standard should be prepared in
such form and materials as should ensure the most absolute
permanency ; that this standard should approximate as nearly as

possible, in the present state of science, to ten millions of j-,

but that, instead of being called by that name, it should be known

simply as the unit of 1862, or should receive some other simpler

name, such as that proposed by Sir Charles Bright and Mr Latimer

Clark in the paper above referred to; that from time to time,

as the advance of science renders this possible, the difference

metre

between this imit of 1862 and the true ten millions of i-

seconds

should be ascertained with increased accuracy, in order that the

error, resulting from the use of the 1862 unit in dynamical

calculations instead of the true absolute unit, may be corrected by

those who require these corrections, but that the material standard

itself shall under no circumstances be altered in substance or

definition.

By this plan the first condition is fulfilled; for the absolute
magnitude of this standard will differ by only 2 or 3 per cent.
fix>m Dr Siemens's mercury standard.

The second and third conditions will be fulfilled with such
accuracy as science at any time will allow.

The fourth condition, of permanency, will be ensured so far as
our knowledge of the electrical qualities of matter will permit ;
and even the fifth condition, referring to the reproduction, is
rendered comparatively easy of accomplishment.

8 PRACTICAL STANDARDS

There are two reasons for desiring that a standard should be
reproducible : first, in order that if the original be lost or destroyed
it may be replaced ; second, in order that men unable to obtain
copies of the true standard may approximately produce standards
of their own. It is indeed hoped that accurate copies of the
proposed material standard will soon be everywhere obtainable^
and that a man will no more think of producing his own standard
than of deducing his foot-rule from a pendulum, or his metre from
an arc of the meridian; and it will be one of the duties of the
Committee to facilitate the obtaining of such copies, which can be
made with a thousandfold greater accuracy than could be ensured
by any of the methods of reproduction hitherto proposed.

It is also hoped that no reproduction of the original standard
may ever be necessary. Nevertheless great stress has been lately
laid upon this quality, and two methods of reproduction have
been described by Dr Werner Siemens and Dr Matthiessen
respectively; the former uses mercury, and the latter an alloy of
gold and silver, for the purpose. Both methods seem susceptible
of considerable accuracy. The Committee has not yet decided
which of the two is preferable ; but their merits have been dis-
cussed, fix)m a chemical point of view, in the appended Report C,
by Prof Williamson and Dr Matthiessen. An interesting letter
from Dr Siemens on the same point will also be found in the
Appendix E. This gentleman there advocates the use of a metre
of mercury of one square millimetre section at 0° C. as the
resistance-unit; but his arguments seem really to bear only on the
use of mercury in constructing and reproducing the standard, and
would apply as well to any length and section as to those which
he has chosen.

When the material 1862 standard has once been made,
whether of platinum, gold an alloy, or mercury, or otherwise, the
exact dimensions of a column of mercury, or of a wire of gold-
silver alloy, corresponding to that standard can be ascertained,
published, and used where absolutely necessary for the purpose
of reproduction.

It should at the same time be well understood that, whether
this reproduction does or does not agree with the original
standard, the unit is to be that one original material permanent
standard, and no other whatever, and also that a certified copy
must always be infinitely preferable to any reproduction.

FOR ELECTRICAL MEASUREMENTS ^

The reproduction by means of a fresh determination of the
absolute unit would never be attempted, inasmuch as it would be
costly, difficult, and uncertain; but, as alrefidy mentioned, the
difference between new absolute determinations and the material
standard should from time to time be observed and published.

The question whether the material standard should aim at an

approximation to the r or j was much debated. In

second second

&vour of the latter it was argued that, so long as in England feet

and grains were in general use, the -^ would be anomalous,

and would entail complicated reductions in djmamical calculations.

In favour of the , it was argued that, when new standards

second °

were to be established, those should be chosen which might be
generally adopted, and that the metre is gaining universal accept-
ance. Moreover the close accordance between Dr Siemens's unit

and the decimal multiple of the 3 weighed in favour of this

^ second "

unit; so that the question was decided in favour of the metrical

system.

In order to carry out the above views, two points of essential

importance had to be determined. First, the degree of accuracy

with which the material standard could at present be made to

correspond with the -j ; and second, the degree of permanency

which could be ensured in the material standard when made.

The Committee is, unfortunately, not able yet to form any
definite opinion upon either of these points.

Resistance-coils, prepared by Professor W. Thomson, have been
sent to Professor Weber ; and he has, with great kindness, deter-
mined their resistance in electro-magnetic units as accurately as
he could. It is probable that his determinations are very accurate;
nevertheless the Committee did not feel that they would be
justified in issuing standards based on these determinations alone.
In a matter of this importance, the results of no one man could
be accepted without a check. Professor Weber had made some
similar determinations with less care some years since, but he has
unfortunately not published the difference, if any, between the
results of the two determinations. Indirect comparisons between

10 PRACTICAL STANDARDS

the two determinations show a great discrepancy, amounting
perhaps to 7 per cent. ; but it is only fair to say that this error
may have been due to some error in other steps of the comparison,
and not to Professor Weber's determination. Meanwhile it was
hoped that a check on Weber's last result would by this time
have been obtained by an independent method due to Professor
Thomson. Unfortunately, that gentleman and Mr Fleeming
Jenkin, who was requested to assist him, have hitherto been
unable to complete their experiments, owing chiefly to their
occupation as jurors at the International Exhibition. The
apparatus is, however, now nearly complete, and it is hoped will
before Christmas give the required determinations.

If Professor Weber's results accord within one per cent, with
these new determinations, it is proposed that provisional standards
shall be made of German-silver wire in the usiial way, and that
they should be at once issued to all interested in the subject,
without waiting for the construction of the final material
standard.

The construction of this standard may possibly be delayed for
some considerable time by the laborious experiments which
remain to be made on the absolute permanency of various forms
and materials. An opinion is very prevalent that the electrical
resistances of wires of some, if not all, metals are far from
permanent; and since these resistances are well known to vary as
the wires are more or less annealed, it is quite conceivable that
even the ordinary changes of temperature, or the passage of the
electric current, may cause such alterations in the molecular
condition of the wire as would alter its resistance. This point is
treated at some length in the two Reports B and C, appended,
by Professor Williamson and Dr Matthiessen. The experiments
hitherto made have not extended over a sufficient time to
establish any very positive results ; but, so far as can be judged
at present, some, though not all, wires do appear to vary in
conducting power.

Mercury would be free from the objection that its molecular
condition might change ; but, on the other hand, it appears firom
Report C that the mercury itself would require to be continually
changed, and that consequently, even if the tube containing it
remained unaltered (a condition which could not be absolutely
ensured), the standards measured at various times would not

FOB ELECTRICAL MEASUREMENTS 11

really be the same standard. A possibility at least of error would
thus occur at each determination, and certainly no two successive
determinations would absolutely agree. If, therefore, wires can
be found which are permanent, they would be preferred to
mercury, although, as already said, no conclusion has been come
to on this point.

Some further explanation will now be given of the resolutions
passed from time to time by the Committee, and appended to this
Report.

Dr Matthiessen was requested to make experiments with the
view of determining an alloy with a minimum variation of resis-
tance due to change of temperature. The object of this research
was to find an alloy of which resistance-coils could be made
requiring little or no correction for temperature during a series
of observations. A preliminary Report on this subject is appended
(A), in which the curious results of Dr Matthiessen's experiments
on alloys are alluded to, and, in particular, the following fact con*
nected with the resistance of alloys of two metals is pointed out.

Let us conceive two wires of the two pure metals of equal
length, and containing respectively the relative weights of those
two metals to be used in the alloy. Let us further conceive these
two wires connected side by side, or, as we might say, in multiple
arc. Then let the difference be observed in the resistance of this
multiple arc when at zero and 100^*0. This diflference will be
found almost exactly equal in all cases to the difference which will
be observed in the resistance of a wire drawn from the alloy
fonned of those two metal wires at zero and 100"", although the
actual resistance at both temperatures will in most cases be very
much greater than that of the hypothetical multiple arc.

Li order to obtain a minimum percentage of variation with a
change of temperature, it was consequently only necessary to
make experiments on those alloys which offer a very high resist-
ance as compared with the mean resistance of their components.
The results of a few experiments are given in the Report, but
these are only the first of a long series to be undertaken.
Hitherto an alloy of platinum and silver is the only one of which
the conducting power and variation with temperature are less
than that of German-silver.

Professor W. Thomson and Dr Matthiessen were requested to
examine the electrical permanency of metals and alloys. A pre-

12 PRACTICAL STANDARDS

liminary Report on the subject by Dr Matthiessen is appended (B),
in which he shows that, after four months, one copper and two
silver hard-drawn wires have altered, becoming more like annealed
wires, but that no decided change has yet been detected in the
great majority of the wires.

Several eminent practical electricians were requested to advise
the Committee as to the form of coil they considered most suitable
for a material standard, and also to furnish a sample coil such as
they could recommend. Sir Charles Bright informed the Com-
mittee that he was ready to comply with the request. The point
is one of considerable importance, respecting which it was thought
that practical men might give much valuable information. Coils
of wire may be injured by damp, acids, oxidation, stretching and
other mechanical alterations. They may be defective from im-
perfect or uncertain insulation; and they may be inconveniently
arranged, so that they do not readily take the temperature of the
surrounding medium, or cannot be safely immersed in water or
oil baths, as is frequently desirable. No definite conclusion as
to the form of coil to be recommended, even for copies, has been
arrived at.

It was resolved ''That the following gentlemen should be
informed of the appointment of the present Committee, and
should be requested to furnish suggestions in furtherance of its
object: —

Professor Edlund (XJpsala).
Professor T. Fechner (Leipzig).
Dr Henry (Washington).
Professor Jacobi (St Petersburg).
Professor G. Kirchhoff (Heidelberg).

Professor Neumann (Konigsberg).
Prof essor J. C. Poggendorff (Berlin).
M. Pouillet (Paris).
Werner Siemens, Ph.D. (Berlin).
Professor W. K Weber (Gottin-

Professor G. Matteucci (Turin). S^^)"

A letter, appended to this Report, was consequently addressed
to each of these gentlemen. Answei-s have been received from
Professor Kirchhoff and Dr Siemens, which will be found in the
Appendix. The resolution arrived at by the Committee to con-
struct a material standard will entirely meet Professor Eirchhoff's
views. The Committee have been unable entirely to adopt
Dr Siemens's suggestions ; but his statements as to the accuracy
with which a standard can be reproduced and preserved by
mercury will form the subject of further special investigation^

FOR ELECTRICAL MEASUREMENTS 13

and the Committee will be most happy to take advantage of
his kind offer of assistance.

A letter was also received from Sir Charles Bright, containing
an ingenious method of maintaining a constant tension or difference
of potentials. This point will probably come before the Committee
at a later period, when Sir Charles Bright's suggestion will not be
lost sight of.

The Committee also received, on the 29th ultimo, after the
present Report had been drawn up, a letter from Dr Elsselbach, a
well-known electrician, who had charge of the electrical tests of
the Malta and Alexandria Cable during its submergence. In this
letter Dr Esselbach arrives at substantially the same conclusions
as thoee recommended by the Committee. Thus, his first conclusion
is ''to adopt Weber's absolute unit substantially, and to derive
from it, by the multiple 10^^ the practical unit." This practical
unit is precisely that recommended by your Committee. Dr Essel-
bach uses the multiple 10*®, starting from the ^ — , where

your Committee recommend the multiple 10\ starting from the

-;: the result is the same.

second

Dr Esselbach's next conclusion is also of great practical value

He points out that the electro-magnetic unit of electromotive

force, also multiplied by 10^^ differs extremely little from that

of the common Daniell's cell, and that, without doubt, by proper

care such a cell could be constructed as would form a practical

unit of electromotive force. This suggestion has the approval of

the Committee. Dr Esselbach next points out that the unit of

resistance which he proposes differs very little from Dr Siemens's

mercury unit, which he, like your Committee, considers a great

advantage; and the difference is, indeed, less than he supposes.

He also proposes to use Weber's absolute unit for the unit of

current — a suggestion entirely in accordance with the foregoing

Report; and he further points out that this current will be of

convenient magnitude for practical purposes. He next approves

of the suggestions of Sir Charles Bright and Mr Latimer Clark

with reference to nomenclature and terminology. In the body of

the Report he gives some valuable data with reference to the

nnit of quantity, which he defines in the same manner as your

Committee. This result will be analyzed in the Report which

14 PRACTICAL STANDARDS

Professor W. Thomson and Mr Fleeming Jenkin will make on the
fresh determination of the absolute unit of resistance.

The Committee attach high importance to this communication^
showing as it does that a practical electrician had arrived at many
of the very same conclusions as the Committee, quite indepen-
dently and without consultation with any of its members.
Dr Esselbach has omitted to point out, what he no doubt was
well aware of, that, if, as he suggests, two equal multiples of
the absolute units of resistance and electromotive force are
adopted, the practical unit of electromotive force, or Daniell's cell^
will, in a circuit of the practical unit of resistance, produce the
unit current.

Mr Fleeming Jenkin was requested to furnish an historical
summary of the various standards of resistance, but he has been
unable to complete his Report in time for the present meeting.

Professor Williamson and Dr Matthiessen were requested to
put together the facts regarding the composition of the various
materials hitherto used for standards of resistance, and the
physical changes they were likely to undergo. Wires of pure
solid metals, columns of mercury, and wires of alloys have been
used for the purpose. The Report of the above gentlemen is
appended (C). In it they arrive at the following conclusions : —

Firstly, with reference to pure metals in a solid state, they
consider that the preparation of those metals in a state of
sufficient purity to ensure a constant specific resistance is ex-
ceedingly difficult, as is proved by the great discrepancy in the
relative conducting powers obtained by diflferent observers. Elec-
trotype copper is excepted from this remark. They also point
out that the influence of annealing on the conducting powers of
pure solid metals is very great, and would render their use for the
purpose of reproducing a standard very objectionable, inasmuch
as it is impossible to ensure that any two wires shall be equally
hard or soft. They observe that errors of the same kind might
be caused by unseen cavities in the wires, and give examples of
the actual occurrence of these cavities. They point out another
objection to the use of pure solid metals as standards, in the fact
that their resistance varies rapidly with a change of temperature,
so that slight errors in a thermometer or its reading would
materially affect the results of an experiment.

Secondly, with reference to mercury, they show that it is

FOR ELECTRICAL MEASUREMENTS 15

comparatively easily purified, varies little in resistance with a
change of temperature, and can undergo no change analogous to
that caused by annealing, but that, on the other hand, measure-
ments of its conducting-power by different observers vary much,
that the tube used cannot be kept full of mercury for any length
of time, as it would become impure by partial amalgamation
with the terminals, and that consequently each time a mercury
standard is used it has, practically, to be remade. The accuracy
with which different observers can reproduce mercury-standards
has not been determined

Thirdly, with reference to alloys, they say that there is better
evidence of the indejpendent and accurate reproduction of a
standard by a gold-silver alloy of certain proportions than by pure
solid metal or by mercury. They point out that annealing and
changes of temperature have far less effect on alloys than on pure
metals, and that consequently any want of homogeneity or any
error in observing the temperature during an experiment is, with
alloys, of little consequence, but that, on the other hand, the
existence of cavities must be admitted as possible in all solid
wires. They are of opinion that the permanence of jewellery
affords strong ground for believing that a gold-silver alloy will be
quite as permanent as any solid pure metal ; and in the course of
the Report they point out some curious facts showing that a
great change in the molecular condition of some pure metals and
alloys may occur without any proportional change in their con-
ducting powers.

Finally, they recommend that practical experiments should be
made independently by several gentlemen to determine whether
mercury or the gold-silver alloy be really the better means of
reproducing a standard.

The main resolution arrived at by the Committee, viz. that a
material standard shall be adopted which, at the temperature of

17- C. shall approximate to 10' ^^^. as far as present data

allow, has been already fully explained. It was not arrived at
until after several meetings had been held, and the merits of the
various proposals fiilly discussed.

This resolution was passed (unanimously) at a meeting when
five out of the six members of the Committee were present

It was at the same time resolved that provisional copies

16 PRACTICAL STANDARDS

should be distributed at the present meeting. The circumstances
have been akeady explained which have prevented this resolution
from being carried into effect.

It was thought desirable that an apparatus should be designed
which could be recommended by the Committee for use in
copying and multiplying the units to be issued, since it is certain
that some of the glaring discrepancies in coils intended to
agree must have been due to defective modes of adjustment.
Mr Fleeming Jenkin has consequently designed an apparatus for
the purpose, of which a description is appended. Messrs Elliott
Brothers have kindly constructed a couple of these instruments,
which may be seen in action by members interested in this
subject.

The present Report was drawn up by Mr Jenkin, and adopted
at a meeting of the Committee on the 30th ultimo.

Appendix to Report on Standards of Electrical Resistance.

A. On the variation of the electrical resistance of alloys due
to change of temperature, by Dr Matthiessen, F.R.S.

B. On the electrical permanency of metals and alloys, by
Dr Matthiessen, F.R.S.

C. On the reproduction of electrical standards by chemical
means, by Professor Williamson, F.R.S., and Dr Matthiessen, F.R.S.

D. Professor Kirchhoff's letter.

E. Dr Siemens's letter.

F. Dr Esselbach's letter.

G. Circular addressed to foreign men of science.

H. Description of apparatus for copying and multiplying the
units of resistance.

Appendix A. — On the Variation of the Electrical Resistance of Alloys
due to Change of Temperature. By Dr Matthiessen, F.R.S.

It has been shown* that the influence of temperature on the
electric conducting power of the metals amounts to 29*3 per cent,
on their conducting power between 0° and 100° C. : an exception
to this law has been found in ironf, the conducting power of
which decreases between those limits 38*2 per cent. It was,

• Phil Tram, 1862, pt. 1.

t MatthieBsen and Vogt, unpublisbed researches.

FOR ELECTRICAL MEASUREMENTS 17

therefore, useless to try any of the other pure metals, as they
would, in all probability, have decreased by the same amount, as
well as from the (act that the metals which would have suited the
purpose had already been tried. I therefore turned my attention
to the alloys, and, in conjunction with Dr C. Vogt, have made a
long series of experiments respecting the influence of temperature
on their electric conducting power. After having determined the
conducting power of a few of them at different temperatures,
together with the help of the few experiments which have already
been made by different observers, it became obvious that the
percentage decrement in their conducting power stands in some
relation to the &ct that, when a solid metal is alloyed with
another (with the exception of lead, tin, zinc, and cadmium
amongst each other), a lower conducting power is observed than
the mean of that of the components*. The law which we found
to regulate this property was with most alloys the following, viz.:— >

" The percentage decrement between 0^ and IW in the conduct-
ing power of an alloy in a solid state stands in the same raJtio to
the mean percentage decrement of the components between 0^ and
IW as the conducting power of the alloy at 100° does to the mean
conducting power of the components at 100*"'; or, in other words,
*'the absolute difference in the observed resistance between 0"" and
100° of an aUoy is equxd to the absolute difference between the
means of the resistance of the component metals between 0° and 100°."

For example, the conducting power of the hard-drawn gold-
silver alloy was found equal to 15'03 at 0° (taking silver equal
100 at 0""), and decreases 6*49 per cent, between 0° and 100°. The
mean decrement of the components between 0° and 100° being
29*3 per cent., the conducting power of the alloy is 14'05 at 100°,
and that of the mean of the components is 62*58 at 100°. If we
now calculate the percentage decrement in the conducting power
of the alloy between 0° and 100° from the above data, we find it
equal to 6*58 per cent., and by experiments it was found equal to
6*49 per cent. Or, taking the resistance of silver at 0° = 100, and
that of gold at 0° = 128*3, we find the resistance of the alloy at
0°» 665*3,- and at 100° » 711*7, and that calculated from a mean of
the volumes of its components at 0°a 113*2, and at 100° = 159*8;

* Awmfnc thftt the eondqating power or neiatoiioe of an alloy is equal to that
of parallel wiiea of the eomponeiita Ittmuiig it

K JL 2

18 PRACTICAL STANDARDS

therefore the absolute difference between the observed resistance
at O"" and KW is 46*4, and that between the calculated at 0° and
100° = 46-8.

Knowing already, from my experiments on the electric con-
ducting power of alloys*, that when two metals are alloyed
together in any proportion, if the alloy is merely a solution of the
two metals in one another, its conducting power may be approxi-
mately foretold, and that, from the above law, it is necessary that
if the conducting power of an alloy should vary between the limits
of O'' and 100"" to a minimum extent, the alloy itself must have a
minimum conducting power as compared with that calculated
from its components, — I at once foresaw that it would be useless,
as was afterwards proved by the research made in conjunction
with Dr Vogt, to make any experiments with the two metal-
alloys, which may be looked upon as a solution of one metal in
the other, as no practical alloy would be found which would vary
in its conducting power between 0° and 100^ to a small extent.
It must also be borne in mind that the alloy sought for must be
a ductile one, capable of being drawn into wire, not too soft, as
would easily be damaged by covering and winding, easily pro-
duced, and cheap in price. Bearing this in mind, we turned our
attention to some three metal-alloys, thinking that we had some
chance there of obtaining a good result; for it is well known that
the conducting power of German-silver wire varies in such a slight
extent between 0° and 100°.

It also appeared worth while to experiment with some of
those alloys which may perhaps be considered chemical com-
binations, or to contain such, as, for instance, platinum and silver ;
and, on account of their other physical properties, the platinum-
iridium alloys were also experimented with.

In the following Table I give the results obtained in conjunc-
tion with Dr Vogt. The unit here taken for comparison is that of
a hard-drawn silver wire at 0°, The normal wires were made of
German-silver, and in order to obtain their values in terms of
hard-drawn silver, they were compared with the gold-silver alloy.
In these experiments it was thought better first to use those
pure metals which are easily obtained, so as to learn something
regarding the manner in which the three metal-alloys behave,
and then try some alloys made of the cheaper commercial metals.

• Phil. Trans. 1860, p. 161.

FOB ELECTBICAL MEASUBEMENTS

As will be seen by the Table, only the first part has been as yet
carried out.

Table.

(With each Series, the formula deduced from the observations
for the correction of the conducting power of the alloy for
temperature is given, when \ is equal to the conducting power
at the temperature f C.)

Composition of alloy Weight

(1) Gold 68-3

Copper 26'5

Silver 15'2

Made from pure metals.

Hard -drawn.

Gompotttioii of alloy

(2) Gold

Silver ,,

Copper

Made of pure metals,
fiard-diawn.

Weight

e6-5

18-1
15-4

10-5637
10-4341
10-3130
101846
10D862

Length 532 mm. ; diameter 0-625 mm.

Conducting power

T. Found

9-0 11-956

53-5 11-674

100-0 11-438

X= 12-017 - 0-0069033<+0-0000111 A

This alloy was taken, as Elarmarsch states it is the hardest and
most elastic of all the gold-silver-copper alloys.

Length 341*5 nun. ; diameter 0*618 mm.

Conducting power
T. Found

10-95
33-52
55-15
78-35
97-52

X= 10-6220 - 0-0056248« + 0-0000009863<3.

This alloy was tried, as it corresponded to equal volumes of
gold-copper and gold-silver, and these again correspond to an
alloy possessing the lowest conducting power of any of those made
of gold-copper or gold-silver.

Length 764 mm. ; diameter 0*553 mm.

Conducting power
T. Found

11*0 45-591

55-5 40*333

100*0 37*560

X = 44-472 - 0-081525i + 0-00b3240<*.

This alloy was taken to see the effect such a combination

would have.

Length 244 mm. ; diameter 0*682 mm.

Conducting power
T. Found

120 4-506

560 4*384

lOOO 4^71

Composition of alloy Weight

(3) Copper 78-3

Silver 14*3

Gold 7-4

Made from pure metals.

Hard -drawn.

Composition of aUoy

(4) Platinum
Iridium
Commercial alloy.
Hard-drawn.

Weight

66-6
33-4

Xt«4-541 - 0*0029307^ +0000002724^.

2—2

PBACTICAI, 3TANPARDS

This alloy was tried, as it possesses very great elasticity and
does not become softer on annealing. On account of these
properties, as well as its permanency in air (not oxidizing on its
sur&ce), it would serve exceedingly well for makkig springs and
contacts for electric and telegraphic apparatus.

Lengtb 381*5 mm. ; diameter 0*451 mm.

Condaoting power
T. Found

120 31-173

56-0 29-560

100-0 28-068

X = 31 -640 - 0039363< + 0-00003642<«.

This and the following two alloys were taken, as they probably
contain chemical combinations.

Composition of alloy Weight

(6) Silver 95-0

Platinum ... 6*0
Made from pure silver and
commercially pure platinum.
Hard-drawn.

Gomposition of alloy Weight

(6) Silver 90*2

Platinum ... 98
The metals emjdoyed were

the same as in No. 5.
Hard-drawn.

LoDgth 708 mm. ; diameter 0-26 mm.

Oondacting power

9-0

54-5

100-0

Found

17-920
17-319
16-767

Ck>mpo8ition of alloy

(7) SUver...
Platinum
Commercial alloy.
Hard-drawn.

X= 18-045 -0-013960* -f 0-00001 183^.

Length 169 mm. ; diameter 0*408 mm^

Weight

66-6
33-4

Condaoting power
Found

8-270 6-6850

54-00 6-5826

99-90 6-4987

X= 6-7032 - 0-0022167^ + 0*000001394^.

In the following Table I have given the results in such a.
manner that they may be easily compared.

Table.

•

Condaoting

power

atO°

Peroentage variation in.
conducting power be-
tween QP and I00»

Pure iron

> • . •

• *•

38-2

Other pure metals in a solid state...

— -

29-3

Alloy 3

» ...

44-5

16-5

» 5

1 ...

31-6

11-3

j> " •••

...

18-0

,^ Gold-silver* ..

• ...

15-0

6-5

» 4

I ...

4-5

5-9

» 2

1 ...

10-6

5-2

yi 1 ... ..<

• a.

12-0

4-8

,y German-silver t

• ■•

7-8

4-4

H f •.• ..a

• . •

6-7

3-1

* Phil Mag, Feb. IS6I.

t Phil Tram. 1862, pt 1.

FOR ELBCTRIGAL MEASUREMENTS 21

The method and apparatus employed for the above deter-
minations, together with the precautions taken to ensure correct
results, have already been described*. We have made only three
observations between 0"* and 100'', for it was found that they gave
almost exactly the same formula for the correction of the con«
ducting power for temperature as if we had taken seven or more
observations between 0"" and lOO"". Each of the above values for
the conducting power, at those temperatures, is the mean of three
or more observations. It was easy to obtain the desired tempera-
tures as a mean of several observations, after very little practice.
I have no doubt that, in the course of our experiments, we shall
be able to find an alloy the conducting power of which will
decrease between 0° and 100° even less than that of silver-
platinum. The experiments are being continued, and I hope,
before the next meeting of the Association, to be able to lay
before you results which will throw more light on the subject, as
well as to propose an alloy with a minimum variation in its
conducting power due to change of temperature, which may be
made commercially in a cheap manner of the common commercial
metals, and possessing those properties which are essential that it
should have.

Appendix B. — On the Electrical Permanency of Metals and
Alloys. By Dr Matthiessen, F.R.S.

Having, in conjunction with Prof. Thomson, been requested by
your Committee to make some experiments on this subject, we
thought it advisable for one of us to undertake some preliminary
experiments in which all possible disturbing causes were isolated.
The chief of these are : — oxidation by the oxygen of the air, as well
as by acids produced by the oxidation of the oil or grease with
which a wire is almost always covered when drawn, as the holes in
the draw-plates are generally oiled or greased ; stretching during
the process of covering and winding ; and after being wound on the
bobbin, elongation by expansion or contraction, owing to variations
of temperature, etc. These, I think, have been obviated in the
following manner : — The wires were carefully wound round a glass

* PMl. Tram. 1863, pt. 1.

22 PRACTICAL STANDARDS

■tube in ord^r to bring them into a smaller compass, and after taking
them off, they were placed inside wide glass tubes, and soldered
to two thick copper wires, these having been previously passed
through corks which fitted into the ends of the glass tube ; through
jeach of the corks a small glass tube passed, drawn out in the
middle to enable it to be drawn off easily, and sealed hermetically
by a lamp. The wire being soldered to the thick copper connectors,
and the corks fitted into the tube, dry c€irbonic-acid gas was led
through it for the space of about six hours, for the purpose of
drying it perfectly, as well as of displacing the air contained in it ;
after which the small glass tubes were melted off at the points,
when they had been previously drawn out. Tin caps, filled with
melted marine glue, were then fitted over the corks and the ends of
the tube, to prevent diffusion of the carbonic acid and air through
the corks. The whole of the tin caps outside, as well as those
parts of the copper- wire connectors which dipped in water of the
bath in which they were placed whilst being tested, were covered
with a thick coating of marine glue.

The wires experimented with were as follows : —

1. Silver: hard-drawn 1 n i. a. j.^

^ a'l 1 J r ^ut from the same piece: pure.

2. Silver: annealed J xr > r

3. Silver: hard-drawn 1 Cut from the same piece, but different

4. Silver: annealed J from 1 and 2; pure.

6. Copper: hard-drawn 1 ri x a. xi_

^ rt 1 J r Cut from the same piece: pure,

6. Copper: annealed J r ? ir

7. Copper: hard-drawn...

8. Copper: annealed

9. Qold: hard-drawn .

10. Qold: annealed

11. Qold: hard-drawn

12. Qold: annealed

13. Platinum: hard-drawn ...) i^ j./. xi • i
,,-.,. 1^ J, . r Cut from the same piece ; commercial.
14 Platinum: haid-drawn ...J ^ »

15. Qold-silver alloy : hard-drawn 1 Cut from the same piece. Made by

16. Qold-silver alloy : hard-drawn J Messrs Johnson and Matthey.

I Cut from the same piece. No. 19 ar-
ranged with longer connectors, and
used as normal wire with which the
rest were compared.

...1 Cut from the same piece, but different
...j from 5 and 6; pure.

)Cut from the same piece ; pure.

... ) Cut from the same piece, but different
...J from 9 and 10; pure.

17. Qerman-silver : annealed

18. Qerman-silver: annealed

19. Qerman-silver: annealed

The reason why duplicates were made in each case was that,
in case any of them should by any cause get damaged, the experi-
ments might be continued with the duplicate. When being tested,

FOR ELECTRICAL MEASUREMENTS 23

they were placed in a large bath containing from 40 to 50 litres of
water. By testing the wires at 20° it was found easy to keep
that temperature in the bath, during the experiments, to 0°*1
or 0°-2.

Up to the present time, that is to say, four months since they
were first tested, the conducting power of the wires 1, 3, and 5 has
altered, owing to becoming, in all probability, partially annealed.
Wire 8 has also altered materially, having decreased in conducting
power 3'5 per cent. : this decrement may be possibly due to bad
soldering. The differences found with the other wires are so very
small, that it is impossible to say whether they have altered or
not; for O**'! or 0**'2 will account for them. It was, therefore,
thought better to wait for another two or four months before
giving an opinion as to whether they alter or not; for as the wires
are in tubes and only surrounded by carbonic acid, we can never
be absolutely sure that the wire has exactly the same temperature
as the bath, more especially when it is coQsidered that each time
the connexion with the battery is made the wire becomes somewhat
heated.

If, two or four months hence, they still show no difference in
their conducting powers, it is proposed to expose the one set to
variations of temperature such as may occur (for instance, from 0^
to 40°), and then, should no change occur in their conducting
powers, to lead a weak current through them, say, for a month ;
for it has been asserted that a current passing through wire causes
a permanent change in its conducting power.

If, after these experiments the conducting power of the wires
remcdns unaltered, the different forms of resistance-coils, made
from those wires which have shown themselves permanent, will
then be tested in order to prove which is the best form of coil for
the British-Association unit.

24 PBACnCAL STANDARDS

Appendix C.—On the Reproduction of Electrical Standards by
Chemical Means. By Professor Williamson, F.R.S., and
Dr Matthiessen, F.RS.

In the following Report we have discussed, more especially
from a chemical point of view, the relative merits of the different
propositions which have been made to reproduce standards of
electric resistance, and have treated them under three heads : —

I. Those reproduced by a given length and section or weight, at
a given temperature, of a pure mstal in a solid state,

II. Those reproduced by a given length and section or weight, at

a given temperature, of a pure metal in a liquid state,

III. Those reproduced by a given length and section or weight, at

a given temperature, of an alloy.

The points on which we shall speak will be : —

1. On their preparation in a state of purity.

2. On their homogeneity and their molecular condition.

3. On the effect of annealing on their conducting power,

4. On the influence of temperature on their conducting power.

L Those reproduced by a given length and section or weight, at
a given temperature, of a pure metal in a solid state.

As type of this class we have chosen copper, for it has been
more extensively used as a unit of electric resistance, both by
scientific as well as by practical men, than any other metal or
alloy ; but what we are about to say regarding copper will hold
good in almost every case for all pure metals in a solid state.

1. On its preparation in a state of purity. — As traces of foreign
metals materially affect the conducting power of most pure metab,
it is of the utmost importance that those used for the reproduction
of units of electric resistance should be absolutely chemically pure.
The difficulty in obtaining absolutely pure metals even by chemists
is very great. Thus, for instance, Becquerel* found the conducting
power of pure gold at 0** equal to 68*9, compared with that of pure
silver at 0"" equal to 100 ; whereas, under the same circumstances,
Matthiessen and Von Bosef found it equal to 77*9 — showing

* Ann. de Chim. et de Phys. (1846), t. xvn. p. 242.
t Phil. Tram. 1862, pt. 1.

FOB ELECTRICAL MEASUREMENTS

a difference of about 12 per cent, in the values observed for
the conducting power of gold, prepared pure by different
chemists. This difference may be due to the silver not being
pore, or to all of them being more or less pure. Now when we
consider that these standards are required by electricians and
other physicists who have little or no acquaintance with chemical
manipulation, and that the cost of the preparation of absolutely
pure metals by scientific chemists would be very expensive on
account of the time and trouble they require, we think that this
tact alone constitutes a very serious drawback to their use as
a means for the reproduction of standards of electric resistance.
From the experience which one of us has had on this subject, it is
more than probable that if pure metals be prepared by different
chemists in the ordinary way of business, variations in their con-
ducting power would be found equal to several per cent. Thus,
copper supplied as pure by a well-known assayer had a conducting
power equal to 92, whereas pure copper conducts at the same
temperature 100*. Again, the pure gold of the assayer conducts
only 65'5, whereas pure gold at the same temperature would have
a conducting power equal to 73f. In order to show that the
conducting power of commercial metals varies to a great extent,
we give in the following Table (X) the values found for that of
the different coppers of commerce ; and it will be evident from it,
that to take a given length and weight or section of a commercial
metal as unit, as has often been done, is very wrong, and can only
lead to great discrepancies between the results of different observers.

Table X,+

(All the wires were annealed.)

Pure copper ... ... ... ...

Lake Superior native, not fused

Ditto, fused, as it oomes in oommeroe

Bum Burra ...

Best selected ...

Bright copper wire

Tough copper...

Demidoff

Rio Tinto

Similar variations will be found with most other metals, and
we shall give examples of these further on.

* Proeeedingi of the Royal Society , vol. zi. p. 126.

t Phil. Tram. 1S60, p. 176.

X Report of the Oovemment Submarine Cable Committee^ p. 835.

Condaeting power

... 100-0 at 15-5

... 98-8 ,

, 15-5

... 92-6 „

, 15-0

... 887 „

, 14-0

... 81-3 „

, 14-2

... 72-2 „

» 16-7

... 71-0 „

, 17-3

... 69-3 „

» 12-7

14-2 ,

, 14-8

26 PBACTICAL STANDARDS

2. 071 its homogeneity and its molecular condition. — It is well
known that the wires of some metals require much more care in
drawing than in others : thus, copper and silver, if not annealed
ofben enough during the process of drawing, will often become
quite brittle, and break oflF short when bent. Now, if the fracture
be closely observed, it will be seen that the wire is hollow ; in
fact, wherever it is broken, cavities will be found, and sometimes
of a millimetre or two in length ; so that such a wire may almost
be regarded as a tube with a very fine bore. The reason of this is
simply that in not annealing the wire often enough, the interna)
part of it becomes hard and brittle, whilst the outside remains
annealed from the heat evolved by its passage through the holes
of the draw-plates ; after a time, however, the inside, being very
brittle, will give way, whilst the outside is still strong enough to
bear the force used in drawing it through the draw-plates. These
places in the wires are easily discovered on drawing the wire finer;
for then at these points the wire slightly collapses, owing to the
quicker elongation of the weak points by the force used in drawing.
Silver and copper are the only metals which have been experi-
mented with in this manner; we ej^ therefore unable to say
whether it may occur with the other metals. However, although
no such wires could be used for experiments, yet what has been
shown possible to occur to such a marked extent when purposely
trying to obtain such results, may occur to some slight extent,
especially when great care is not used, and when the wires are
drawn by difiierent persons This may explain why, with some
metals and alloys of the same preparation, conducting powers are
often obtained which vary several per cent. For instance, W. Thom-
son* found the conducting power of several alloys of copper which
he had had made and tested to alter considerably on being drawn
finer; some of them were faulty from the cause we have just
mentioned, and, on their being drawn finer, these places showed
themselves and were then cut away.

It has also been shown f that when copper wire is heated to
100° for several days, a permanent alteration takes place in its
conducting power: thus, with the first wire experimented on, it
increased almost to the same extent as if it had been annealed.
With the second wire the increment was not nearly so large as

• Proceedings of the Royal Society, vol. xi. p. 126.
t Phil, Traru. 1862, pt. 1.

FOR ELECTRICAL MEASUREMENTS 27

with the first, and with the third it hardly altered at all. That
this is not due to one or the other of the wires being faulty in the
just-mentioned manner is proved,

1st, By the close agreement in the conducting powers.

2nd, By the close agreement between the differences in the
values found for the conducting powers of the hard-drawn and
annealed wires. They were: —

iBt wire at 0° 2nd wire at 0° Srd wire at 0°

Hard-drawn 99*6 lOOO 100*3

Annealed lOl'S 1021 1022

The values given for the hard-drawn wires are those which
were observed before the wire was heated at all.

Srd, That the same occurs with pressed wires: thus, with bismuth
it was found that the pieces of the same wire behaved differently ;
wire 1 showing, after 1 day's heating to 100'', an increment in the
conducting power of 16 per cent., whereas wire 2 increased, although
a piece from the same length of wire, 9 per cent.

Again, take the case of tellurium, and taking the conducting
power of each bar at first equal to 100, we find that the conducting
power of bar 1 had decreased after 13 days' heating to 4, where it
then remained constant, that of bar 2 after 32 days became
constant at 19, and that of bar 3 after 33 days at 6.

The cause of these marked changes in the conducting power
must therefore be looked for in the molecular arrangement of the
wires or bars employed. In the case of copper, they may be, and
probably are, due to the partial annealing of the wires ; for we find
that wire 1, although the conducting power increased afber having
been kept at 100° for several days almost to the same extent as if
it had been annealed, yet, on annealing it, it only gained as follows
(the results obtained with wires 2 and 3 are added) : —

l8t wire at 0"" 2nd wire at (P Srd wire at 0°

Hard-diawn 995 1000 1003

Aaer being kept several) ^^^.^ ^^^.^ ^qq.q

days at 100 )

After anneaUng 101-8 1021 102-2

The above shows that, in all probability, the annealing plays
here a part, but not the whole, in the change ; for otherwise why
do the wires behave differently ? This point will be fully discussed
in another Report which will be laid before your Committee, and

28 PRACTICAL STANDARDS

in which it will be shown where the hard-drawn wires become
partially annealed, and annealed wires partially hard-drawn, by age.

It is a curious fact that a change in the molecular arrangement
of the particles of wire of some metals which may be considered
homogeneous has very little effect on its electric conducting power.
Thus pure cadmium*, which when cold is exceedingly ductile,
becomes quite brittle and crystalline at about 80°, and returns
again to its ductile condition on cooling, shows no marked change
in its conducting power at that temperature ; in fact, it behaves
as if no such change had taken place. Again, when iron wire is
heated in a current of ammonia it becomes perfectly brittle and
crystalline, without altering its conducting power to any marked
extent.

That a wire which changes its molecular condition in becoming
crystalline does not necessarily materially alter in its conducting
power, is an important as well as a very interesting point, and has
also been proved in the case of German-silver.

3. On the effect of annealing on the condvcting power. — When
hard-drawn wires of silver, copper, gold, etc. are heated to redness
and cooled slowly, they become much softer, and on testing their
conducting powers they will be found to have increased thus : —

Silver Copper Gold Aooording to

Taking the hardHirawn) ^q^^ jq^^ jq^^

Wire aa )

The annealed will be... 107*0 102*6 101*6 Becquerelt

Ditto 109-0 102*3 102*0 j^VonB^l"^

Ditto 110*0 106*0 — Siemens§

Now there ia a certain difficulty in drawing a wire which is
hard-drawn ; and if annealed wires be used for the reproduction
of standards, the molecular condition, or perhaps the process of
annealing, has an influence on the increment of the conducting
power. Thus, according to Siemens ||, the difference in the con-
ducting power between hard-drawn and annealed silver varies
between 12*6 and 8 per cent., and that of copper between 6 and
— 0*5 per cent. ; according to Matthiessen and Von BoselT, that of
silver varies between 10 and 6 per cent., and that of copper
between 2*6 and 2 per cent.

♦ Phil. Tram. 1862, pt. 1.

t Aim. de Ckim. et de Phys. 1846, t. xvii. p. 242.

J Phil Trans. 1862, pt. 1. § Phil. Mag. Jan. 1861. |I Ibid.

IF Matthiessen and Vogt's anpublished researohes.

FOR ELECTRICAL MEASUREMENTS 2&

Again, the annealed wires of pure metals are so soft that they
would easily get damaged in covering them with silk or winding
them on the bobbins, so that in using them the utmost care would
have to be employed in order to prevent their getting injured.

4. On the influence of temperature on the electric conducting
power. — It ha? been shown that the conducting power of most
pure metals decreases, between 0** and 100*, 293 per cent. : pure
iron has been found to form an exception to this law, its conducting
power decreasing between those temperatures 38'2 per cent. If
pure metals be tiierefore used as standards, very accurate thermo-
meters are necessary, as an error of 0*1° in comparing two standards
would cause an error in the resistance of about 0*04 per cent. Now
there is great difficulty in obtaining normal thermometers ; and we
must bear in mind that supposing the zero-point of the thermo*^
meter is correct to-day, we are not at all justified in assuming that
it will be so in six months time ; so that we ought to redetermine
the zero-point of the thermometer before using it for the above
purpose. Again, it has been proved that the influence of tempera-
ture on the conducting power of wires of the same metal is not
always the same*. Thus, for the conducting power of annealed
copper wires the following values were found : —

No. 1 No. 8

100-0

20'

92-8

92-4

40*

86-3

86-6

«r

80-4

79-6

80"

75-1

74-4

100* 70-5 70-0

showing therefore that if standards of pure metals be used, the
influence of temperature on the conducting power of each would
have to be ascertained. It must also be borne in mind that it is
not at all easy to maintain a standard, even in a bath of oil or
water at a given temperature, for any length of time.

IL Those reproduced by a given length and section or weight, at
a given temperature, of a pure metal in a liquid state.

The only metal which has been proposed to be used in a liquid
state for the reproduction of units of resistance is mercury. We
shall only have to speak of its preparation in a state of purity, and
on the influence of temperature on its conducting power. For

* Phil, Trans. 1862, pt. 1.

30 PRACTICAL STANDARDS

a tube, carefully filled with mercury, will certainly form a homo-
geneous column, and its molecular condition will always be the
same at ordinary temperatures.

On its preparation in a pure state. — ^Although this metal is one
of the most easily purified, yet the use of it as a standard is open
to the same objections, although in a less degree, as have been
advanced against the use of pure metals in a solid state when
speaking of their preparation. We there stated that metals pre-
pared by different chemists conducted differently. Now although
the same manipulator may obtain concordant results in purifying
metals from different sources, yet that by no means proves that the
results of different observers purifying the same metal would show
the same concordance. Thus we find that the values obtained by
one experimenter* for the resistance of mercury, determined in six
different tubes, varied 1*6 per cent. This difference, he says, is
not greater than was to be expected. The resistances found were
as follows : —

Tubes I II m IV V . VI

Experiment... 1016-52 427-28 55538 21773 ld4-70 1142*3
Calculated ... 102554 427*28 55587 21601 19356 1148-9

Again, the values found for the conducting power of different
preparations of pure hard-drawn gold, by the same observerf , were
found equal to

78-0 at O'* 78-2 at 0" 76-8 at 0'

79-5 „ 78-3 „ 76-7 „

77-0 „ 78-0 „ 77-3 „

These values agree together as well as might be expected, con-
sidering that O'Ol per cent, impurity would cause these differences.
Now the values obtained by different observers vary between the
numbers 59 and 78.

If we now take the case of copper, the values found by the
same experimentersj for different preparations of the pure hard-
drawn metal were : —

99-9 at 0°
101-0 „
99-8 „
99-9 „

* Phil. Mag. Jan. 1861. The same experimenter (Dr SiemeuB) states, however,
in a later paper {Pogg. Ann. oxin. p. 96), that he is able to reproduce standards of
resistance by means of mercury with an accuracy equal to 0-05 per cent., but does
not indicate what other precautions he takes (see remarks on the above, PhiL Mag.
Sept. 1861).

t Phil. Trant. 1862, p. 12. t Ihid, p. 9.

99-4 at 0**

99-8 at 0**

99-4 „

100-3 „

99-9 „

100-0 „

FOB ELECTRICAL MEASUBEMENTS

* They were drawn by themselves, and all, with one exception,
electrotype copper.

It is well known how differently the so-called pure copper
condacts when prepared by different experimenters. In the
following Table, in order to show these &cts more clearly, we have
given the conducting powers of the metals, taking that of silver
equal 100 at 0^ Silver, copper, gold, and platinum were hard-
diawn. All values given, except where the contnuy is mentioned,
have been reduced to 0"^.

Siemens

Lenz

Becquerel

Matthiessen

Silver*

100
96-9

14-2
1-72

100
73-4

58-5

22-6
13-0
10-7
10-4
3-42 at 18-9

100

95-3

66-9

26-3

25-7

15-0

13-1

8-8

1-86

100
99-9
78-0
23-7
29-0
12-3
14-4 at 20-4

8-3
10-5 at 20-7
1-65

Copper

Gold

Cadmium

Zinc

Tin

Iron

Lead

Platinum

Mercury

If, now, mercury be taken as unit, we find the following values: —

Siemens

Lenz

Beoquerel

Matthiessen

Silver

58-20
56-40

8-25
100

29-24
21-46
1710

6 59
3-80
312
3-04
100 at 18-7

53-76

5123

37-04

1414

13-82

8-10

7-04

4-73

4-62

1-00

60-60
60-55
47-27
14-42
17-70
7-45
8-72 at 20-4

5-03
6-36 at 20-7
1-00

CoDDcr

o^?fr...:::::...

Cadmium

ZiDc

Tin

Iron

J/md

Platinum

Mercury

A glance at the foregoing Tables will sufiice to show how badly
Lenz's series agrees with the rest when mercuiy is taken as unit ;
and, in £Bu:t, we obtain more concordant results if, in the above
aeries, we take any other metal as unit. These facts therefore

* This and the following Table have been copied from a paper pablished in the
PkU. Mag. for Sept. 1861.

32 PRACTICAL STANDARDS

seem to indicate that mercury is not yet proved to be a safe ineiftLs
of reproducing standards of electric resistance.

The influence of temperature on the conducting power of
mercury, between 0"" and 100°, is, comparatively speaking, small,
being only 8*3 per cent., whereas that of the metals in a solid state
decreases between those limits 29*3 per cent. This property
would, of course, render the use of very accurate thermometers
unnecessary; for I'' would only cause a difference in the con-
ducting power of about 0*08 per cent., and therefore 0**! only
0*008 per cent., so that an error of 1 or 2 tenths of a degree might
almost be overlooked.

A fact has just come to our knowledge through Mr Jenkin.
He informs us that, having to make a report on the electrical
apparatus in the International Exhibition, he tested, amongst
other things, several resistance-coils. Now he found two sets of
coils made by the same firm, the one exhibited in the Prussian,
the other in the English department. Both were said to be
multiples of the mercury unit proposed by Siemens*, and their
resistances determined by comparing a coil in each set with that
of a tube filled with mercuiy. Taking each set by itself and com-
paring the coils in it with one another in the proper combination,
they were found to be perfect ; in fact, the adjustment of them
was perfectly accurate. When, however, Mr Jenkin compared
coils of the two sets with each other, instead of being equal, they
were found to show a difference of 1*2 per centf

III. TJiase reproduced by a given length and section or weight,

at a given temperature, of an alloy*

The alloy on which we have to speak is that composed of two
parts by weight of gold and one of silver. The reason why this
alloy was proposed is that the use of (say) 1 per cent, more or less
gold does not materially alter its conducting power.

1. On its preparation in a state of purity. — It has been shown
that the alloy may be made of commercially pure metals and have
the same conducting power as that made firom chemically pure
ones; for the maximum differences in the conducting power between

♦ Phil Mag. Feb. 1861.

t This discrepancy may perhaps be aitribatod to some inaccuracy in the repro*
daotion of the mercury standard.

FOR ELECTRICAL MEASUREMENTS 33

thoee made in different parts of the world are not greater than
those of a pure metal, either in a solid or liquid state, prepared by
the same experimenter. But it may be urged that part of the
differences obtained by different observers are due to the different
methods employed in determining their conducting powers, and
therefore had the conducting power of these alloys been determined
by different persons, much greater differences would have been
found. In answer to this, we give, in the following Table, the
determination of the conducting power of several alloys by Thomson
and Matthiessen*, independently of one another. The alloys were
made by Messrs Johnson and Matthey.

AUoy

Thomson

Matthiessen

100-0

100-05

95-8

95-0

102-9

102-7

100-8

99-1

98-1

97-7

89-9

92-7

80-6

80-06

Pore copper

Thornton

MatthieBsen

107O

107-2

107-5

105-9

108-7

106-9

107-7

1081

The differences here, with the exception of alloy 6 and copper
2, may he due to the temperature at which the observations were
made not being in both cases the same ; for 2 or 3 degrees' differ-
ence will account for them. The Table, however, shows that
different observers do obtain the same values for the conducting
power of the wires.

The values obtained for the conducting power of the gold-silver
alloy, made by different persons, of different gold and silver, are
given in the following Table : —

Alloy

Hard-drawn

Annealed

100-3

100-6

100-2

100-7

98-8

99-2

100-2

100-4

100-7

99-7

99-8

100-3

100-8

1001

100-4

* Procudingi of the Bayal Society, Feb. 1861.
B. A.

34 PAACTICAL STANDARDS

which shows, therefore, that the alloy may be prepared in a com-
mercial way, and still have a conducting power which varies less
than that of a pure metal prepared at different times by the same
experimenter. If we look at the hard-drawn series, we find five
out of the seven wires tested agree together exceedingly well, the
greatest difference being only 0'3 per cent. These five alloys were
made, three in London, by scientific chemists, one in Frankfort-
on-the-Maine, and ooe in Brussels. Those which agree least with
the others were made in New York (No. 3) and by a well-known
assayer in London (No. 6).

2. On its homogeneity, and its molecular condition, — If the
wires of the alloy made and drawn by different persons were not
homogeneous, the values obtained for the conducting power could
not have agreed so well together. It has been already mentioned
that some of the alloys determined by Thomson, when redrawn,
were found to have a different conducting power*.

Gondaotizig power of wire as GoDdnoting power after
Alloy received from the wire-drawer being redrawn

1 100-0 100-0

2 100-7 95-8

3 103-9 102-9

4 94-6 100-8
6 96-0 981

6 92-0 89-9

7 74-7 86-0
Pure copper 100-0 98*6

Of course, here again, some of these differences are due to the
temperature in each case not being the same ; but the differences
found with the alloys 2, 4, and 6 were undoubtedly due to &ulty
wires. It was for this reason that care was taken to have the
alloy drawn by different persons, in order to see if this would
influence the results obtained with them, as well as to ascertain
whether the wires would show the same faults as silver and copper
does when not carefully drawn. It has been argued that the
molecular condition of all alloys is liable to undergo a change by
age, and that, therefore, alloys are not fit to be used as standards.
Thus it is well known that brass and German-silver become brittle
and crystalline by age, and that the same may occur with the gold-
silver alloy ; but on looking at the composition of the alloy, it will
be found to be nearly the same as that of the gold chains of
commerce. Now, we do not know of a single instance where such

* Proeeedingi of the Royal Society, Feb. 1861.

FOR ELECTRICAL MEASUREMENTS 35

a chain, even after years of use, becomes brittle or crystallii^e ; so
that we think it more than possible that the alloy will not change
its molecular condition by age. It must also be remembered that
even when German-silver becomes brittle, it does not materially
alter in its conducting power. The same has already been
proved, and mentioned in this Report, to be the case with iron
and cadmium.

3. On the effect of annealing on the conducting power of the
aUoy. — When the alloy is heated to redness and cooled slowly, its
conducting power was found to have increased only 0*3 per cent. —
this value being the mean of eight wires annealed in different
ways, — ^proving, therefore, that if the wires may be only partially
hard-drawn, it will make but little difference in the conducting
power.

4. On the influence of temperature on the conducting power of
the allay. — When wires of this alloy are heated from 0'' to 100°, a
decrement in the conducting power, amounting to 6'5 per cent.,
will be found. The same arguments may, therefore, be put forward
in &vour of the use of the alloy as a standard, as in the case of
mercury when speaking of this property.

To sum up, therefore, the arguments in favour of and against
the use of the three propositions made to reproduce standards of
electric resistance, we find in feivour of a pure metal in a solid
state: —

That it appears that all descriptions of electrotype copper, when
carefully drawn, have the same conducting power.

Against it : —

1. That their preparation, with the exception of the electro-
type copper in a state of purity, is exceedingly difficult ; so that
independent persons preparing the same metal find, on comparing
the conducting powers obtained for them, that they vary several
per cent.

2. That the influence of annealing on their conducting powers
18 so great that differences may occur simply because the wires are
partially haid-drawn«

3. That the influence of temperature on their conducting
power is very great ; so that slight errors in thermometers, or in
the reading of them off, would materially affect the result.

In favour of using mercury as a means of reproducing stsmdards
the following may be said : —

3—2

36 PRACTICAL STANDARDS

1. That no molecular change can take place in the metal, nor
can any alteration occur in its conducting power, on account of
annealing; for its temper is always the sama

2. That the influence of temperature has only a small effect
upon its conducting power.

And against it : —

1. That there is a difficulty in obtaining absolutely pure
mercury ; so that the results obtained by different observers show
great variations.

2. That the standard tube cannot be kept full of mercury for
any length of time, owing to the diffusion of impure metal, arising
fix>m the amalgamated terminals, into the narrow tube ; so that
each time the standard has to be used, it must piactically be
remade.

3. If the tube be broken during the process of cleaning or
otherwise, it is not yet certain with what exactitude the standard
could be reproduced.

4. It is doubtful whether the resistance of a tube filled with
mercuiy to-day will have the same resistance if filled a year hence ;
for we have no proof if the dimensions of the tube will not alter
by being kept. It is well known that the bulbs of thermometers
are liable to change, and are continually changing, in capacity.

In fiivour of the gold-silver alloy it may be said : —

1. That this material, when prepared and drawn by different
persons, was found not to vary in its conducting power more than
1*6 per cent. ; whereas the variations found with the metals in a
solid state, prepared and drawn by different persons, amount to
several per cent., and those found for mercury by different observers
amount also in all cases to several per cent.

2. That the homogeneity and molecular condition of this
alloy are alwaj^ the same.

3. That the effect of annealing on the conducting power is
very small, being only 0*3 per cent. ; so that if a wire be partially
hard-drawn, its conducting power will not suffer to any appreciable
extent.

4. That the influence of temperature on its conducting power
between 0** and 100°, viz. a reduction of 6*5 per cent., is smaller
than either that of the metals in a solid state, viz. 29*3 per cent.»
or that of mercury, viz. 8*3 per cent.

And against it : —

FOB ELECTRICAL MEASUBEMENTS 37

That the conducting power may alter hj age, as the phjrsical
properties of alloys are more likely to change than those of metals.

From the foregoing statements, based on facts at present
known, it would appear that the best method of reproducing
standards, for those who are unable to procure copies of the
British- Association unit of electrical resistance, is that they should
make, or have made, a certain amount of the gold-silver alloy (as
described in the PhiL Mag. Feb. 1861), by two or three different
persons, in order to ensure a correct result, and take a given
length and section or weight of it, at a given temperature, which
has been foimd equal in resistance to the British-Association unit.
We would recommend, in order further to test what we have
stated in the foregoing Report, that three or more scientific men
and electricians be requested to compare the resistances of pure
mercury (obtained by them fix>m the best sources they are able)
and of the gold-silver alloy (made in the manner described in the
PhU. Mag^ with a Qerman-silver standard supplied to them by
your C!ommittee. If this be done, results would be obtained which
would put an end to many disputes on the subject, as well as
decide which of the above means is practically the best for repro-
ducing standards of electrical resistance where no copies of the
British- Association unit can be obtained.

Appendix D. — ^Professor Kirchhoff's Letter.
To Fteeming Jenkin, Esq.

Heidelberg, June 8, 1862.

Dear Sir, — I have the honour to acknowledge the receipt of
your letter of the Slst of May, in which you inform me of the
labours of the Committee appointed by the British Association, to
try and bring about the general introduction of one unit of electrical
resistance. I gladly respond to the invitation to express my view
on the manner in which the desired object might be best attained.

To define the unit of resistance by the resistance of a wire of
given dimensions of a pure metal appears to me impossible, for the
leasons which have been urged by the Committee ; hence, of the
three proposals discussed by the Committee, there only remain
two for our consideration.

38 PRACTICAL STANDARDS

(1) To adopt the unit proposed by Weber ; or (2) To establish,
as unit of resistance, the resistance of a column of pure mercury of
given dimensions and at a given temperature.

I do not think that to these a third of equal value can be
added ; for to define the unit of resistance by the thermal action
of an electrical current would certainly never answer the purpose,
because this thermal action cannot be measured with the necessary
accuracy, and the resistance of any wire which is to be permanently
kept cannot be fixed as unit ; for the resistance of any wire for a
given temperature certainly undergoes changes if electrical currents
are transmitted through it, and it is exposed to fluctuations of
temperature.

Of the above two units, the first recommends itself by coming
up more satisfactorily to the demands of science ; the second, as I
think, by being capable for the present of being practically carried
out with greater accuracy. But is it really necessary to decide
for one and against the other of these two units ? I think not. If
the ratio between them is established with the accuracy which is
now attainable, there can, I think, arise no more confusion firom
their simultaneous use, than from the practice of expressing lengths
sometimes in metres and sometimes in millimetres. You say, " It
is proposed that the unit adopted shall be represented by one
particular standard, constructed of veiy permanent materials, laid
up in a national repository " ; and further, " The Committee will
probably endeavour to devise some plan by which copies of the
actual material standard adopted may.be easily procured at a
reasonable cost." This plan, the execution of which I consider
highly desirable, might evidently be realized in all its essential
points without its being necessary to give the preference to one of
these units over the other : it would only be necessary to measure
the resist€utice of the normal standard in both units, and to add to
ea<^h copy its resistance expressed in both units.

In choosing the metal or the alloy of which the normal standard
and the copies are to be made, care must undoubtedly y{r«^ be
taken that the resistance is as unalterable as possible for one
temperature. It is undoubtedly desirable that the resistance
shall not vary rapidly with the temperature. This is, however,
not very important, provided that the temperature of the wire can
be accurately observed at any moment. To satisfy this condition,
the wires must not be coiled upon cylinders, but fastened so that.

FOR ELECTRICAL MEASUREMENTS 39

for the greater part of their extent, they lie clear, and hence
rapidly assume the temperature of the surrounding air or of the
non-conducting liquid in which they may have been immersed.

You request me to point out to you any researches of mine,
which refer to a unit of electrical resistance. I have to mention a
short treatise only, which appeared in vol. LXXVI. of Poggendorff's.
Annalen, under the title "Determination of the Constants on
which the Intensity of Induced Electrical Currents depends," and
which formed the answer to an academical prize-question which
Professor Neumann, in ESnigsberg, had proposed in the year 1846.
In this treatise a unit of electrical resistance has not been suggested ;
but in it the resistance of a wire has been measured by the unit
(or rather by double the unit), which was afterwards proposed by
Weber in his EUctrodynamic Measurements^ Professor Weber has
subsequently had the kindness to compare the copper wire whose
resistance I measured with those whose resistances he himself had
determined (Pogg, Ann. vol. Lxxxii. p. 360) ; he thereby found the
resistance of my wire about one-seventh greater than I had found
it The reason of this want of agreement consists partly in the
imperfection of the instruments which I had used, and partly in
the feet that in my experiments the temperature was little above
0** R., while in Weber's experiments it was about 20'' R.

Allow me, my dear Sir, to record the very great respect with
which I have the honour to be.

Yours very truly,

G. KiRCHHOFF.

Appendix E. — Dr Sibmens's Letter. ^Suggestions for the adoption
of a Common Unit in measurement of Electrical Resistance.

To the Committee appointed by the British Association to report on

Standards of Electrical Resistance.

Qentlemen, — I beg to acknowledge, with thanks, the honour
you have done me, in requesting me to furnish you with sugges-
tions in furtherance of your endeavours to procure the adoption of
a common unit of electrical resistance.

I proposed in Poggendorff*s Annalen (vol, ex. p. 1) to supply
this want by the adoption of the conducting power of mercury as
unit, and of the resistance which a prism of that metal a metre

40 PRACTICAL STANDARDS

long and a square millimetre section, at 0"" C, opposes to the passage
of a current, as unit of resistance.

The method by which I constructed standards in this unit was
as follows : —

From the ordinary glass tubes of commerce, pieces were
selected whose calibre was found to vary most regularly. After the
selected tubes had been ground to the length of a metre, they were
carefully cleaned and filled with pure mercury — the temperature
being measured. The contents were then weighed, and the values
reduced to 0° C. for expansion of glass and metal The resistances
of the tubes were calculated by the formula

1 + Va + ^
T«- Ira yg

9 3

which represents the resistance to a current in the longer axis of a
prismatic conductor either in the above unit or in 0*001 unit,
according as 2 is expressed in metres and g in grammes, or 2 in
millimetres and g in milligrammes respectively, o-s 13*557, the
specific gravity of mercury, at 0° C.

l + Va + -7=

is the coefBcient for conicalness, which in good tubes equals 1 very
nearly, a is the ratio of the greatest to the least transverse
section of the tube.

All the data therefore necessary for the value of W are exact
measures of length and weight. Measurements of the same tube,
at different times, gave results corresponding within O'Ol per cent,
with each other.

The first objection which is raised against the adoption of
mercury as unit, " that the tubes cannot be made of uniform or
similar wires, and that the standard once broken is lost for ever,"
is clearly untenable, since the tubes are not required to be uniform,
and the breakage of the standard involves only the necessity of a
new tube, and the determinations of length and weight anew, to
put the operator in possession of a new standard, whose agreement
with the broken one will depend solely on his own handiness in
manipulating. Everv standard, of whatever material, is liable to

FOR ELECTRICAL MEASUREMENTS 41

injuiy ; but the breakage of a glass is infinitely to be preferred to
the treacherous results of a bruised wire.

Mercury is, of all metals, that which is best suited to supply
a reproducible standard.

In the first place, it is procurable pure in sufficient quantities.
I heated for some hours samples of commercial mercury under
sulphuric acid containing a few drops of nitric acid, and found
their conducting powers afterwards to be precisely the same as
that of a quantity of chemically pure mercury reduced from the
oxide.

Secondly, mercury has always the same molecular structure,
and has therefore, at the same temperature, always the same
resistance.

From these two grounds it is possible to couple with this unit
a geometrical conception which is indispensable in practice.

Thirdly, of all metals capable of being used for resistances,
mercury has the lowest conducting power ; and of all pure metals
capable of the same application, its resistance varies least with
Tariations of temperature.

Having formed such original standards, it only remained to
copy them in a convenient form for employment in practice. This
I have done, —

1. In mercury contained in glass spirals, and

2. In German-silver wire.

The resistance-bridge which I made use of in these measure-
ments, with a reflecting galvanometer in its circuit, enabled me to
attain a precision of within O'Ol per cent.

The mercury spirals, as may be seen by the accompanying
drawing*, are provided with cups at their ends, for convenience
of filling and for receiving the contacts of the measuring apparatus.
They are either of known resistances, approximating only to a
multijde of the unit, or may be adjusted to an exact multiple by
boring out one of the ends of the tube, which, in this case, must
stand up half an inch inside the cup. The resistances of the
bridge must then be arranged so that no current passes through
the instrument only when the desired resistfiutice in the fourth side
is reached. When the spiral is filled, a vulcanized india-rubber
ring is put round the cups, and the spiral is suspended in a vessel

* The drawings have been omitted, the descriptions being intelligible without
them.

42 PRACTICAL STANDARDS

of ice- water or water kept in circulation by passing a current of air
through it, and the temperature measured by a delicate thermometer.

The electrical value of each spiral which I have made has been
determined by comparing it with at least two of the straight
normal tubes, both being kept during the measurement in ice-
water. The greatest differences which I have found between such
determinations do not exceed 0*05 per cent., to which limit the
copies may be trusted.

In answer to the objection that an admixture takes place
between the mercury and the solid metal used for the terminals, I
must remark that I have found this occasion really less incon-
venience than is generally believed. I kept the copper connexions
immersed in the mercury a whole week, but could not perceive
the slightest decrease in its resistance. Platinum electrodes of
considerable surface might be employed ; but I believe that the
removal of the copper connexion after each test, and the removal
of the old mercury from their surfaces before using them again,
are a sufficient safeguard against error arising from this source.
Besides, it is easy to fill the spiral with fresh mercuiy whenever it
is suspected to have dissolved any quantity of copper, or even on
every occasion when a measurement with it is to be made. Nor
does mercuiy change its resistance in the least by standing in the
air. This I have proved by keeping a spiral six months filled
without changing the mercury, and found its resistance to be
constant.

The material which I have extensively employed in copying
this measure, viz. German-silver, may be classed under the same
head as the expensive gold-silver alloy of Dr A. Matthiessen, over
which it has, however, the considerable advantages of a greater
specific resistance, and that its resistance varies less with tempera-
ture variations.

As a preventive against alteration of resistance by the influence
of the air, I have usually had the resistances made of this metal
covered with a coating of silk and lac.

Intermediate between the resistances to be measured and the
measure itself I have introduced resistance-scales. These contain
each a series of resistances (multiples of the unit), and are so
arranged that each resistance is exact when it stands stopped
alone in the circuit. When carefully made, these scales may be
Hp.pended on to 0*1 per cent.

FOR ELECTRICAL MEASUREMENTS 4S

Being convinced of the sufficiency of the method I have
described of reproducing a standard of electrical resistance, I have
the honour to suggest to you : —

1st. To recommend the universal adoption of the conducting
power of mercury as unit, and of the resistance which a prism of
that metal, a metre long and square millimetre section, at O"* C.>
opposes to a current of electricity as common unit of resistance.

2nd. To have the value of this measure ascertained, with the
greatest possible exactness, in absolute units.

3rd. To have copies of this unit constructed in mercury
contained in glass spirals for preservation in scientific repositories.

In the event of my suggestions being adopted, the mercury
unit should be determined again with the greatest possible care,
and with all the help which pure and applied science offers, and
copies of it made with equal exactness.

According to a late determination by Weber, the mercury unit
is only about 2^ per cent, greater than 10'^ absolute units, or one
mercary unit at — 26"* C. would equal 10,000,000,000 absolute units.

Since those cases in which the expression of resistances in
absolute measure is of advantage in £su;ilitating calculations occur
only very seldom, and only in purely scientific exercises, a single
determination of the relation of the two measures would be amply
sufficient. Should the absolute unit or any multiple of it be
adopted as common unit of resistance, there would still be wanted
a unit for expressing the conducting powers of bodies ; and mercury
is indisputably the best calculated for this purpose. And for
practical purposes, which in adopting a universal unit should be
principally taken into consideration, it is indispensable to define
the resistance-measure as a geometrical body of that material
which is selected as unit of conducting power. Every other
definition would not only burden uunecessarily the calculations
which occur in common life, but also confuse our conception of the
measure.

The reason why the arbitrary unit proposed by Jacobi (a length
of copper only approximately defined) found no admittance into
general use is to be sought in the £su;t that it failed to fulfil this
condition, and because the conducting power of all solid bodies is
too dependent on their molecular structure.

The same objection renders the adoption of the gold-silver alloy
proposed by Dr A. Matthiessen equally incapable.

44 PRACTICAL STANDARDS

Another disadvantage in the way of a solid metal unit is the
impossibility to solder thick connexions into the ends of a defined
length of any wire without altering its resistance.

Should the adoption of the mercury unit be deemed advisable,
I would place at the service of the British Association any further
information or assistance in my power.

I have the honour to be. Gentlemen,

Your most obedient Servant,

W. Siemens.

Appendix F. — Extrdctsfrom a Letter addressed to Professor

Williamson hy Dr Esselbach.

The two objections against the practical applications of Weber's
absolute unit have been sufficiently pointed out as being —

1. Its minuteness; and

2. That the electromotive force of galvanic elements does not
allow of variation (as strength of current, tension, and resistance
do), but that we have to accept certain constants as nature has
fixed them.

I take it for granted that the standard of absolute unit would

not lose in authority if a plain multiple of it were adopted. I need

not point out that the French metre itself \a only a submultiple,
1
A AAA AAA^^ ^^ ^ uatural unit — ^the ecuth's quadrant. The multiple

of the natural electro-magnetic unit I am about to suggest for

practical use is 10'*, therefore very simple (which is of no little

importance) ; and it is a multiple which leads us to those standards

which are practically used.

M. Bosscha gives the electromotive force of his Daniell's cells

in absolute measure as

1025-80 . 10«,

and calculates the one used by Mr Joule to be

10451 . 10«.

It will therefore be practicable to determine such concentration of
sulphuric acid as to make the electromotive force equal to

10 . 10»« ;

FOR ELECTRICAL MEASUREMENTS 45

and I believe the concentration required would be very near what
is actually used in telegraphy.

Besistanoe. — ^The different copies of Jacobi's ^talons are well
known to differ as much between each other as Daniell's cells ; and
if Siemens had done nothing else for galvanometry than to give us
copies which agree among themselves within a quarter per cent.»
the progress is obvious.

Weber's copy of Jacobi's etalon is

698. 1(K;
and that of M. Bosscha was

607 . lO'

in absolute measure.

Other statements (of Eirchhoff and others) give a much
smaller value.

In comparing Mr Siemens's mercury standard with three copies
of Jacobi's Etalon in his possession, I found two of them agreeing^
tolerably well with each other, and with a third one copied by my
friend Dr Teddersen, at Leipzig, from the original of M. Leyser>
which I took therefore to be the more correct ones. I found the
absolute value of Siemens's unit to be

or 1*1 Siemens's unit = 10".

We should therefore only have to multiply all observations

10"
expressed in Siemens's units by y^ to reduce them to absolute

measure, and the suggested multiple for the future standard would
not be far from I'l of Siemens's units, which every one admits to
be for metallic conductors a practical unit

For the resistance of insulating materials the figures become
impracticably high; but it would be a matter of professional
telegrajAy to adopt, in conformity with the system, the " resist-
ance " 10" and, besides, another " great resistance " containing 10**
•* resistances."

While the resistance of a mile of copper in an ordinary cable
would be (say) 4 R. (four resistances), the insulation-resistance of
a mile ot cable would be about 0*04 G. R. (great or gutta-percha
resistances).

46 PRACTICAL STANDARDS

My suggestion would therefore be : —

1. To adopt Weber's absolute unit, and to derive from it, by
the multiple 10^» (or 10,000,000,000), the practical unit.

2. To adopt 10*® of Weber's electro-magnetic units as the
** practical absolute unit " for electromotive force and resistance.

(10 of these units would be exactly 1 Daniell's cell.)

3. 1 of these units would be I'l of Siemens's units.

4. To allow, besides, a " practical great unit," viz. 10" of the
** practical units," for resistances in order to express the insulation-
resistance of cables in convenient figures.

5. To allow also a " practical small unit " of jzrr absolute units

to express insulation-currents and charge-quantities of cables in
convenient figures.

6. To adopt, in order to avoid confusion, for such " practical
units " a terminology as proposed by Messrs Bright and Clark.

London, September 18, 1862.

Appendix Q. — Circular addressed to Foreign Men of Science.

Sir, — I am requested to inform you that a Committee was
appointed by the British Association, which met last year at
Manchester, to report on Electrical Standards of Resistance.

The Committee consists of the following gentlemen : —

Professor A. W. Williamson,

F.R.S. (University College,

London).
Professor Charles Wheatstone,

F.RS. (London).
Professor William Thomson,

F.RS. (Glasgow>

Professor W. H. Miller, F.RS.
(Cambridge).

A Matthiessen, Ph.D., F.RS.
(London).

Fleeming Jenkin, Esq. (Lon-
don).

The Committee met on December 6th, 1861, and on April 3rd,
1862. On the latter occasion the following Resolution was
passed: —

" Resolved, — That the following gentlemen be informed of the
appointment of the present Committee, and be requested to
furnish suggestions in furtherance of its object.

FOR ELECTRICAL MEASUREMENTS 47

Professor Edlund (Upsala). Professor Neumann (Konigs-

Professor Th. Fechner (Leipzig). berg).

Dr Hemy (Washington).
Professor Jacobi (St Petersburg).
Professor G. Kirchhoff (Heidel-
berg).
Professor C. Matteucci (Turin).

Professor J. C. PoggendorfF

(Berlin).
M. Pouillet (Paris).
Werner Siemens, Ph.D. (Berlin).
Professor W. Q. Weber

(Gottingen)."

I have, in consequence, the honour of addressing you the
present letter.

The Resolutions passed at the two meetings are enclosed, and
from them you will gather the general scope of the Committee's
inquiry. I add some further explanation as to the objects and
intentions of the Committee.

Great inconvenience has been felt from the absence of any
generally adopted unit for the measurement of electrical resistance,
and it was thought that the influence of the British Association
might be successfully exerted to procure the adoption of a common
standard. The present time was thought especially favourable,
since, although the methods of observation have been brought to
great perfection, no local units have as yet taken very deep root.

The units which up to the present time have been considered
by the Committee may be classed under three heads : —

Ist. A given length and weight or section of wire made of
some pure metal, and observed at a given temperature, as originally
proposed by Professors Wheatstone, Jacobi, and others.

2nd. Units based on Weber's and Gauss's system of absolute
measurement.

3rd. A given length and section of pure mercury at a given
temperature.

Whatever basis is adopted for the unit, it is proposed that the
unit adopted shall be represented by one particular standard,
constructed of very permanent materials, laid up in a national
repository ; and it has been proposed to use Dr A. Matthiessen's
gold-and-silver alloy for this purpose. The arguments which have
been used for and against these systems are as follows : —

In &vour of the use of a wire of some pure metal it is said —

That the plan is the simplest possible, and admits of indepen-
dent observers forming their own standard.

48 PRACTICAL STANDARDS

Against this plan it is said —

1st. That even when pure, two apparently similar wires do
not resist equally unless their temper or molecular condition be
the same — a condition which cannot practically be ensured.

2nd. That there is reason to believe that the resistance of a
given wire is not constant even at a constant temperature.

3rd. That the resistance of all pure metals varies veiy rapidly
with the temperature.

4th. That great difficulty is found in obtaining any metal
pure, and that the attempt of most persons to reproduce the unit
for their own use would be attended with incorrect results. This
is evidenced by the different relative results as to the resistance of
pure metals published by different observers.

In favour of Weber's units it is urged —

1st. That their use will ensure the adoption of a complete
system of corresponding standards for electrical currents, quantities,
and tension or difference of potential

2nd. That their use is essential in the djnoamic treatment of
any problem connected with electricity; for instance, in deter-
mining the heat generated, the force exerted, the work done,
and the chemical action required or produced under any given
circumstances.

3rd. That their use would be a simple extension of the system
already universally adopted in magnetic measurements.

4th. That the unit is independent of the physical properties
of any material.

Against the system it is urged that the unit cannot be deter-
mined with sufficient accuracy, and that even its approximate
reproduction, where copies cannot be obtained, is difficult and
expensive.

In favour of the mercury standard the following arguments
are used: —

1st. No change can occur in the molecular structure or temper
of the material, and therefore the same tube filled with pure
mercury will certainly always conduct alike.

2nd. Change of temperature causes only a slight difference in
resistance.

Against this plan it is said —

1st. That tubes cannot be made of uniform or similar wires,
and that, therefore, the standard once broken is lost for ever.

FOR ELECTRICAL MEASUREMENTS 49

2nd. That the standard tube cannot be kept full of pure
mercury, owing to the admixture which would take place of the
solid metal used for the terminals, so that each time the standard
has to be used it has practically to be remade.

3rd. That the attempt, by most observers, to reproduce the
unit for their own use would be attended with incorrect results, as
is shown by the different results obtained by different observers.

In &your of Dr Matthiessen's alloy, as compared with wires of
pure metal, or with mercury, as a material for the standard, it is
said —

1st. That the variations of resistance, corresponding with
variations of temperature or temper, are small.

2nd. That a unit expressed in this material can be more
readily and certainly reproduced than one expressed by a pure
metal, because the presence of slight impurities in the component
metals, or a slight change in their proportion, does not sensibly
affect the result.

Against this plan it is said that the physical properties of an
alloy are more likely to change than those of a pure metal.

Against all the plans for standards, based on an arbitrary length
and section of an arbitrary material, the supporters of the absolute
units state that the adoption of such an arbitraiy standard would
lead to great confusion and complication in the measurement of all
other electrical properties, and in the expression of the relation of
such measurements to those of force, work, heat, etc.

This objection does not, of course, apply to the expression of
the absolute unit by means of a wire of pure metal, of an alloy, or
by mercury : but it is urged that no observer should ever attempt
the reproduction of a standard when a copy of the proposed
universal standard can possibly be obtained ; and the Committee
will probably endeavour to devise some plan by which such copies
of the actual material standard adopted may be easily procured at
a reasonable cost.

It will be seen, from the resolutions passed, that the Committee
are now engaged in investigating the degree of accuracy with
which Weber 8 units can be obtained, and the degree of permanency
which may be expected from the use of the metal or alloy forming
the material standard expressing these or other units.

The Committee will feel greatly indebted to you if you will
afford them the benefit of your valuable advice and experience on

& A. ^

60 PRACTICAL STANDARDS

the above points, and on any others which may occur to you.
They also venture to hope that such a standard may be selected
as jwill give very general satisfaction ; and, if approved by you,
that you will kindly take an interest in procuring its general
adoption.

Personally being charged with the duty of preparing an his-
torical summary of the various units proposed, I shall be grateful
if you will favour me with any remarks as to your own labours in
this field, or if you could oblige me with references to any papers
or works in which the subject is treated.

I am. Sir,

Your obedient Servant,

Fleeming Jenkin.

Appendix H. — Description of the Electrical Apparatus arranged
by Mr Fleeming Jenkin /w the production of exact copies
of the Standard of Resistance,

This apparatus is a simple modification of that generally known
as " Wheatstone's bridge." It contains, however, some special
arrangements, in virtue of which various practical difficulties are
avoided, so that very great accuracy can be ensured with compara-
tive ease. The usual bridge-arrangement is shown in Plate I.
fig. 9, where the irregular scrolls. A, C, R, 8, represent the four
conductors of which the resistance is to be compared ; the thick
black lines show those portions of the circuit which join the coils
with the four comers, U^ F, Z, F, and are supposed to have no
sensible resistance in comparison with the coils ; finally, the thin
lines show connexions, the resistance of which in no way affects
the accuracy of the comparison between the four coils. By this
arrangement the four conductors. A, C, R, 8, are so connected with
the galvanometer, 0, and the battery, B, that no current passes
through the galvanometer when the conductors bear such a

A a

relation to one another that the equation 77=0 holds good ;

whereas a current in one or other direction passes so soon as 7^

FOB ELECTRICAL MEASUREMENTS 61

is greater or less than ^*. Thus the direction and strength of

the current observed serve as guides by which the resistance of
any one of the conductors may be gradually adjusted by shortening
or lengthening the Mrire, until on the completion of the circuit no
deflection whatever can be observed on. the galvanometer, however
delicate it may be, or however powerful the battery used. When
this has been done, we may be sure that the above relation exists
between the four conductors. In practice, it is seldom desirable to
use powerful batteries ; the test is made delicate by the use of an
extremely sensitive astatic galvanometer.

In speaking of the four conductors. A, C, R, S, which are
generally all coils of wire of similar construction, although each
fulfilling a distinct function, some difficulty often occurs in
explaining readily which coil or conductor is referred to. They
can of course be distinguished by letters; but this requires
reference to a diagram on every occasion, and the writer has
therefore been in the habit of distinguishing the four coils by
names drawn from a very obvious analogy existing between this
electrical arrangement and the common balance in which one
weight is compared with another. The equality between the two
weights on either side of a balance, when the index is at zero,
depends on the equality of the arms of the balance ; and if the
arms are unequal, the weights required to bring the index to zero
are proportional to the arms (inversely). Let A and C be called
the arms of the electrical balance, while 8 and JB are looked on as
analogous to the standard weight and mass to be weighed respec-
tively, and let the galvanometer-needle stand for the index of the
balance. Then all the above statements, with respect to the
weights and arms, hold good for the electrical arrangement (except
that the proportion between the electrical arms and weights is
direct instead of inverse). The writer therefore calls this arrange*
ment an electric balance — A and C the arms, 8 the standard, and
12 the resistance measuredf. In the adjustments of resistance-

* This stotement holds good ftlao if the battery and galyanometer wires, as
shown in the diagram, are interchanged.

t The name of parallelogram, sometimes given to the arrangement, is objeotion-
able, inasmneh as the relation obtaining between the fonr condactors is not that
wUdi exists between the four sides of any parallelogram, except in the one case of
equality between all foor conductors. The connexions are, howeyer, most easily

4—2

58 PBACTICAL STANDARDS

coils or copies of a standard, the object is to produce a second coil,
iJ, exactly equal to the first or standard, 8; and the arms, A, C,
must therefore be absolutely equal, before, by this arrangement, an
exact copy can be made. Hitherto it has often been the practice
to use for the arms, A, 0, two coils made as equal as possible, and
placed so close as to remain at sensibly equal temperatures; so
that the equality between 12 and S is dependent on the equality
between A and C, and cannot be determined with greater accuracy
than that between these coils. This limit to the accuracy is a
defect for our present purpose, and the writer has, moreover, found
it undesirable to depend on the permanent equality of two coils.
It is by no means certain that, without very extraordinary pre-
cautions, the two arms will remain unaltered in their original
equality. A slight molecular change, or a slight chemical action
on the surface of the wires, disturbs this equality permanently ;
and even if the coils are so constructed as to remain really equal
at equal temperatures, the accidental passage of a current through
one arm, and not through the other, for a very short time, will
disturb their accuracy very sensibly for a considerable time. There
are various devices by which the equality to be established between
R and 8 may be rendered independent of the absolute equality
between A and (7, and the writer has adopted a plan, now to be
explained with the aid of the diagrams (figs. 7, 8). This plan
allows the approximation to equality between R and £> to be almost
indefinitely increased.

It will be seen that fig. 7 does not differ from fig. 9, except by
the addition of a wire, WX, of sensible resistance, between the
two coils A and G. The point U is no longer fixed, but can be
moved along WX, The arms of the balance are therefore no>
longer A and C, but A-i-XU and C+WU. Thus the movable
point D affords the means of slightly altering or adjusting the
ratio of the two arms. A and C are made as equal as possible,,
independently of WX, which is a very short wire.

The test is made as follows : — When the standard and coil to be
measured have been put in their places as in fig. 7, the point U ia
moved along the wire WX until the galvanometer-index is not

foUbwed in a drawing when arranged as the four sides of a quadrilateral figure.
Professor Wheatstone's original name of Differential Resistance Measurer does not,
as it seems to the writer, sufficiently distinguish this arrangement trom other
differential methods.

FOB ELEOTRICAL MEASUREMENTS 53

deflected when the circnit is closed. The position of the point U
is noted by a scale. R and S are then reversed, so as to occupy
the position relatively U> A, C shown in fig. 8. The point U is
again moved until the galvanometer-needle remains undeflected on
the circuits being closed. The new position of {7 is again observed
by a scale. If the point U does not require to be moved at all, we
may be quite sure that R is exactly equal to 8, and that

^+xt7=c7+Trtr,

since it would be quite impossible that the ratio >^ — =p^ should be

equal to both ^ and -p, unless this ratio were equal to 1. More-
over, if WX be made of the same wire as the coils A and (7, and
if those coils are formed of about 100 inches of wire, and if the
observed positions of U differ by a given distance, x, this length,
c, measured in inches, will express very necu'ly the difference
between R and £f in a percentage of the whole length of R. Thus, if
X be one inch, the standards S and 22 differ by about one per cent.
If the point 27, when adjusted in each case, be found nearer R than
8, then R is the smaller of the two, and vice versa. The percentage
of error in i2, thus measured, is not of course strictly accurate,
inasmuch as the ratio between the two arms is not exactly |^ ;
but if WX be not more than three or four inches long, the percent-
age of error measured in this way is quite sufficiently accurate to
allow the new coil to be so exactly adjusted after very few trials,
that no greater movement of 27 than (say) ^th of an inch is
required to prevent any deflection of the galvanometer when R
and 8 are reversed. We may then be sure that no greater error
than (say) about 0*1 per cent, exists in the equality between the
new coil and the standard. Two firesh coils, ili, (7i, are then
taken, containing each about 1000 inches of wire similar to WX, or
an equivalent resistance. It will then be found that, to maintain the
index at zero when R and 8 are reversed, 27 must be moved about
ten times as much as before, or (say) one inch. 22 can then be
still further adjusted till 27 is not moved more than ^th of an
inch, when a new degree of approximation to equality, with an
error of not more than 0*01 per cent., will have been reached.
Then the coils A^ Ci are changed for a fresh pair, il„ (7„ with a
resistance equal to about 10,000 inches of the wire WX : one-tenth
of an inch on WX will then represent an error of only 0*001 per

54 PBACTIOAL STANDABDS

cent. By a repetition of this process, quite independently of any
absolute equality between the pairs A, C; Ai, 6*,; A^, C,; etc., a
gradual approximation to any required extent may be ensured. The
delicacy of the galvanometer used, and the nicety of the means
available for increasing or diminishing the resistance of R, form
the only limits to the approximation. A slight want of equality
between any pair of arms will simply bring the point U a little to
one side or the other of the centre of WX, as the final adjustment
with that pair is made, but will not affect the truth of the
comparison between R and 8, Each pair must, however, be so
nearly equal that the addition of part of the short wire, WX, to
one side will be suflScient to correct the other; otherwise the
adjustable point U would not bring the index to zero, even when
at one end of the wire.

This arrangement, besides rendering us independent of the
accuracy of any two arms, has some incidental advantages of con-
siderable practical importance. At each test it gives a measure of
the amount by which the new coil to be adjusted must be
lengthened or shortened. The test is at first comparatively rough,
or adapted to errors of one or two per cent., and only gradually
increases in delicacy as the desired equality is more and more
nearly approached. It is not necessary that the resistance of WX
should remain absolutely constant, since it is only used (numerically)
to give a rough approximation to the percentage of error. It is
desirable that the battery should remain in circuit as short a time as
possible ; the circuit is therefore broken between 1 and 2, figs. 7 and
8, by a key, K, with which contact should be only momentarily made,
when all the other connexions are complete. The direction of the
jerk of the galvanometer-needle to one side or the other need alone
be observed ; no permanent deflection is required with this arrange-
ment as a guide to the amount of error. This is a considerable
advantage, inasmuch as it avoids heating the wires, and saves
time. The induction of the coils on themselves might lead to
some false indications, unless special precaution were taken against
it, as pointed out by Professor W. Thomson*. To avoid this
source of error, the galvanometer circuit is broken between 3 and 4,
figs. 7 and 8, at Ki, and should only be closed after the battery
circuit has been completed at K and equilibrium established
throughout all the conductors.

* Vide Phil Mag. ADgast 1862.

FOR ELEGTBICAL MEASUREMENTS 55

Before passing to a detailed description of the apparatus as
actually constructed, some remarks are required as to the means
of making temporary connexions. All connexions which require to
be altered may be the means of introducing errors, inasmuch as
the points of contact are very apt to offer a sensible but uncertain
resistance. In measuring small resistances, the resistance at the
common binding-screws is found to create very considerable errors.
Binding-screws have therefore to be avoided at all points where an
uncertain resistance could cause error. Mercury-cups, made as
follows, have been found in practice very suitable for temporary
connexions, and have been adopted in the apparatus. The bottom
of each cup is a stout copper plate, with its surface well amal-
gamated, forming one of the two terminals to be joined. A stout
copper wire, | inch in diameter, with a flat end well amalgamated,
forms the other terminal. When the amalgamation is good, and
care is taken that the wire shall rest on the plate, this form of
connexion offers no sensible resistance. The amalgamated wire is
easily kept bright and clean by being dipped from time to time in
a solution of chloride of mercury and wiped. The copper plate
should also be removed from the cup, cleaned, and re-amalgamated
occasionally. All permanent connexions should be soldered.

The apparatus itself, as actually constructed, will now be
described (figs. 1 to 6), Plate 1. It consists of a wooden board*,
about 12 in. X 7 in., containing the mercury-cups, the adjusting
wire, yfX, the key, K, and the terminals to which the battery and
galvanometer are connected. The letters in the figures 1 to 6
correspond exactly to those used in the diagrams 7 and 8 ; and the
apparent complexity of the connexions can thus be easily dis«
entangled c, Ci ; a, ax are two pcdrs of mercury-cups, into which the
terminal wires on the bobbin, (7, A, dip. This bobbin contains
the two coils, C and A^ forming the arms of the balance, r, r^ and
«, 9x are mercury-cups, into which the terminals of the standard and
coil to be adjusted are placed. These mercury-cups are so con«
nected with the four cups, (2, duf^fi^ that when d is connected
with dj, and / with /i, by a couple of wires in a small square of
wood, D, then A,C,8, and R are connected as in fig. 7 ; but when
D is turned round, so as to connect d with /, and di with /i. A, C,

* Ezperieooe has shown that this board should be made whoUj of yaloanite,
and not, as shown in the drawing, partly of wood and partly of vnloanite. —
F. J., 1S72.

56 PRACTICiiL STANDAJtDS

M, and 8 are connected as in fig. 8. D is called the commutator.
The same end might be effected without a commutator by simply
interchanging R and 8; but it is frequently inconvenient to do
this. All these connexions cure made by short stout copper bars,
dotted in fig. 2. The wire WX, the sliding brass piece H, carrying
a spring for the contact at U (fig. 4), and the scale E, by which the
position of jET is observed, will be readily understood from the
drawing. The sliding piece, H, is connected with the proper
points by the helix of copper wire, h, and the screw, /. OQi and
BBi are common binding-screws, to which the wires from the
galvanometer and battery are attached. K is the key, by depressing
which, first, the battery is thrown into circuit, and then the
galvanometer. It consists of three brass springs, 1, 2, 3 (fig. 6),
each insulated one fix>m the other, and connected by three screws,
1, 2, 3 (fig. 2), with the necessary points of the arrangement. A
fourth terminal, 4 (figs. 2 and 6), is immediately under the free
end of the springs, and is armed with a small platinum knob or
contact-piece. The three springs are also all armed with platinum
contact-pieces, all in a line one above the other (fig. 6). When
the finger-piece, T, is pressed down, 1 and 2 are first joined, and
then 3 and 4 ; 3 is insulated from 2 by the vulcanite, Q. All the
connexions which are permanently made, and under the board, are
shown in fig. 2. Those which have no sensible resistance are stout
copper bars, and form the bottoms of the mercury-cups : those of
which the resistance is immaterial are made of wire, insulated by
gutta-percha, and are simply shown as dotted irregular lines in
fig. 2 ; they will be found, on comparison, to correspond with the
thin lines on fig. 7. It will also be found that those parts shown
by thick lines in the diagram are made by thick bars or rods and
mercury-cups.

Three sets of arms, CA, CiA^, (7,ila, are provided; the shortest
pair is first used, and U adjusted by the slide, H, till the galvano-
meter does not deflect when T is pressed down. The commutator,
D, is then turned round, and U adjusted afiresh. The coil, R, is
then altered according to the two positions of U, and this process
repeated, using the second and third pair of arms as required,
until the desired approximation between R and 8 has been
obtained. An astatic galvanometer, with a very long coil, will, for
most purposes, give the best results ; and one or two elements will
be found a sufficient battery. The construction of B and 8

Plate 1,

^mposUum.<iraj^(T^.JL) ci connects wUk. d.Ji.f„^auf

•^.tf I>ta^frtMnvofccnnjtcpio>uf with. €Xfmmueaior Ti
plactMl aano9» board,, dboonruuXfd wvOt, /•# d, wiOt f^

Fig. 9.

Comntotv Bridgt/

^ .o'-^

-»->'* ■>*» vNiifyw*"^ ">'» ^N"^ !%xi».N^A

•»«..

•^^r

T
O

1 \\

•♦t«»\-» , ^^\ *

V»V\ ". V»*\

••V .-

•rr

1 X

1 "•:

:*\

.'T

•/ 0

.^.-,

,/ »■

_»_ * — .

-v>c

• • I

\N *\

FOR ELECTRICAL MEASUREMENTS 57

recommended, and the precautions to ensure perfect equality of
temperature, will form part of next year's Report.

The apparatus, although specially designed for the production
of equal coils, is applicable to ordinary measurements of resistance
by comparison with a set of resistance-coils ; for this purpose the
terminals of the resistance-coils should be put in the place of the
standard 8, and any conductor of which the resistance is to be
measured in the place of R. If a comparison by equality is to be
made, the wire WX can be used as already described ; it is, how-
eyer, frequently desirable to make a comparison with one arm
tenfold or a hundredfold greater than the other, by which means
measurements of resistances can be made ten or a hundred times
greater or smaller than could be done if equality alone between R
and S were measured ; for this purpose the three pairs, AC, AiCi,
AiC^, are made exactly decimal multiples one of the other, and
then, by taking A and (7i, or A and Cq, etc., in the cups a, a, and
c, Ci, the required decimal ratio is obtained. The resistance of the
wire TTJT would, however, falsify this ratio, and it is eliminated by
a simple copper rod, which is placed for the purpose between the
two cupe, e, ei, and maintains the whole wire WX at sensibly one
potential. The commutator also is useless in measurements of
this kind, and should be left untouched in the position shown in

fig. 1.

The apparatus exhibited was manufactured for the Committee
by Messrs Elliott Brothers, of London, and gives excellent results.

SECOND REPORT— NEWCASTLE-ON-TYNE, 1863.

The Committee on Electrical Measurements, appointed in
1862, have not confined their attention to determining the best
unit of electrical resistance, the point to which the duties of the
Committee of 1861 were nominally restricted, but have viewed
this comparatively limited question as one part only of the much
larger subject of general electrical measurement. The Committee,
after mature consideration, are of opinion that the system of
so-called absolute electrical units, based on purely mechanical
measurements, is not only the best system yet proposed, but is the
only one consistent with our present knowledge both of the relations
existing between the various electrical phenomena and of the
connexion between these and the ftmdamental measurements of
time, space, and mass. The only hesitation felt by the Committee
was caused by doubts as to the degree of accuracy with which this
admirable system could be or had been reduced to practice.

The measurements of voltaic currents, electromotive force, and
quantity would offer little difficulty, provided only electrical resist-
ance could be measured in absolute units ; and for this purpose it
would be sufficient that the resistance of a single standard con-
ductor should be so determined, since copies of this standard could
be multiplied at will with any desired precision, and fix)m com-
parison with these copies the absolute resistance of any circuit
whatever could be obtained by methods requiring comparatively
little skill and well known to all electricians. The practical adop*
tion of the absolute system was felt therefore to depend on the
accuracy with which the absolute resistance of some one standard
conductor could be measured ; and while doubts existed on this
point, it was thought premature to make any extended experiments
on the application of the absolute system to voltaic currents,
electromotive force, or quantity. The Committee are happy to
report that these doubts have been dispelled by the success of the

PBACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 59

experiments made for the Committee by Professor J. Clerk Max-
well, Mr Balfour Stewart, and Mr Fleeming Jenkin, according to
the method devised by Professor W. Thomson. These experiments
have been actively prosecuted at King's College for the last five
months with continually increasing success, as, one by one,
successive mechanical and electrical improvements have been
introduced, and the various sources of error discovered and
eliminated.

The Sub-committee are confident that considerably greater
accuracy can yet be obtained by the further removal of slight
defects, the importance of which only became apparent when the
main difficulties had been overcome. In order, therefore, to secure
the best attainable result, and still fijrther to test the accuracy
and concordance of the experiments before taking any irrevocable
step, the Committee have decided not to issue standard coils at the
present Meeting ; but the results already obtained leave no room
for doubt that the absolute system may be adopted, and that the
final standard of resistance may be constructed without any serious
delay. Over-haste might eventually entail corrections as incon-
venient as those which would follow an arbitrary and unscientific
choice of units, and the very experiments made by the Sub-com-
mittee prove that the hesitation of many to adopt the absolute
units as hitherto determined was well founded. It is certain that
resistance-coils purporting to have been constructed fi*om previous
absolute determinations do not agree one with another within 7>
8, or even 12 per cent.

Before further alluding to the results obtained by the Sub-
committee, it is desirable that the experiments themselves should
be understood ; and to this end the Committee have thought fit
that a full explanation of the meaning of absolute measurement,
and of the principles by which absolute electrical units are deter-
mined, should form part of the present Report, especially as the
only information on the subject now extant is scattered in detached
papers by Weber, Thomson, Helmholtz, and others, requiring con-

•

aiderable labour to collect and understand. In order to make this
account as clear as possible, it has been thought best to disregard
entirely the chronological order of the discoveries and writings on
which the absolute system is founded; and this has rendered it
very difficult to refer to the original source of each statement or
conclosion. In the Appendix (C) this want is, it is hoped^
remedied

60 PRACTICAL STANDARDS

The word '' absolute " in the present sense is used as opposed
to the word ''relative/' and by no means implies that the measure-
ment is accurately made, or that the unit employed is of perfect
construction ; in other words, it does not mean that the measure-
ments or units are absolutely correct, but only that the measure-
ment, instead of being a simple comparison with an arbitrary
quantity of the same kind as that measured, is made by reference
to certain fundamental units of another kind treated as postulates.
An example will make this clearer. When the power exerted by
an engine is expressed as equal to the power of so many horses,
the measurement is not what is called absolute; it is simply
the comparison of one power with another arbitrarily selected,
without reference to units of space, mass, or time, although these
ideas are necessarily involved in any idea of work. Nor would
this measurement be at all more absolute if some particular horse
could be found who was always in exactly the same condition and
could do exactly the same quantity of work in an hour at all
times. The foot-pound, on the other hand, is one derived unit of
work, and the power of an engine when expressed in foot-pounds
is measured in a kind of absolute measurement, x.e. not by
reference to another source of power, such as a horse or a man,
but by reference to the units of weight and length simply — units
which have been long in general use, and may be treated as
fundamental. In this illustration, chosen for its simplicity, the
unit of force is assumed as fundamental, and as equal to that
exerted by gravitation on the unit mass ; but this force is itself
arbitrarily chosen, and is inconstant, depending on the latitude of
the place of the experiment.

In true absolute measurement the unit of force is defined as
the force capable of producing the unit velocity in the unit of
mass when it has acted on it for the unit of time. Hence this
force acting through the unit of space performs the absolute unit
of work. In these two definitions, time, mass, and space are
alone involved; and the units in which these are measured,
i.e, the second, gramme, and metre, will alone, in what follows,
be considered as fundamental units. Still simpler examples of
absolute and non-absolute measurements may be taken from the
standards of capacity. The gallon is an arbitrary or non-absolute
unit. The cubic foot and the litre or cubic decimetre are absolute
units. In fine, the word absolute is intended to convey the idea
that the natural connexion between one kind of magnitude and

FOR ELECTRICAL MEASUREMENTS 61

another has been attended to, and that all the units form part of
a coherent ^stem. It appears probable that the name of "derived
units" would more readily convey the required idea than the word
"absolute," or the name of mechanical units might have been
adopted ; but when a word has once been generally accepted, it is
undesirable to introduce a new word to express the same idea.
The object or use of the absolute system of units may be expressed
by saying that it avoids useless coefficients in passing from one
kind of measurement to another. Thus, in calculating the
contents of a tank, if the dimensions are in feet, the cubic contents
are given in cubic feet, without the introduction of any coefficient
or divisor ; but to obtain the contents in gallons, the divisor 6'25
is required. If the power of an engine is to be deduced from the
pressure on the piston and its speed, it is given in foot-pounds
or metre-kilogrammes per second by a simple multiplication ; to
obtain it in horse-power, the coefficients -^jj or. ^ must be used.
No doubt all the natural relations between the various magnitudes
to be measured may be expressed and made use of, however
arbitrary and incoherent the units may be. Nevertheless the
introduction of the numerous factors then required in every calcu-
lation is a very serious annoyance; and moreover, where the
relations between various kinds of measurement are not imme-
diately apparent, the use of the coherent or absolute system will
lead much more rapidly to a general knowledge of these relations
than the mere publication of formulae.

The absolute system is, however, not only the best practical
system, but it is the only rational system. Every one will readily
perceive the absurdity of attempting to teach geometry with a
unit of capacity so defined that the contents of a cube should be
6| times the arithmetical cube of one side, or with a unit of
sarbce of such dimensions that the surface of a rectangle would
be equal to 0'000023 times the product of its sides ; but geometry
00 taught would not be one whit more absurd than the science of
electricity would become unless the absolute system of units were
adopted.

In determining the unit of electrical resistance and the other
electrical units, we must simply follow the natural relation existing
between the various electrical quantities, and between these and
the fundamental units of time, mass, and space. The electric^
phenomena susceptible of measurement are four in number —

62 PRACTICAL STANDARDS

current, electromotive force, resistance, and quantity. The
definitions of these need not now be given, but will be found
in the Appendix C (§§ 14, 15, 16, and 17). Their relations one to
another are extremely simple,and maybe expressed by two equations.
First, by Ohm's law, experimentally determined, we have the
equation

<^=i <i)

where C = current, E = electromotive force, and R = resistance.
From this formula it follows that the unit electromotive force
must produce the unit current in a circuit of unit resistance ; for
if units were chosen bearing any other relation to each other,

C would be equal to a? -^ , where x would be a useless and absurd

factor, complicating all calculation, and confusing the very simple
conception of the relation established by Ohm's law.

Secondly, it has been experimentally proved by Dr Faraday
that the statical quantity of electricity conveyed by any given
current is simply proportional to the strength of the current,
whether electro- magnetically or electro-chemically measured, and
to the time during which it flows; hence, in mathematical
language, we have the equation

e = <», (2)

where t = time, and Q =» quantity. From this equation it follows
that the unit of quantity must be the quantity conveyed by the
unit current in the unit of time; otherwise we should have
Q sz yCt, where y would be a second useless and absurd coefficient.
From equations (1) and (2) it follows that only two of the elec-
trical units could be arbitrarily chosen, even if the natural relation
between electrical and mechanical measurements were disregarded.
Thus if the electromotive force of a Daniell's cell were taken as the
unit of electromotive force, and the resistance of a metre of mercury
of one millimetre section at 0° were taken as the unit of resistance,
it would follow fix)m equations (1) and (2) that the unit of current
must be that which would be produced by the Daniell's cell in a
circuit of the above resistance, and the unit of quantity would be
the quantity conveyed by that current in a second of time. Such
a system would be coherent ; and if all mechanical, chemical, and
thermal effects produced by electricity could be neglected, such a

FOB ELECTRICAL MEASUREMENTS 63

system might perhaps be called absolute. But all our knowledge
of electricity is derived fix)m the mechanical, chemical, and thermal
effects which it produces, and these effects cannot be ignored in a
true absolute system. Chemical and thermal effects are, however,
now all measured by reference to the mechanical unit of work ;
and therefore, in forming a coherent electrical system, the chemical
and thermal effects may be neglected, and it is only necessary to
attend to the connexion between electrical magnitudes and the
mechanical units. What, then, are the mechanical effects observed
in connexion with electricity? First, it has been proved that
whenever a current flows through any circuit it performs work, or
produces heat or chemical action equivalent to work. This work
or its equivalent was experimentally proved by Dr Joule to be
directly proportional to the square of the current, to the time
during which it acts, and to the resistance of the circuit ; and it
depends on these magnitudes only. In mathematical language
this is expressed by the equation

W-^C^Rt (3)

where W = the work equivalent to all the effects produced in the
circuit, and the other letters retain their previous signification.
This is the third fundamental equation affecting the four electrical
quantities, and represents the most important connexion between
them and the mechanical units. From equation (3) it follows
(unless another absurd coefficient be introduced) that the unit
current flowing for a unit of time through a circuit of unit re-
sistance will perform a unit of work or its equivalent. If every
relation existing between electrical and mechanical measurements
were expressed by the three fundamental equations now given,
they would still leave the series of units undefined, and one unit
might be arbitrarily chosen from which the three other units
would be deduced by the three equations; but these three
equations by no means exhaust the natural relations between
mechanical and electrical measurements. For instance, it is ob-
served that two equal and similar quantities of electricity collected
in two points repel one another with a force {F) directly propor-
tional to the quantity Q, and inversely to the square of the distance
(d) between the pointa This gives the equation

^'l' w

64 PRACTICAL STANDARDS

from which it would follow that the unit quantity should be that
which at a unit distance repels a similar and equal quantity with
unit force. The four equations now given are sufficient to measure
all electrical phenomena by reference to time, mass, and space
only, or, in other words, to determine the four electrical units by
reference to mechanical units. Equation (4) at once determines
the unit of quantity, which, by equation (2), determines the unit
current; the unit of resistance is then determined by equation
(3), and the unit electromotive force by equation (1). Here, then,
is one absolute or coherent system, starting from an effect pro-
duced by electricity when at rest. The units based on these four
equations are precisely those called by Weber electrostatical units^
although it may be observed that he chose those units without
reference to what is here called the third fundamental equation,
or, in other words, without reference to the idea of work, intro-
duced into the system by Thomson and Helmholtz*.

The four equations are sufficient to determine the four units,
and into this system no new relation can be introduced. The first
three equations may, however, be retained, and a distinct absolute
system established by substituting some other relation between
electrical and mechanical magnitudes than is expressed in equa-
tion (4); and, indeed, the electrostatic system just defined is not
that which will be found most generally useful It is based on a
statical phenomenon, whereas at present the chief applications of
electricity are dynamic, depending on electricity in motion, or on
voltaic currents with their accompanying electro-magnetic effects.
Now the force exerted on the pole of a magnet by a current in
its neighbourhood is a purely mechanical phenomenon. This
force (/) is proportional to the magnetic strength (m) of the pole
of the magnet, and to the strength of the current C; and if the
conductor be at all points equidistant from the pole, or, in other
words, be bent in a circle of the radius k round the pole, the force
is proportional to the length of the conductor (L) ; it is also in-
versely proportional to the square of the distance (k) of the pole
from the conductor, and is affected by no other circumstances than
those named. Hence we have

/=^ (5)

* Vide Appendix C, § 81.

FOB ELECTRICAL MEASUREMENTS 65

From this equation it follows that the unit length of the unit
current must produce the unit force on a unit pole at the unit
distance. If the equations (1), (2), (3). and (5) are adopted as
fundamental, they give a distinct absolute system of units, called
by Weber the electro<magnetic units. Equations (4) and (5) are
incompatible one with another, if equation (2) be considered
fundamental ; but the electro-magnetic units have a constant and
natural relation to the electrostatic units. It will be seen that in
the fundamental equation (5) of the electro-magnetic system,
besides the measurement of time, space, and mass, alone entering
into the other equations, a fourth measurement (tn) of a magnetic
pole is required ; but this measurement is in itself made in terms
of the mechanical units, for the unit pole is simply that which
repels another equal pole at unit distance with unit force. Thus
in the electro-magnetic as in the electrostatic system all measure-
ments are ultimately referred to the fundamental units of time^
space, and mass. The electro-magnetic units are found much the
more convenient when dealing, as we have now chiefly occasion to
do, with electro-magnetic phenomena.

The relations of the electro-magnetic units one to another and
to the mechanical units may be summed up as follows : — The unit
current conveys a unit quantity of electricity through the circuit
in a unit of time. The unit current in a conductor of unit re-
sistance produces an effect equivalent to the unit of work in the
unit of time. The unit current will be produced in a circuit of
unit resistance by the unit electromotive force. The unit current
flowing through a conductor of unit length will exert the unit
force on a unit pole at unit distance. (In the electrostatic
system all the above propositions hold good except the last, for
which the following must be substituted: — the unit quantity of
electricity will repel a similar quantity at the unit distance with
unit force.)

It remains to be explained how electrical measurements can be
practically made in electro-magnetic units. Of all the magnitudes,
currents are the most easily measured, provided the horizontal
force (H) of the earth's magnetism be known. Let a length (L)
of wire be wound so as to form a circular coil of small section as
compared with its radius (£).

Let a short magnet be hung in the centre of the coil placed in
the magnetic meridian, as in the ordinary tangent galvanometer,

B. A« 6

66 PRACTICAL STANDARDS

and let the deflection produced by the current C be called d, then
it is easily* proved from the fundamental equation (5) that

G^^tAud (6)

Thus, where the value of H is known, a tangent galvanometer
only is required to determine the magnitude of a current in
electro-magnetic absolute measure, although neither the resist-
ance of the circuit nor the electromotive force producing the
current may be known. The measurement of quantity can be
obtained from that of a current by a make-and-break apparatus,
or ** Wippe," in a well-known manner, or by measuring the swing
of a galvanometer-needle when a single instantaneous discharge is
allowed to pass through it (Appendix C, § 25). If, therefore, we
could measure resistance in absolute measure, the whole system of
practical absolute measurement would be complete, since, when
the current and resistance are known, equation (1) (Ohm's law)
directly gives the electromotive force producing the current. The
object of the experiments of the Sub-committee (made at King's
College, by the kind permission of the Principal) was therefore to
determine, in the absolute system, the resistance of a certain piece
of wire, in order from this one careful determination to construct
the material representative of the absolute unit with which all
other resistances would be compared by well-know^n methods.

There are several means by which the absolute resistance of
a wire can be measured. Starting from equation (3), Professor
Thomson, in 1851, determined the absolute resistance of a wire
by means of Dr Joule's experimental measurement of the heat

* The resaltant electro-magnetic force (/) exerted at the centre of the coil by

CL
a current (C) will, by equation (5), be /= -r,- , and the short magnet hong in the

centre will experience a couple acting in a direction perpendicular to the plane of

the coil equal to .^ , where m/=the product of the strength of one of the poles

into the length of the magnet, or, in other words, its magnetic moment. The
strength of the couple acting perpendicularly to the axis of the magnet, when it

has deflected to an angle d under the influence of the current, will be cos d ;

at the same time the equal and opposite couple exerted on the magnet by the earth^s
magnetism wiU be sin d Hml ; hence

o = — F~ X a = — T~ tan a.

L cos d L

FOB ELECTRICAL MEASUBEMENTS 67

developed in the wire by a current*; and by this method he
obtained a result which agrees within about 5 per cent, with our
latest experiments. This method is the simplest of all, so far as
the mental conception is concerned, and is probably susceptible of
very considerable accuracy.

Indirect methods depending on the electromotive force induced
in a wire moving across a magnetic field have, however, now been
more accurately applied ; but, before describing these methods, it
will be necessary to point out the connexion between the electro-
motive force induced in the above manner and the fundamental
equations adopted for the absolute system. The exact sense in
which the terms are employed is defined in the accompanying
footnote, along with some simple corollaries from those defi-
nitions f-

* Ptdl. Mag. Tol. n. ser. 4, 1851, p. 551.

t DefiniHcn 1. — A magnetic field is anj space in the neighbonrhood of a magnet.

DejinUicn 2. — The unit magnetic pole is that which, at a unit distance from a
similar pole, ia repeUed with unit force.

Defiimtum 3. — The intensity of a magnetic field at anj point is eqoal to the force
whieh the unit pole would experience at that point.

Corollary 1. — A pole of given strength {S) will produce a magnetic field which
(if oninfluenoed by other magnetic forces) will at the unit distance from the pole be
of the intensity S, t.«. numerically equal to the strength of the pole ; for, at that
distance, the force exerted on a unit pole would, by def. 2, be equal to S, and hence,
by def. 3, the intensity of the magnetic field at that point would be equal to S.

Definition 4. — The direction of the force in the field is the direction in which
any pole is urged by the magnetism of the field ; this is the direction which a short,
balanced, freely suspended magnet would assume.

Hrmark. — The properties of a magnetic field, as shown by Dr Faraday, may be
eonveniently and accurately conceived as represented by lines of force (each line
rapraaenting a force of constant intensity). The direction of the lines will indicate
the direction of the force at all points; and the number of lines which pass through
the unit area of cross section will represent the magnetic intensity of the field
resolved perpendicularly to that area.

Definition 5. — A uniform magnetic field is one in which the intensity is equal
throughout, and hence, as demonstrated by Professor W. Thomson, the lines of
force parallel.

Example. — The earth is a great magnet. The instrument-room, where experi-
ments are tried, is a magnetic field. The dipping-needle is an instrument by which
the dixcetion of the lines of force is found. The intensity of the field ia found by a
method described in the Admiralty Afamiat, Srd edit., article ** Terrestrial Magnetism."
The number of lines of force paasing through the unit of area perpendicularly to the
dipping-needle in the room must be conceived as proportional to this intensity, and
the direction to correspond with that of the dipping-needle. The magnitude and
diraetion of the earth's force at a point are generally expressed by resolving it into
two oomponenta, one horizontal and the other vertical The mean horizontal

6—2

68 PRACTICAL STANDARDS

A current (0) in a straight conductor of length (i) crossing
the lines of force of a magnetic field of the intensity (8) at right
angles will experience the same force (/) as if all the points of
the conductor were at the unit distance fix)m a pole of the strength
(8). The force in this case exerted on the magnet is, by equation
(5), equal to 8LC, and, conversely, an equal force is exerted by the
magnet on the current. Hence we have equation (7), expressing
the value of the force (/) exerted on a current crossing a magnetic
field at right angles,

f^8LG. (7)

Let us imagine this straight conductor to have its two ends
resting on two conducting-rails of large section in connexion with
the earth, and let the whole sensible resistance (R) of the circuit
thus formed be constant for all positions of the conductor. Let
us further imagine the rails so placed that when the conductor
slips along them it moves perpendicularly to the magnetic lines of
force and to its own length. By experiment we know that when
the conductor is moved along the rails cutting these lines of force,
a current will be developed in the circuit, and that the action of
the magnetic force on this current will cause a resistance (/) to
the motion (due to electro-magnetic causes only) ; and, by equa-
tion (7), we find that this resistance /= SLC.

Let the motion be uniform, and its velocity be called V; and
let the work done in the unit of time in overcoming the resistance
to motion due to electro-magnetic causes be called W; then
W = VSLC. But this force produces no other effect than the
current, and the work done by the current must therefore be IT,
or equivalent to that done in moving the conductor against the
force /; but, by equation (3), W = C'R, and hence

R-^ (8)

It has already been shown that C and 8 can be obtained in
absolute measure ; hence the second member of equation (8) con-
tains no unknown quantities, and, by the experiment described>

component in England for 1862 was at Kew =3*8154 British units, or 1*7592
metrical ; i,e, a nnit pole weighing one gramme, and free to move in a horiiontal
plane, would, under the action of the earth's horizontal force, acquire, at the end
of a second, a velooitj equal to 1*7592 metres per second. {Vide also Appendix G»
S§ 5 to 12.) If the centimetre is taken as the fundamental unit of length, -17592
will be the mean value of the horizontal force.

FOR ELECTRICAL UEASUREMENTS 69

the absolute resistanoe (iZ) of a wire might be determined. One
curious consequence of these considerations is, that the resistance
of a conductor in absolute measure b really expressed by a
velocity ; for, by equation (8), when 8L = C we have 12 = F, that
is to say, the resistance of a conductor may be expressed or de-
fined as equal to the velocity with which it must move, if placed
in the conditions described, in order to generate a current equal
to the product of the length of the conductor into the intensity
of the magnetic field ; or more simply, the resistance of a circuit
is the velocity with which a conductor of unit length must move
acroes a magnetic field of unit intensity in order to generate a
unit current in the circuit. Moreover it can be shown that this
velocity is independent of the magnitude of the fundamental units
on which the expression of the magnetic intensity of the field or
strength of the current is based, and hence that electrical re-
sistance really is measured by an absolute velocity in nature, quite
independently of the units of time and space in which it is expressed

V8L
(Appendix C, § 39). By equation (8) we have (7=s p , but by

equation (1) C^ = "» > hence

E^VSL; (9)

that is to say, the electromotive force produced between two ends
of a straight conductor moved perpendicularly to its own length
and to the lines of force of a magnetic field is equal to the product
of the intensity of the field into the length of the conductor and
the velocity of the motion ; or, more simply, the unit length of a
conductor moving with unit velocity perpendicularly across the
lines of force of a magnetic field will produce a unit electromotive
f<»ce (or difference of potential) between its two ends. This was
by Weber made a fundamental equation, in place of equation (3),
first shown by Thomson and Helmholtz to be consistent with
Weber's electro-magnetic equation. These simple and beautifiil
relations between inductive effects and the simple voltaic effects
first described are well adapted to show the rational and coherent
character of the absolute system.

The experiment last described, as a method of finding the
absolute resistance of a conductor by measuring the velocity of
motion of a straight wire, would be barely practicable ; but it will
be eadly understood that we can, by calculation, pass from this

70 PRACTICAL STANDARDS

simple case to the more complex case of a circular coil of known
dimensions revolving with known velocity about an axis in a
magnetic field of known intensity. Weber, from these elements,
determined the absolute resistance of many wires; but this
method requires that the intensity of the magnetic field be known;
and the determination of this element is laborious, while its value,
for the earth at least, is very inconstant. A method due to
Professor Thomson, by which a knowledge of this element is
rendered unnecessary, has therefore been adopted in the experi-
ments of the Sub-committee at King's College. In this plan a
small magnet, screened from the effect of the air, ia hung at the
centre of a revolving coil, which is divided into two parts to allow
the suspending fibre to pass freely.

By calculation it can be shown that when the coil revolves
round a vertical axis, the couple exerted on a magnetic needle of
the moment ml, when deflected to the angle d, will be

-jj^m/cosd

The equal and opposite couple caused by the earth's magnetism
will be Hnd sin d. Hence

tan (2 =

4A:*i2

an equation fix)m which the earth's mlBignetic force and the moment
of the suspended magnet have heeh eliminated, and by which the
absolute resistance (R) can be calculated in terms of the length
(i), the velocity (F), the radius (fc), and the deflection {d\ The
resistance thus calculated is expressed in electro-magnetic absolute
units, because equation (10) is a simple consequence of equations
(1), (3), and (5) — fundamental equations in the electro-magnetic
system. The essence of Professor Thomson's method consists in
substituting, by aid of the laws of electro-magnetic induction, the
measurements of a velocity and a deflection for the more complex
and therefore less accurate measurements of work and force
required in the simple fundamental equations. But, however
simple in theory the method may be, the practical determination
of the absolute resistance of a conductor by its means required
great care and very numerous precautions, — some of an obvious

FOR ELECTRICAL MEASUREMENTS 71

character, while the need of others only became apparent during
the course of the experiments.

The apparatus consisted of two circular coils of copper wire,
about one foot in diameter, placed side by side, and connected in
series ; these coils revolved on a vertical axis, and were driven by
a belt from a hand-winch, fitted with Huyghens's gear to produce
a sensibly constant driving-power. A small magnet, with a mirror
attached, was hung in the centre of the two coils, and the deflec-
tions of this magnet were read by a telescope from the reflection
of a scale in the mirror. A frictional governor controlled the
speed of the revolving coil. The details and a drawing of the
apparatus are given in Appendix D and Plate 2; but a short
account may fitly be given here of the points of chief practical
importance, the difficulties encountered, and the improvements
stiU desirable.

It is essential that the dimensions of the coil be very accurately
known, that the axis on which it revolves should be truly vertical,
and that, except in the coil itself, no currents affecting the position
of the magnet be induced in any part of the apparatus. To
measure the angular deflection the distance of the scale fix)m the
mirror is required, and the scale must be truly perpendicular to
the line joining its middle point with the suspension-fibre. All
these conditions were fulfilled without difficulty ; but the scale by
the reflection of which the deflections were measured was, towards
the end of the experiments, found not to be very accurately
divided ; and although a correction for this inaccuracy has been
applied in the calculations, an improvement can in future experi*
ments be effected by the use of a more perfect scale. The magnet
was suspended by a single silk fibre, eight feet long, inside a
wooden case, and by suitable adjustments was brought very care*
fiilly to the centre of the coils. The whole suspended system was
so screened fr*om currents of air, and so well protected from
vibration, that when the coil revolved at its full speed of 350
revolutions per minute, the reflection in the mirror was as clear
and undisturbed as when the coil was at rest. The torsion of the
long fibre was determined by experiment, and the slight necessaiy
corrections applied in the calculations. The Huyghens's gearing
for the driving hand-winch was somewhat roughly constructed,
and could certainly be improved; nevertheless there was little
difficulty in maintaining a sensibly constant driving-power for

72 • PRACTICAL STANDARDS

twenty minutes at a time. The speed of the coil was controlled
by a frictional governor of novel form, designed by Mr Jenkin for
another purpose, and lent for the experiments in question. The
action of this governor, combined with that of the driving-gear,
was such that in many experiments the oscillations in deflection
due to a change of speed were not so great as those due to the
passage of steamers in the river when all parts of the apparatus
were at rest ; so that the deflections during twenty minutes could
be quite as accurately observed as the slightly imperfect zero-point
from which they were measured. Still better results are expected
with a larger governor, made specially for the apparatus, on the
joint plans of Professor Thomson and Mr Jenkin The oscillations
produced by the passage of steamers on the Thames at no great
distance from the place of experiments were of veiy sensible
magnitude; and although by carefully observing the limit of every
oscillation during every experiment the error due to this cause
was in great part eliminated, it is desirable that any future
experiments should be conducted in some spot free from all local
magnetic disturbance.

The speed of the coil was determined by observing on a chro-
nometer the instant at which a small gong was struck by a detent
released once in every hundred revolutions. Mr Balfour Stewart's
skill in this kind of observation enabled him thus to determine
the velocity with great accuracy, especially as the observations
frequently lasted for twenty minutes without material alteration
in the speed.

During the operation of coiling the wire, the circumference of
the core and of each successive layer was carefrilly measured by
means of a steel riband applied first to the coil, and then to a
standard scale, allowance being made for the half thickness of the
steel. From this the mean radius and depth of the coil and the
effective length of the wire were determined. It was considered
advisable, however, in order to check any error in counting the
number of windings of the coil, to measure the length of the wire
when uncoiled. This was effected without stretching the wire, in
a manner amusing from its simplicity. At the conclusion of the
experiments, the wire to be measured was uncoiled in the Museum
at King's College and lay in awkward bends on the planked floor.
The straight planks formed an obvious contrast to the crooked
wire, and a joint between the planks was found where the opening

FOR ELECTRICAL HEASUREMEKTS 73

was just sufficient to hold the wire when pushed into this little
groove. Held in this way, the wire when measured was quite
straight, and yet was never stretched.

No other measurements than those already described are re-
quired by the simple theory; but this theory, as hitherto stated,
stands in need of various slight corrections. The currents induced
l>y the earth's magnetism are modified by the currents induced by
the little suspended magnet, and also by the induction of the coil
on itself. The force deflecting the magnet is also modified by the
lateral distance of the coils from the vertical axis. An elaborate
analysis of the corrections required on these grounds was made by
Professor Maxwell (Appendix D), and to allow of these corrections,
the moment of the suspended magnet was measured, and the
position of every turn of the copper coil carefully observed. An
experimental determination of the induction of the coil on itself,
by a method due to Professor Maxwell, agreed with the calculated
correction within one-quarter per cent.

The resistance of the copper coil measured by these laborious
experiments varied each day, and during each day, according to
the temperature; and, moreover, this temperature could at no
time be determined with sufficient accuracy. It was therefore
intended that at each experiment a small German-silver coil, at a
known temperature, should have been prepared exactly equal in
resistance to the copper coil during that experiment, and these
omall coils were to have been kept as permanent records of the
resistance of the copper coil on each occasion ; but this resistance
was found to vary so rapidly that the little copies could not be
accurately adjusted with sufficient rapidity, and the resistance of
the copper coil was therefore simply measured at the beginning
and end of each experiment, in terms of an arbitrary unit. This
proportional measurement was made with rapidity and precision
by a new method, which, it is believed, is superior to the usual
plan depending on the division or calibration of a comparatively
short wire in the Wheatstone balance. (Appendix D, Part II.)

One unforeseen difficulty was caused by the change of direction
of the earth's magnetic force during each experiment. Our method
is indeed independent of the intensity of the earth's magnetism,
but depends essentially on its direction, since it depends on the
value of a deflection from the magnetic meridian. When this
aouroe of error was discovered by the continual and gradual change

74 PRACTICAL STANDARDS

of zero observed, the absolute time of each experiment was noted,
and a continuous correction obtained from the contemporaneous
records at Kew, which agree closely with the total changes observed
at the beginning and end of each experiment. As the change of
zero frequently reached three or four divisions in the course of the
day, and as the whole deflection seldom exceeded 300 divisions,
the importance of this correction is apparent.

The presence of stationary masses of iron does not aflfect the
experiments injuriously, so long as the uniformity of the magnetic
field in which the coil revolves is undisturbed — a point carefully
tested before the experiments began; but a change in the position
of iron in the neighbourhood during any experiment produces a
corresponding error in the result, and the serious eflfect of moving
very small masses of iron at a great distance from the coil was
only fully appreciated in the later experiments.

When it is considered that the method described is the simplest
known, the discrepancy between the few determinations hitherto
made in absolute measurement will cause no surprise. The time,
labour, and money required could hardly be expected to be given
by any one person, and in researches of this kind the value of the
co-operation secured by the committees of the Association is
especially evident.

The absolute unit of the Sub-committee is about eight per
cent, larger than the unit as derived from a German-silver coil
lately measured by Professor Weber. It is about six and a half
per cent, larger than the unit as derived from a value published
by Professor Weber of Dr Siemens's mercury units. It is about
five per cent, smaller than the unit as derived from coils issued
by Professor Thomson in 1858, based on Jacobi's standard and a
previous determination by Professor Weber. It is about five per
cent, smaller than Thomson's determination from Joule's silver
wire. It agrees most closely with an old determination of a copper
standard made by Weber for Professor Thomson, which it exceeds
by only a very small fraction.

The experiments of the Sub-committee agree much better than
the above, one with another. Owing to the gradual improvement
in the method and apparatus, the experiments of the last three
days are alone considered satisfactory. On the first day the
maximum deviation in six distinct experiments from their mean
result was 2*4 per cent. On the second day the maximum

FOR ELECTRICAL MEASUREMENTS 75

deviation in four experiments from their mean was 1'3 per cent.
On the third day the maximum deviation in five experiments
from their mean was 1'15 per cent. The maximum deviation in
the means of the three days' experiments from the mean of the
whole is only four-tenths per cent.

These results are not unsatisfactory, and are perhaps more
accurate than any measurement yet made of the relative values of
heat and work — a measurement corresponding to a great extent
in its nature with that undertaken by the Committee. Never-
theless, considering the discrepancy of the various independent
results, the Committee are of opinion that it is essential that
the results of the Sub-committee should be checked by a fresh
series of experiments with a new coil in a distinct place, when
every separate measurement will necessarily be repeated. The
Sub-committee especially urge the repetition of the experiments,
as with the improvements already enumerated, and other minor
alterations, they confidently expect a considerably closer approxi-
mation to the absolute unit than they have hitherto obtained.
It will be well here to remark that, according to the resolution of
the Committee of 1861, the coils, when issued, will not be called
absolute units, but the units of the British Association ; so that
any subsequent improvement in experimental absolute measure-
ment will not entail a change in the standard, but only a trifling
correction in those calculations which involve the correlation of
the physical forces.

It is now time to leave the question of absolute measurement
and pass to some of the other points under the consideration of
the Committee. Dr Matthiessen has, by careful experiment,
proved the permanence for a year at least of the electrical resist-
ance of certain wires; but he has detected a change in others, due,
apparently, to the influence of time. Certain specimens of silver,
gold, and copper have varied; but other specimens of the same
metals have remained constant. All the specimens of platinum
and gold-silver alloy have remained constant, and all the specimens
of German-silver have changed considerably. It is proposed to
continue and extend these experiments, and it is much to be
hoped that the defect observed in the German-silver tested will
not be found common to all the varieties of this alloy, in other
respects so well adapted for the construction of resistance-coils.
Dr Matthiessen found no difference in the resistance of wires of

76 PRACTICAL STANDARDS

any of the above metalB before and after the passage of a powerful
current transmitted through them continually for a fortnight.
The details of these experiments are given in Appendix A.
Dr Matthiessen has also continued his experiments with the
object of finding an alloy with a minimum variation of resistance
due to change of temperature, but has been unable to produce
a wire superior in this respect to the silver-platinum alloy men-
tioned in Appendix A of the Report of last year, as decreasing in
conducting power 31 per cent, between 0° and 100° Centigrade.
German-silver was found to decrease under the same circumstances
4*4 per cent.

The valuable experiments by Mr Sabine, for Dr Werner
Siemens of Berlin, on the reproduction of standards by means of
mercury, although not undertaken for the Committee, yet bear
80 directly on the subject before them that the results cannot be
allowed to pass unmentioned. Dr Siemens has conclusively proved
that he can, in his laboratory, reproduce a standard by means of
mercuiy with an error of less than 0*05 per cent. This admirable
result, while it seriously affects the question of the best material
for the construction and reproduction of the standard, leaves, of
course, the question of the best magnitude for the standard quite
untouched. Dr Matthiessen thinks that several of the solid metals
are equally fitted for the purposes of reproduction, and, if aided
by the Association, is disposed to put his conviction to experi-
mental proof. It is especially desirable that the various methods
proposed should be tested by the concordance of the results obtained
from a number of independent observers.

With reference to the construction of the material standard, it
is proposed that the British- Association unit shall be represented
by several equal standards made of the different metals, which, so
iar as our limited experience goes, show the greatest signs of
constancy. Two at least of those standards would be made of
mercury, in the manner proposed by Dr Siemens. The permanent
agreement between several of these standards would afford the
strongest possible proof of their constancy.

Passing to other electrical measurements, the Committee have
to report that Professor W. Thomson has successfiilly constructed
a material standard gauge by which electromotive force or differ-
ence of potentials can be directly measured. This instrument is
founded on a measurement of the electrical attraction exerted on

FOR ELECTRICAL MEASUREMENTS 77

a small movable portion of a large conducting-plaoe by another
laige parallel plane fixed at a constant distance, and electrified to
a different potential. The force exerted is ultimately measured
by the torsion of a platinum wire; but the difference of potential
corresponding to any one gauge is simply indicated by the motion
of an index to a sighted position. If the planes are brought
sufficiently close, with a given torsion in the platinum wire, the
movable piece will be in a condition of unstable equilibrium when
its index is in the sighted position, but if moved to a greater
distance the equilibrium will be stable ; hence, by a correct choice
of the distance between the two planes, or initial torsion in the
platinum wire, as compared with the difference of potential to be
measured, any required delicacy of indication is obtained. The
constancy of the gauge, like that of all standards, depends simply
on the constancy of the materials of which it is constructed, and
there is no reason to apprehend any special difficulty in the
present case.

Professor Thomson has also on the same principle constructed
an electrometer in which the distance between the parallel plcmes
is made variable, and is adjusted by a micrometer-screw. The
plane conductor, of which the small movable index forms part, is
in this instrument permaiiently maintained at a high potential
by connexion with the inner coating of a Leyden jar, and the
other plane is connected with the body to be tested. Calculation^
oonfirmed by experiment, shows that in these instruments the
difference of potentials between any two bodies, successively tested^
is directly proportional to the difference of the distances between
the parallel planes required in each case to bring the index to its
sighted position. This difference of distance is the same whatever
be the charge of the Leyden jar, provided only it remains constant
during the comparison of the two bodies. With this limitation,
the indications of the instrument may be called independent of
the charge of the Leyden jar. There can be little doubt that
gauges of electromotive force and electrometers, fulfilling the above
conditions, will shortly become as necessary to all practical elec-
tricians as standards of resistance and sets of resistance-coils.

No progress has been made in the measurement of currents^
and mnch remains to be done in this respect. The method already
described^ depending on the use of a tangent galvanometer, requires
a knowledge of the horizontal force of the earth's magnetism^

78 PRACTICAL STANDARDS

and is, therefore, in most cases beyond the reach of observers where
greater accuracy is required than can be obtained by taking their
value from the scientific almanacs. Next year it is hoped that
this want may be remedied; and the present Report may fitly
conclude by the enumeration of objects to be pursued by the
Committee, if reappointed at the present Meeting: —

1st. The experiments on the determination of the absolute
unit of resistance will be continued.

2nd. Immediately on the conclusion of these experiments,
equal standards, constructed of such metals as promise the greatest
constancy, will be deposited at Kew, where the permanence of
their equality will be rigorously tested.

3rd. Unit resistance-coils of the best known construction will
be issued to the public.

4th. The experiments already begun on the permcmence of
the electrical resistance of wires and alloys under various circum-
stances will be continued and extended.

5th. The experiments on the reproduction of standards by
chemical means will be continued.

6th. Experiments on the best construction of gauges of electro-
motive force or diiFerence of potential, and on electrometers, will
be continued.

7th. A standard galvanometer, for the measurement of
currents in absolute measure, will be constructed, and electro-
dynamometers for the same purpose compared with the standard
instrument, and issued to the public.

8th. Experiments on the ratio between the electrostatic units
and the electro-magnetic units will be undertaken.

9th. Experiments will be made on the development of heat in
conductors of known absolute resistance with currents of known
absolute magnitude. The results of these experiments will give, by
equation (3), a new and very accurate determination of the mechani-
cal value of the unit of heat.

The conclusion of the experiments on absolute resistance, and
the adoption of the absolute system as the basis of all electrical
measurement, will, it is hoped, allow considerable progress to be
made in most of these researches.

FOR ELECTRICAL MEASUREMENTS 79

Appendix A. — On tlie Electrical Permanency of Metals
and Alloys. By A, Matthiessen, F.R.S.

The following are the results obtained with the metals and
alloys described in Appendix B of the First Report on Standards
of Electrical Resistance by your Committee : —

The wires to be experimented on were : —

1. Silver: hard-drawn ) a^ ^ -^ x.

2. saver; annealed } ^* **•" *•"* '^'^ P"^ ' P'^'*-

3. Silver: hard-drawn 1 Cut from the same piece, but different

4. Silver: annealed j from 1 and 2; pure.

5. Copper: hard-drawn •••

6. Copper: hard-drawn 1 /^ i. -_ xi.

e. Copper: amiealed / ^""^ ^^°^ *^" ^°^" P*^' P^

7. Copper: hard-drawn \ Cut from the same piece, but different

8. Copper: annealed J from 5 and 6; pure.

9* Gold: hard-drawn 1 ^ ^ - .,

10. Gold : annealed } ^* '"" **"» *'""' P'^' P'^

11. Gk>ld: hard-drawn \ Cut from the same piece, but different

12. Gold: annealed j from 9 and 10; pure.

13. Platinum: hard-drawn ...1 ^ . - ^, . ,
, - ^, ^. 1 jj y Cut from the same piece : commercial.

14. Flatmum: hard-drawn ...J ^

15. Gbld-silver alloy : hard-drawnl Cut from the same piece. Made by Messrs

16. Gold-silver alloy : hard-drawnj Johnson and Matthey.

' Cut from the same piece. No. 19 ar-
ranged with longer connectors, and
used as normal wire with which the
rest were compared.

17. German-silver: annealed

18. German-silver: annealed

19. German-silver: annealed

•.«

These were first tested on May 9th, 1862, and at intervals
between that date and June 14th, 1863, when they were last
tested. During the time when not used, they were hung up
in a room where in the winter a fire was kept all day, so that the
temperature may have varied at times some 10 or 12 degrees in
the twenty-four hours.

The following Table contains the results of the first and last
comparisons. I have taken the conducting power in the first in
all cases equal to 100 as compared with No. 19; in the last I have
assumed that the conducting power of No. 15 has remained
unaltered: —

PBACnCAL STANDARDS

Conducting powers foun
compared with No. 19=

d, as
= 100

— !
1

Conducting 1
power found, as i
compared with

No. 15 « 100

May 9,
1862

June 14,
1863

1
1

1. Silver: hard-drawn

100-00
100-00
100-00
100-00
100-00
100-00
100-00
100-00
100-00
100-00
100-00
100-00
100-00
100-00
100 00
100-00
100-00
100-00

20-2
20-2
20-2
20-2
20-1
20-1
200
20-0
20-0
20-0
20-0
20-0
20-0
20-0
20 0
19-9
20-3
20-3

103-700
99-740

102-590
99-825

100-040
99-807
99-941
■95*358
99-838
99-855
99-662
99-670
99-744
99-792
99-793
99-766
99-955
99-938

20-0
20-1
20-2
20-0
20-2
20-0
19-8
20-4
20-2
20-0
20-2
20-3
20*2
20-2
20-2
20-3
20-0
20D

103-915

99-947

102-807

100-031

100-248

100-015

100149

95-556

100-045

100062

99-869

20-0
201
20-2
20-0
20-2
20-0
19-8
20-4
20*2
20-0
90-2

2. Silver: annealed

3. Silver: hard-drawn

4. Silver: annealed

5. Copper: hard-drawn

6. Copper: annealed

7. Copper: hard-drawn

8. Copper: annealed

9. Gola: hard-drawn

10. Gold: annealed

11. Gold: hard-drawn

1 12. Gold: annealed

13. Platinum: hard-drawn

14. Platinum: hard-drawn

99-877 20-3
99-951 20-2
99*999 20*2

15. Gold-silver alloy : hard-drawn

16. Gbld-silver alloy: hard-drawn

17. German-silver: annealed

18. German-silver : annealed

19. German-silver: annealed

100-000
99-963
100162
100-145
100-217

20-2
20-3
20-0
20-0
20*2

From the above it would appear that if the conducting power
of No. 19 has remained constant, that of all the others has altered;
but supposing such to be the case, it will be found on comparing^
the values that the conducting powers have all altered in a like
extent. Is it probable? Is it not more probable that the con-
ducting power of the German-silver has changed, than that that of
ail the others should have altered in the same degree ? If that of
the gold-silver alloy (No. 15) be called 10000 instead of 99*793,
then, as will be seen from column 3, very few show any change in
their conducting power. Those which show no sensible change are
as follows : —

No. 2. Silver: annealed...

No. 4. Silver: annealed...

No. 6. Copper: annealed

No. 9. Gola: hard-drawn

No. 10. Gold: annealed ...

No. 13. Platinum: hard-drawn

No. 14. Platinum: hard-drawn

No. 15. Gold-silver alloy : hard-drawn ...

No. 16. Gold-silver alloy : hard-drawn ...

Values taken from column 8
99-947

ioo-a3i

100-015
100-045
100-062

99-951

99-999
100-000

99-963

FOB SLECTBICAL MEASUBEHENTS 81

The differences in the above are probably due to temperature;
for as the wires are in tubes filled with carbonic-acid gas, we can never
be abeolutely sure that the wire has exactly the same temperature
as the bath. In properly made resistance-coils this source of error
is materially diminished, and in some experiments which are about
to be made to further test the electrical permanency of metals
and alloys this source of error will be almost entirely obviated.
It may be here again mentioned, that the reason for placing the
wires in glass tubes filled with carbonic-acid gas was to obviate
the oxidation of the metal or alloy by the oxygen of the air, or
firom the acids produced by the oxidation of the oil or &t with
which the wires are covered when drawn, as the holes in the
draw-plates are generally oiled or greased.

Those whose conducting powers have changed are as follows: —

Values taken from oolomn 3

Ka 1. Silver: hard-drawn 103*915

No. 3w Silver: hard-drawn 102-807

No. 5. Copper: hard-drawn ... ... 100-248

No. 7. Copper: hard-drawn 100*149

No. 8. Copper: annealed 95*556

No. 11. Qold: hard-drawn 99-869

No. 12. Qold: Annealed ...- 99-877

No. 17. German-silver: annealed ... 100*162

No. 18. Oermau-silver : annealed ... 100*145

No. 19. G^ermau-silver : annealed ... 100*217

The cause of the change in the conducting powers of the alloys
Noe. I, 3, 5, 7 is undoubtedly due to their becoming somewhat
annealed by age*. With No. 8 the alteration may be attributed to
&ulty soldering. That the conducting power of the German-silver
experimented with has altered is not a proof that all German-silver
will do so; for we find the gold wires Nos. 9 and 10 not altered,
but Nob. 11 and 12 (which were cut firom the same piece, but a
different one firom the one from which Nos. 9 and 10 were taken)
have altered. Further experiments are, however, required to find
whether the metals and alloys given above as constant in their
conducting power are so or not.

Schroder van der Kolk states f that the conducting power of
copper wire undergoes a change when even weak currents are
allowed to pass through it In order to see whether that of the

* Svpra, p. 28 ; and Brit. Auoc, Report, 1862, p. 140.
t Pogg. Ann, voL ex. p. 452.

a A. 6

82 PRACTICAL STANDARDS

above wires would suffer any change, the following experiment
was arranged: — Nos. 1, 2, 5, 6, 9, 10, 13, 15, 17 were connected
together, and a current from two Bunsen's cells was allowed to
pass through day and night for six days. The cells were cleaned
every morning and evening, and the dilute sulphuric acid renewed.
The experiment was carried out soon after June 14, 1863. In the
subjoined Table the conducting powers are given as found before
and after the trial, compared with No. 19.

Gonduoting power observed, as compared
with No. 19=100

Before T. After T.

No. 1 103-700 20-0 103-776 20-2

No. 2 99-740 201 99733 202

No. 6 100-040 20-2 100-045 20*2

No. 6 99-S07 20-0 99-865 20O

No. 9 99-838 20-2 99860 202

No. 10 99-865 20-0 99807 202

No. 13 99-744 202 99766 20-2

No. 15 99-793 202 99762 202

No. 17 99-956 20-0 99*926 20-2

From the above numbers it will be seen that the conducting
power has not chcmged, the differences in the values being in all
probability due, as above stated, to temperature.

If the passage of a current really altered the conducting power
of a wire, then of what use would resistance-coils be ? The above
experiments prove that a much stronger current than is used for
testing the resistance of a wire has no effect on it.

Appendix B. — On the Vai-iation of the Electric Besistance of
Alloys d/we to Change of Temperature. By A. Matthiessen, F.R.S.

In the Appendix to the Report of your Committee read at the
Meeting held last year, I gave a Table containing the results of
experiments with some alloys, made with a view to find out the
alloy whose conducting power decreases least with an increase of
temperature. With the same apparatus, etc., I have, in conjunction
with Dr C. Vogt, experimented with the following alloys.

(With each series the formula deduced from the observations
for the correction of the conducting power for temperature is
given, where X is equal to the conducting power at the tempera-
ture f C. Silver (hard-drawn) is taken at 0® ■= 100.)

FOB ELECTRICAL HEASUREMENTS

Composition of alloy by weight.

(1) Gold 96-3

Iron 4'7

Made from pure metals.

Hard-drawn.

Length 226 mm.; diameter 0*470 mm.

T. Conducting power
12'0 2-3573

660 2-3138

100-0 2-2798

X=2-3708 - 00011566*-fO-000002454^.

(2) Gold
Iron
Hard-drawn.

95-0
5-0

Length 284 mm.; diameter 1*217 mm.

T. Conduoting power

16-0 2-0819

67-6 2-0424

100-0 2W67

X = 2-0967 - 0-0010067* -H 0-000001052A

No. 2 and the two follo¥mig alloys were made by Messrs Johnson
and Matthey. No. 2 was made to check the results obtained with
Na 1, for those given with Nos. 3 and 4 appeared to show that
some mistake had been made with No. 1. That this was not the
case is proved by No. 2. It is, however, a veiy curious fact that
the percentage decrement increases in this manner, for in no other
series of alloys has this behaviour been noticed. Its cause may be
attributed to the existence of chemical combinations in the solid
alloys of gold and iron.

Nos. 3 and 4 are veiy brittle, and therefore difficult to draw.

(3) Gold
Iron
Hard-drawn.

•••

90O
10-0

Length 184 mm.; diameter 0-943 mm.

T. Condacting power

14-0 1-9822

670 1-7961

100-0 1-7010

X = 2-0632 - 0-0061367* -H 000002513^.

(4) Gold
Iron
Haid-drawn.

••■

•••

86-0
160

Length 146 mm.; diameter 0*768 mm.

T. Condnoting power

16-0 2-6239

67-6 2-2732

lOOO 1-9926

X = 2-7646 - 0-0096686* -I-0-00001940A

(5) Silver 75-0

Palladium ... 26-0
Made by Messrs Johnson

and Matthey.
Hard-drawn.

Length 620 mm.; diameter 0*802 mm.

ll-O

66*6

100-0

Conducting power
8-4846
8-3677
8-2266

X»8-6162 -0-0027644/ -0-000001313^,

This alloy was formerly used by dentists on account of its
elasticity. It was tested, as it appeared to answer some of the
conditions required*

6—2

PfiACTICAL STANDARDS

Composition of alloy by weight.

(6) Copper 63-3

Zinc 36*7

Made from pure metals.

Hard -drawn.

Length 296*6 mm.; diameter 0*576 mm.

(7) Copper 76-0

Zinc 25^0

Made from pure ilietals.
Hard-drawn.

Condocting power

15-72

21-807

23-75

21*562

39-28

21-116

54-38

20-693

69-31

20-300

84-63

19-897

99-43

19-327

190 mm.;

diameter 0*381 mi

GondoctiDg power

13-47

21-704

24-07

21*413

39-21

21O20

53-65

20*647

69^3

20-268

83-71

19*915

98-97

19-566

X=22-076 -0-028100<-f0-00002946««.

These alloys are given, as they approach in composition to that
of brass. It seemed very desirable to test the influence of tem-
perature on the alloy, as it was proposed by Jacobi as a unit of

electric resistance.

Length 322*5 mm. ; diameter 0*524 mm.
(8) Copper 90*3 t. Conduoting power

Copper 90*3

Tin 9*7

Made from pure metals.

Hard-drawn.

15-43 12-058

23*40 11*990

40-35 11-852

54-75 11-737

69-78 11*619

. 84-66 11-499

98-70 11-391

X= 12*186 - 0O084168/+ 0-000003700^.

(9) Copper 89-7

Tin 10-3

Made from pure metals.
Hard-drawn.

Length 429 mm.; diameter 0*627 mm.

T. Condaoting power

11-0 101386

55-5 9*8710

100*0 9-6526

X=10-212-0-0068043^-HO-00001210<«.

These alloys are given, as they approach in composition to that

of ordinary gun-metal.

Length 904*5 mm.; diameter 0-650 nun.
(10) Gun-metal (Austrian).
Copper.
Zinc.
Iron.
A specimen obtained through
the kindness of Mr F. Abel.
Hard-drawn.

T. Condacting power

130 26-336

56-5 24-056

100-0 22-121

X = 27*084 - 0-058750^ -1-0*000091 16^.

FOR ELECTRICAL MEASUREMENTS

The conducting power of this alloy increased by heating to
100** for one day 5*7 per cent. — a larger increment than has been
observed with any alloy. Generally, the conducting power of an
alloy either remains constant, or only varies O'l or 0*2 per cent.
under the same conditions.

Length 1564 mm.; diameter 0*526 mm.

T. Condacting power

16-0 68-969

57-5 60-179

100-0 53-387

X = 72-548 - 0-24692^ + 0-0005631 <».

(U) Pitwfgold.
Hard-drawn.

(12) Standard silver.
Hard-drawn.

Length 2328 mm.; diameter 0-525 mm.

T. Gondncting power

12-0 78-015

56-0 69*301

100-0 61-949

X=80'628 - 0-22196/+0-0003518<2.

In the following Table I have given the results here obtained,
with those of last year, in such a manner that they may be easily
compared: —

Pure iron*

Pure thallium*

Other pure metals in a solid state

Gold, with 15p.a iron

Proof gold

Standard silver

Gun-metal (Austrian)

Gold, with 10p.c. iron

Gold, with 14-3 p.a silver and 7*4 p.c. copper...

Copper, with 36*7 p.c zinc

Copper, with 25p.c. sdnc

Silver, with 5 p.c. platiniun*

Silver, with 9*8 p.c. platinum*

Copper, with 9*7 p.c. tin

The gold-silver alloj*

Platinum, with 33*4 p.a iridium

Copper, with 10*3 p.a tin

Golo, with 18*1 p.c. silver and 15*4 p.c. copper*
Gold, with 15-2 p.c. silver and 26*5 p.c. copper*

German-silver*

Gold, with 5p.c. iron

Gold, with 4*7 p.c. iron

Silver, with 25p.a palladium

Silver, with 33*4 p.c. platinumt

Condnoting

power

atO°

16*81
9*16

2*76

72*55

80*63

27-08

2-06

44*47

22*27

22*08

31*64

18-04

1219

16*03

4*54

10-21

10-6

12-02

7-80

210

2-37

8*52

6-70

Percentage

decrement in

oondncting power

between 0° A 100°

39*2

31*4

29*3

27*9

26-4

23-2

18*3

17*5

15*5

12*4

11-5

11-3

7*1

6-6

6*5

5*9

5*2

4-8
44

4-3
3-8
3-4
31

Proe. /toy. 8oe. xu. p. 472 (186S). t Supra, p. 20 ; and Brit. Assoc, Report, 1862, p. 187.

PRACTICAL STANDARDS

It will be observed that I have not yet been able to find an
alloy whose conducting power decreases between 0° and 100° less
than that of the alloy of silver with 334 p.c. platinum ; and fiixjm
results obtained in this direction in conjunction with Dr Vogt, I
am of opinion there will be great difficulty in doing so. We have
already tested upwards of 100 alloys, and it is curious how few we
have found whose conducting power varies less than that of German-
silver between 0° and 100°.

Appendix C. — On the Elementary Relations between Electrical
Measurements. By Professor J. Clerk Maxv^ll and
Mr Fleeming Jenkin.

TABLE OF CONTENTS.

Part I. — Introductory.

1. Objects of treatise.

2. Derivation of units from funda-

mental standards.

3. Standard mechanical units.

4. Dimensions of derived units.

Part 11 — The Measurement of Magnetic Phenomena.

6. Magnets and magnetic poles.

6. Magnetic field.

7. Magnetic moment.

8. Litensity of magnetization.

9. Coefficient of magnetic induction.

10. Magnetic potentials and equi-

potential surfaces.

11. Lines of magnetic force.

12. Relation between lines of force

and equipotential •surfaces.

Part III. — Measurement of Electric Phenomena by their

Electro-magnetic Effecto.

13. Preliminary.

14. Meaning of the words "electric

quantity.*'

15. Meaning of the words "electric

current,"

16. Meaning of the words "electro-

motive force."

17. Meaning of the words "electric

resistance."

18. Measurement of electric currents

by their action on a magnetic
needle.

19. Measurement of electric currents

by their mutual action on one
another.

20. Weber's Electro^ynamometer.

21. Comparison of the electro-mag-

netic and electro-chemical
action of currents.

22. Magnetic field near a current.

23. Mechanical action of a magnetic

field on a closed conductor
conveying a current.

FOR ELECTRICAL MEASUREMENTS

Part III. — continued.

24. Gieneral law of the mechanical ac-
tion between electric currents
and other electric currents or
magnets.

85. Electro-magnetic measurement of
electric quantity.

26. Electric capacity of a con-

ductor.

27. Direct measurement of electro-

motive force.

28. Indirect measurements of electro-

motive force.

29. Measurement of electric resist-

ance.
90. Electric resistance in electro-
magnetic units is measured
by an absolute velocity.

31. Magneto-electric induction.

32. On material standards for the

measurement of electric mag-
nitudes.

Part IV. — Measurement of Electric Phenomena by

Statical Effects.

33b Electrostatic measure of electric
quantity.

34. Electroetatio system of units.

35. Ratio between electrostatic and

electro-magnetic measures of
quantity.

36. Electrostatic measure of cur-

rents.

37. Electrostatic measure of electro-

motive force.

38. Electrostatic measure of resist-

ance:

39. Electric resistance in electro-

static units is measured by
the reciprocal of an absolute
velocity.

40. Electrostatic measure of the

capacity of a conductor.

41. Absolute condenser. Practical

measurement of quantity.

42. Practical measurement of cur-

rents.

43. Practical measurement of electro-

motive force.

44. Comparison of electromotive

forces by their statical efifiacts.

45. Practical measurement of electric

resistance.

46. Experimental determination of

the ratio v between electro-
magnetic and electrostatic
measures of quantity.

Part v. — Electrical Measurements derived from the
Five Elementary Measurements

47. Electric potential

46. Density, resultant electric force,
electric pressure.

48. Tension.

5(X Conducting power, specific resist-
ance, and specific conducting
power.

51. Spedfic inductive capacity.

52. Heat produced in a conductor by

a current.

53. Electro-chemical equivalents.

54. Electromotive force of chemical

affinity.

55. Tables of dimensions and other

constants.

56. Magnitude of units and nomen-

clature.

88 PRACTICAL STANDARDS

Part I. — Introductory.

1. Ohjecta of Treatise, — The progress and extension of the
electric telegraph has made a practical knowledge of electric and
magnetic phenomena necessary to a large number of persons who
are more or less occupied in the construction and working of the
lines, and interesting to many others who are unwilling to be
ignorant of the use of the network of wires which surrounds them.
The discoveries of Volta and Galvani, of Oersted, and of Faraday
are &miliar in the mouths of all who talk of science, while the
results of those discoveries are the foundation of branches of
industry conducted by many who have perhaps never heard of those
illustrious names. Between the student's mere knowledge of the
history of discovery and the workman's practical familiarity with
particular operations which can only be communicated to others
by direct imitation, we are in want of a set of rules, or rather
principles, by which the laws remembered in their abstract form
can be applied to estimate the forces required to effect any given
practical result.

We may be called on to construct electrical apparatus for a
particular purpose. In order to know how many cells are required
for the battery, and of what size they should be, we require to
know the strength of current required, the electromotive force of
the cells, and the resistance of the circuit. If we know the results
of previous scientific enquiry, and are acquainted with the method
of adapting them to the case before us, we may discover the
proper arrangement at once. If we are unable to make any
estimate of what is required before constructing the apparatus, we
may have to encounter numerous failures which might have been
avoided if we had known how to make a proper use of existing
data.

All exact knowledge is founded on the comparison of one
quantity with another. In many experimental researches con-
ducted by single individuals, the absolute values of those quantities
are of no importance; but whenever many persons are to act

FOR ELECTRICAL MEASUREMENTS 89

together, it is necessary that they should have a common under-
standing of the measures to be employed. The object of the
present treatise is to assist in attaining this common understand-
ing as to electrical measurements.

2. Derivation of Units from fundamental Standards. — Every
distinct kind of quantity requires a standard of its own, and these
standards might be chosen quite independently of each other, and
in many cases have been so chosen; but it is possible to deduce all
standards of quantity from the fundamental standards adopted for
length, time, and mass; and it is of great scientific and practical
importance to deduce them fix)m these standards in a systematic
manner. Thus it is easy to understand what a square foot is when
we know what a linear foot is, or to find the number of cubic feet
in a room from its length, breadth, and height; because the foot,
the square foot, and the cubic foot are parts of the same system of
units. But the pint, gallon, etc. form another set of measures of
volume which has been formed without reference to the system
based on length; and in order to reduce the one set of numbers to
the other, we have to multiply by a troublesome firaction, difficult
to remember, and therefore a finiitful source of error.

The varieties of weights and measures which formerly prevailed
in this country, when different measures were adopted for different
kinds of goods, may be taken as an example of the principle of
unsystematized standards, while the modem French system, in
which every thing is derived fix)m the elementary standards,
exhibits the simplicity of the systematic arrangement.

In the opinion of the most practical and the most scientific
men, a system in which every unit is derived firom the primary
units with decimal subdivisions is the best whenever it can be
introduced. It is easily learnt; it renders calculation of all kinds
simpler; it is more readily accepted by the world at large ; and it
bears the stamp of the authority, not of this or that legislator
or man of science, but of nature.

The phenomena by which electricity is known to us are of a
mechanical kind, and therefore they must be measured by me-
chanical units or standards. Our task is to explain how these
units may be derived from the elementary ones; in other words,
we shall endeavour to show how all electric phenomena may be
measured in terms of time, mass, and space only, referring briefly
in each case to a practical method of effecting the observation.

90 PRACTICAL STANDARDS

3. Standard Mechanical Units, — In this country the standard
of length is one yard, but a foot is the unit popularly adopted.
In France it is the ten millionth part of the distance from the
pole to the equator, measured along the earth's surface, according
to the calculations of Delambre; and this measure is called a
metre, and is equal to 3-280899 feet, or 39*37079 inches.

In the original Report the metre was taken as the fundamental
unit of length ; the gramme, or French standard of weight, is not,
however, systematically derived from the metre, being the weight
not of a cubic metre, but of a cubic centimetre of water. This
consideration has led several Members of the Committee in
subsequent writings to adopt the centimetre as the fundamental
unit of length. To facilitate comparison with these writings,
constants based on the centimetre will be given, besides those
for the metre.

The standard unit of time in all civilized countries is deduced
from the time of rotation of the earth about its axis. The sidereal
day, or the true period of rotation of the earth, can be ascertained
with great exactness by the ordinary observations of astronomers;
and the mean solar day can be deduced from this by our knowledge
of the length of the year. The unit of time adopted in all physical
researches is one second of mean solar time.

The standard unit of mass is in this country the avoirdupois
pound, as we received it from our ancestors. The grain is one
7000th of a pound. In the French sjrstem it is the gramme
derived from the unit of length, by the use of water at a standard
temperature as a standard of density. The weight of one cubic
centimetre of water is a gramme = 15*43235 grains = '00220462 lb.

A Table showing the relative value of the standard and
derived units in the British and metrical system is given in § 55.

The unit of force adopted in this treatise is that force which
will produce a unit of velocity in a free unit mass, by acting on it
during a unit of time. This unit of force is equal to the weight
of the unit mass divided by g, where g is the accelerating force of
gravity, the value of which, as depending on the place of obser-
vation, is given in § 55. In this country it is about 32*2 feet, or
981 centimetres.

A unit of force still very generally used is the weight of the
standard mass. This is called the gravitation unit of force, and
measurements of force, work, etc. in which it is used are called

FOR ELECTRICAL MEASUREMENTS 91

gravitation measurements. The gravitation unit is equal to the
absolute unit multiplied by g.

The unit of work adopted in this treatise is the unit of force,
defined as above, acting through the unit of space {vide § 55).

4. Dimensions of Derived Units. — The name of every quantity
consists of two factors or components, and may be written thus, Q [Q\

The first, or numerical factor, Q, is a number, integral or
fractional. The second, or denominational factor, [Q], is the name
of an individual thing of the same kind as the quantity to be
expressed, the magnitude of which is agreed on among men.
Thus, in the expression 28 lb., 28 is the numerical part represented
by Qy and lb. is the denominational part represented by [Q].
When Q is unity, then the quantity expressed is the unit, 1[Q],
or simply [Q]. In the example [Q] is one pound ; that is, a piece
of platinum preserved in the Exchequer Chambers, and marked
"P.S. 1844, 1 lb.," or some copy of the same.

We shall use the symbols [Z], [Jf], and [T] enclosed in square
brackets to denote the standards or units of length, mass, and time;
and symbols without brackets, such as Z, Jlf, T, to denote the
number of such units in the quantity to be expressed. Thus if
[Z] denotes a centimetre and L the number 978, L \L] denotes
978 centimetres. Similarly, if [2] denotes 1 foot and / the number
32*088, 1 [/] denotes 32*088 feet.

Now these quantities express the same distance measured in
two different ways, so that

but 1 foot is 30*479 centimetres, or

[q = 30-479 [L\.

Hence L = 30479/;

or the numerical &ctor of the expression of a given quantity varies
inversely as the magnitude of the unit employed.

In passing from one system of measurements to another, we
first consider the magnitude of the units employed in the two
systems and then determine the numerical &ctors so that the
quantity expressed may be the same. Every measurement of
which we have to speak involves as factors measurements of time,
space, and mass only; but these measurements enter sometimes
at one power and sometimes at another. In passing from one
set of fundamental units to another, and for other purposes, it

92 PRACTICAL STANDARDS

is necessary to know at what power each of these fundamental
measurements enters into the derived hieasure.

Thus the value of a force is directly proportional to a length
and a mass, but inversely proportional to the square of a time.
This is expressed by saying that the diviensions of a force are

-jfp ; in other words, if we wish to pass from the English to the

French system of measurements, the French unit of force will be

X xu 17 r u •0328x15-43 ^ ^^^ . ^ .

to the Enghsh as : 1, or as '506 to 1 ; because there

are '0328 feet in a centimetre, and 15*43 grains in a gramme. If
the metre be adopted as the unit of length, the French unit of
force will be to the English as 50'6 to 1. If the minute were
chosen as the unit of time, the unit of force would, in either system,
be 7^ of that founded on the second as unit.

A Table of the dimensions of every unit adopted in the present
treatise is given in § 55.

Part II. — The Measurement of Magnetic Phenomena.

5. Magnets and Magnetic Poles, — Certain natural bodies, as
the iron ore called loadstone, the earth itself, and pieces of steel
after being subjected to certain treatment, are found to possess
the following properties, and are called magnets.

If one of these bodies be free to turn in any direction, the
presence of another will cause it to set itself in a position which is
conveniently described or defined by reference to certain imaginary
lines occupying a fixed position in the two bodies, and called their
magnetic axes. One object of our magnetic measurements will be
to determine the force which one magnet exerts upon another.
It is found by experiment that the greatest manifestation of force
exerted by one long thin magnet on another occurs very near the
ends of the two bars, and that the two ends of any one long thin
magnet possess opposite qualities. This peculiarity has caused
the name of ''poles" to be given to the ends of long magnets; and
this conception of a magnet, as having two poles capable of exerting
opposite forces joined by a bar exerting no force, is so much the most
familiar that we shall not hesitate to employ it, especially as many
of the properties of magnets may be correctly expressed in this

FOB ELECTRICAL MEASUREMENTS 93

way; but it mast be borne in mind, in speaking of poles, that
they do not really exist as points or centres of force at the ends
of the bar, except in the case of long, infinitely thin, uniformly
magnetized rods.

If we mark the poles of any two magnets which possess similar
qualities, we find that the two marked poles repel each other, that
two unmarked poles also repel each other; but that a marked and
an unmarked pole attract each other. The pole which is repelled
from the northern regions of the earth is called the positive pole ;
the other end the negative pole. The negative pole is generally
marked N by British instrument-makers, and is sometimes called
the north pole of the magnet, though it is obviously similar to
the earth's south pole.

The strength of the pole is necessarily defined as proportional
to the force it is capable of exerting on any other pole. Hence the
force / exerted between two poles of the strengths m and mi must
be proportional to the product mmi. The force / is also found to
be inversely proportional to the square of the distance, D, sepa<^
rating the poles, and to depend on no other quantity; hence we
have, unless an absurd and useless coefficient be introduced,

/-^ (1)

From which equation it follows that the unit pole will be that
which at unit distance repels another similar pole with unit force ;
/ will be an attraction or a repulsion according as the poles are
of opposite or the same kinds. The dimensions of the unit

magnetic pole are — ~- .

6. Magnetic Field. — It is clear that the presence of a magnet
in some way modifies the surrounding space, since any other
magnet brought into that space experiences a peculiar force. The
neighbourhood of a magnet is, for convenience, called a magnetic
field; and for the same reason the effect produced by a magnet is
often spoken of as due to the magnetic field, instead of to the
magnet itself. This mode of expression is the more proper, inas-
much as the same or a similar condition of space may be produced
by the passage of electrical currents in the neighbourhood, without
the presence of a magnet. Since the peculiarity of the magnetic

94 PRACTICAL STANDARDS

field consists in the presence of a certain force, we may numerically
express the properties of the field by measuring the strength and
direction of the force, or, as it may be worded, the intensity of the
field and the direction of the lines of force.

This direction at any point is the direction in which the force
tends to move a free pole ; and the intensity, i/, of the field is
completely defined as proportional to the force,/, with which it
acts on a free pole ; but this force, /, is also proportional to the
strength, m, of the pole introduced into the field, and it depends
on no other quantities; hence

/=mF, (2)

and therefore the field of rmit intensity will be that which acts
with unit force on the unit pole.

The dimensions of [H] are

LirJ-

The lines of force produced by a long thin bar-magnet near its
poles radiate firom the poles, and the intensity of the field is equal
to the quotient of the strength of the pole divided by the square
of the distance from the pole ; thus the unit field is produced at
the unit distance fix)m the unit pole.

In a uniform magnetic field the lines of force, as may be
demonstrated, are parallel ; such a field can only be produced by
special combinations of magnets, but a small field at a great
distance from any one pole is sensibly uniform. Thus, in any room
unaffected by the neighbourhood of iron or magnets, the magnetic
field due to the earth is sensibly uniform; its direction is that
assumed by the dipping-needle.

7. Magnetic Moment, — When a bar-magnet is placed in a
uniform field two equal opposite and parallel forces act on its
poles, and tend to set it with the line joining those poles in the
direction of the force of the field. When the magnet is so placed
that the line joining the poles is at right angles to the lines of force
in the field, this tendency to turn or "couple," (?, is proportional to
the intensity of the field, H, the strength of the poles, m, and the
distance between them, /; or

G^mlH. (3)

ml, or the product of the strength of the poles into the length

FOR ELECTRICAL MEASUREMENTS 95

between them, is called the magnetic moment of the magnet; and
from equation (3) it follows that, in a field of unit intensity, the
couple actually experienced by any magnet in the above position
measures its moment. The dimensions of the unit of magnetic

moment are evidently — — — .

8. Intensity of Magnetization. — ^The intensity of magnetization
of a magnet is measured by its magnetic moment divided by its
volume.

The dimensions of the unit of magnetization are therefore

—J— I , the same as in the case of intensity of field.

9. Coefficient of Magnetic Induction, — When certain bodies>
such as soft iron, etc., are placed in the magnetic field, they become
magnetized by "induction"; so that the intensity of magnetiza-
tion is (except when great) nearly proportional to the intensity of
the field.

In diamagnetic bodies, such as bismuth, the direction of mag«
netization is opposite to that of the field. In paramagnetic bodies,
such as iron, nickel, etc., the direction of magnetization is the same
as that of the field.

The coefficient of magnetic induction is the ratio of the intensity
of magnetization to the intensity of the magnetic force within the
body, and is therefore a numerical quantity, positive for paramag-
netic bodies, negative for diamagnetic bodies.

10. Magnetic Potentials and EquipoteniicU Surfaces, — If we
take a very long magnet, and, keeping one pole well out of the
way, move the other pole fix)m one point to another of the magnetic
field, we shall find that the forces in the field do work on the pole,
or that they act as a resistance to its motion, according as the
motion is with or contrary to the force acting on the pole. If the
pole moves at right angles to the force, no work is done.

The magnetic potential at any point in a magnetic field is
measured by the work done against the magnetic forces on a unit
pole during its motion firom an infinite distance from the magnet
producing the field to the point in question, supposing the unit
pole to exercise no influence on the magnetic field in question.
The idea of potential as a mathematical quantity having different
values at different points of space was brought into form by

96 PBACTICAL STANDARDS

Laplace*. The name of potential, and the application to a great
number of electric and magnetic investigations, were introduced
by George Green in his Essay on Electricity (Nottingham, 1828).

An equipotential surfiu^ in a magnetic field is a surfiau^ so
drawn that the potential of all its points are equal. By drawing
a series of equipotential suri'aces corresponding to potentials
1, 2, 3 ... n, we may map out any magnetic field, so as to indicate
its properties.

The magnetic force at any point is perpendicular to the equi-
potential surface at that point, and its intensity is the reciprocal
of the distance between one surface and the next at that point.

The dimensions of the unit of magnetic potential are —^p — .

11. Lines of Magnetic Force. — ^There is another way of
exploring the magnetic field, and indicating the direction and
magnitude of the force at any point. The conception and appli-
cation of this method in all its completeness is due to Faraday "f.
The full import€ince of this method cannot be recognized till we
come to electro-magnetic phenomena (§§ 22, 23, and 24).

A line whose direction at any point always coincides with that
of the force acting on the pole of a magnet at that point, is called
a line of magnetic force. By drawing a sufficient number of such
lines we may indicate the direction of the force in every part of
the magnetic field; but by drawing them according to rule, we
may indicate the intensity of the force at any point as well as its
direction. It has been shown :^ that if, in any part of their course,
the number of lines passing through unit of area is proportional
to the intensity there, the same proportion between the number
of lines in unit of area and the intensity will hold good in every
part of the course of the lines.

All that we have to do, therefore, is to space out the lines in
any part of their course, so that the number of lines which start
from unit of area is equal to the number representing the intensity
of the field there. The intensity at any other part of the field will
then be measured by the number of lines which pass through unit
of area there ; each line indicates a constant and equal force.

* Micanique Celeste ^ lit. iii.

t Experimental Researches^ vol. in. art. 3122 et passim.

t Vide Maxwell on Faraday's Lines o£ Force, Cambridge Phil. Trans. 1857.

FOR ELECTRICAL MEASUREMENTS 9t

12. Relation between Lines of Farce and Equipotential Swr-
faoee. — ^The lines of force are always perpendicular to the equi-
potential sur&ces ; and the number of lines passing through unit
of area of an equipotential surface is the reciprocal of the distance
between that equipotential surface and the next in order — ^a state-
ment made above in slightly different language.

In a uniform field the lines of force are straight, parallel, and
equidistant; and the equipotential surfaces are planes perpen-
dicular to the lines of force, and equidistant fix>m each other.

If one magnetic pole of strength m be alone in the field, its
lines of force are straight lines, radiating from the pole equally
in all directions; and their number is Airm. The equipotential
sorfitces are a series of spheres whose centres are at the pole, and
whose radii are m, ^m, ^m, ^m, etc. In other magnetic arrange-
ments these lines and surfaces are more complicated; but in all
cases the calculation is simple, and in many cases the lines and
8iirfiu»8 can be graphically constructed without any calculation.

Part III. — ^Measurebcent of Electric Phenomena by their

Electro-Magnetic Effects.

13. Prdiminary. — ^Before treating of electrical measurements,
the exact meaning in which the words "quantity," ''current,"
** electromotiye force," and ''resistance" are used must be ex-
plained. But in giving these explanations, we shall assume the
reader to be acquainted with the meaning of such expressions as
conductor, insulator, voltaic battery, etc.

14. Meaning of ike words ''Electric Quantity" — When two
light conducting bodies are connected with the same pole of a
voltaic battery, while the other pole is connected with the earth,
they may be observed to repel one another. The two poles
produce equal and similar effects. When the two bodies are
connected with opposite poles, they attract one another. Bodies,
when in a condition to exert this peculiar force one on the other,
are said to be electrified or charged with electricity. These words
are mere names given to a peculiar condition of matter. If a
|Hece of glass and a piece of resin are rubbed together, the glass
will be found to be in the same condition as an insulated body
connected with the copper pole of the battery, and the resin in
the same condition as the body connected with the zinc pole of

& A. 7

•98 PBAOTICAL STANDARDS

the batteiy. The former is said to be positively and the latter
negatively electrified. The propriety of this antithesis will soon
appear. The force with which one electrified body acts on
another, even at a constant distance, varies with different circum-
stances. When the force between the two bodies at a constant
distance, and separated by air, is observed to increase, it is said to
be due to an increase in the quantity of electricity; and the
quantity at any spot is defined as proportional to the force with
which it acts, through air, on some other constant quantity at a
distance. If two bodies, charged each with a given quantity of
electricity, are incorporated, the single body thus composed will
be charged with the sum of the two quantities. It is this fact
which justifies the use of the word " quantity."

Thus the quality in virtue of which a body exerts the peculiar
force described is called electricity, and its quantity is measured
{coBteris paribus) by measuring force,

.The quantity, thus defined, produced on two similar balls
similarly circumstanced, but connected with opposite poles of a
voltaic battery, is equal, but opposite ; so that the sum of these
two equal and opposite quantities is zero ; hence the conception of
positive and negative quantities.

In speaking of a quantity of electricity, we need not conceive
it as a separate thing, or entity distinct fix)m ponderable matter,
any more than in speaking of sound we conceive it as having a
distinct existence. Still it is convenient to speak of the intensity
or velocity of sound to avoid tedious circumlocution ; and quite
similarly we may speak of electricity without for a moment
imagining that any real electric fiuid exists.

The laws according to which this force varies, as the shape of
the conductors, their combinations, and their distances are varied,
have been established by Coulomb, Poisson, Green, W. Thomson,
and others. These will be found accurately described, indepen->>
dently of all hypothesis, in papers by Professor W. Thomson,
published in the Ca/mbridge Maihematical Journal, vol. I. p. 75
(1846), and a .series of papers in 1848 and 1849.

16.. Meaning of the words ''Electric Current," — When two
balls charged by the opposite poles of a battery with opposite and
equal quantities of electricity are joined by a conductor, they lose
in a very short time their peculiar properties, and assume a
neutral condition intermediate between the positive and negative

FOR ELECTRICAL MEASUREMENTS 99

states, exhibiting no electrical symptoms whatever, and hence
described as unelectri£ed, or containing no electricity. But,
daring the first moment of their junction, the conductor is found
to possess certain new and peculiar properties: any one part of
the conductor exerts a force upon any other part of the conductor;
it exerts a force on any magnet in the neighbourhood; and if any
part of the conductor be formed by one of those compound bodies
called electrolytes, a certain portion of this body will be decom-
posed. These peculiar effects are said to be due to a current of
electricity in the conductor. The positive quantity, or excess, is
conceived as flowing into the deficiency represented by the
n^;ative quantity; so that the whole combination is reduced to
the neutral condition. This neutral condition is similar to that
of the earth where the experiment is tried. If the balls are
continually recharged by the battery, and discharged or neutra-
lized by the wire, a rapid succession of the so-called currents will
be sent; and it is found that the force with which a magnet
is deflected by this rapid succession of currents is proportional
{eoBteris paribus) to the quantity of electricity passed through the
conductor per second; it is also found that the amount of chemical
action, measured by the weights of the particular substances
decomposed, is proportional to the same quantity. The currents
just described are intermittent; but a wire or conductor, used
simply to join the two poles of a battery, acquires permanently
the same properties as when used to discharge the balls as above
with great rapidity; and the greater the rapidity with which the
balls are discharged, the more perfect the similarity of the
condition of the wire in the two cases. The wire in the latter
case is therefore said to convey a permanent current of electricity,
the magnitude or strength of which is defined as proportional to
the quantity conveyed per second. This definition is expressed
by the equation

^-?. (*)

where C is the current, Q the quantity, and t the time. A
permanent current flowing through a wire may be measured by
the force which it exerts on a magnet; the actual quantity it
conveys may be obtained by comparing this force with the force
exerted, under otherwise similar conditions, when a known quantity
is sent through the same wire by discharges. The strength of a

7—2

100 PHACTICAL STANDARDS

permanent current is found at any one time to be equal in all
parts of the conductor. Conductors conveying currents exert a
peculiar force one upon another; and during their increase or
decrease they produce currents in neighbouring conductors.
Similar effects are produced as they approach or recede from
neighbouring conductors. The laws according to which currents
act upon magnets and upon one another will be found in the
writings of Ampfere and Weber.

16. Meaning of the words ** Electromotive Force'* — Hitherto
we have spoken simply of statical effects; but it is found that
a cuzrent of electricity, as above defined, cannot exist without
effecting work or its equivalent. Thus it either heats the
conductor, or raises a weight, or magnetizes soft iron, or effects
chemical decomposition; in fine, in some shape it effects work,
and this work bears a definite relation to the current. Work
done presupposes a force in action. The immediate force pro-
ducing a current, or, in other words, causing the transfer of a
certain quantity of electricity, is called an electromotive force.
This force is necessarily assumed as ultimately due to that part of
a circuit where a "degradation" or consumption of energy takes
place : thus we speak of the electromotive force of the voltaic or
thermo-electric couple; but the term is also used, independently
of the source of power, to express the fact that, however caused, a
certain force tending to do work by setting electricity in motion
does, under certain circumstances, exist between two points of a
conductor or between two separate bodies. But equal quantities
of electricity transferred in a given time do not necessarily or
usually produce equal amounts of work; and the electromotive
force between two points, the proximate cause of the work, is
defined as proportional to the amount of work done between
those points when a given quantity of electricity is transferred
from one point to another. Thus if, with equal currents in two
distinct conductors, the work done in the one is double that done
in the other in the same time, the electromotive force in the first
case is said to be double that in the second ; but if the work done
in two circuits is found strictly proportional to the two currents,
the electromotive force acting on the two currents is said to be
the same. Defined in this way, the electromotive force of a
voltaic battery is found to be constant so long as the materials
of which it is formed remain in a similar or constant condition.

FOR ELECTRICAL MEASUREMENTS 101

The above definitions, in mathematical language, give Ws* ECt, or

^'ct <5)

where E is the electromotive force, and W the work done. Thus
the electromotive force producing a current in a conductor is
equal to the ratio between the work done in the unit of time and
the current effecting the work. This conception of the relations
of work, electromotive force, current, and quantity will be aided by
the foUowing analogy : — ^A quantity of electricity may be compared
to a quantity or given mass of water ; currents of water in pipes
in which equal quantities pass each spot in equal times then
correspond to equal currents of electricity; electromotive force
corresponds to the head of water producing the current. Thus
iC with two pipes conveying equal currents, the head forcing the
water through the first were double that forcing it through the
second, the work done by the water in flowing through the first
pipe would necessarily be twice that done by the water in the
second pipe ; but if twice as much water passed through the first
pipe as passes through the second, the work done by water in the
first pipe would again be doubled. This corresponds exactly with
the increase of work done by the electrical current when the
electromotive force is doubled and when the quantity is doubled.

Thus, to recapitulate, the quality of a battery or source of
electricity, in virtue of which it tends to do work by the transfer
of electricity from one point to another, is called its electromotive
force, and this force is measured by measuring the work done
during the transfer of a given quantity of electricity between
these points. The relations between electromotive force and work
were first fiiUy explained in a paper by Professor W. Thomson,
on the application of the principle of mechanical effect to the
measurement of electromotive forces, published in the Philosophical
Magctsine for December, 1851.

17. Meaning of the words "Electric Resistance.** — It is found
by experiment that even when the electromotive force between
two points remains constant, so that the work done by the
transfer of a given quantity of electricity remains constant, never-
theless, by modifying the material and form of the conductor, this
teansfer may be made to take place in very different times ; or, in
other words, currents of very different magnitudes are produced.

102 PRACTICAL STANDARDS

and very different amounts of work are done, in the unit of time.
The quality of the conductor in virtue of which it prevents the
performance of more than a certain amount of work in a given
time by a given electromotive force is called its electrical resist-
ance. The resistance of a conductor is therefore inversely pro-
portional to the work done in it when a given electromotive force
is maintained between its two ends ; and hence, by equation (5), it
is inversely proportional to the currents which will then be pro-
duced in the respective conductors. But it is found by experiment
that the current produced in any case in any one conductor is
simply proportional to the electromotive force between its ends;

hence the ratio 77 will be a constant quantity, to which the resist-

ance as above defined must be proportional, and may with con-
venience be made equal ; thus

E
i2 = g. (6)

an equation expressing Ohm's law. In order to carry on the
parallel with the pipes of water, the resistance overcome by the
water must be of such nature that twice the quantity of water
will flow through any one pipe when twice the head is applied.
This would not be the result of a constant mechanical resistance,
but of a resistance which increased in direct proportion to the
speed of the current ; thus the electrical resistance must not be
looked on as analogous to a simple mechanical resistance, but
rather to a coefficient by which the speed of the current must be
multiplied to obtain the whole mechanical resistance. Thus if
the electrical resistance of a conductor be called J2, the work, IT,
is not equal to CRtt but C x CR x <, or

TF= C*IU*, (7)

where C may be looked on as analogous to a quantity moving at a
certain speed, and CR as analogous to the mechanical resistance
which it meets with in its progress, and which increases in direct
proportion to the quantity conveyed in the unit of time.

18. Measurement of Electric Cwrrents by their Action on a
Magnetic Needle. — In 1820, Oersted discovered the action of an

E
* By equation (5) we have W=CEt; but by equation (6) R=-^; henoe

ir= CBt.— O.B.D.

FOR ELECTRICAL MEASUREMENTS 103

electric carrent upon a magnet at a distance, and one method of
measurement may be based on this action* Let us suppose the
current to be in the circumference of a vertical circle, so that in
the upper part it runs finom left to right. Then a magnet sus-
pended in the centre of the circle will turn with the end which
points to the north away from the observer. This may be taken
as the simplest case, as eveiy part of the circuit is at the same
distance from the magnet, and tends to turn it the same way.
The force is proportional to the moment of the magnet, to the
strength of the current as defined by § 15, to its length, and
inversely to the square of its distance frx>m the magnet.

Let the moment of the magnet be ml, the strength of the
current C, the radius of the circle k, the number of times the
current passes round the circle n, the angle between the axes of
the magDet and the plane of the circle 0, and the moment tending
to turn the magnet O, then

G = nUC.27mk^co8 0, (8)

which will be unity if ml, C, k, and the length of the circuit be
unity, and if 0 = 0''.

The unit of current founded on this relation, and called the
electro-magnetic unit, is therefore that current of which the unit
of length placed along the circumference of a circle of unit radius
produces a unit of magnetic force at the centre.

The usual way of measuring C, the strength of a current, is by
making it describe a circle about a magnet, the plane of the circle
being vertical and magnetic north and south. Thus, if If be the
intensity of the horizontal component of terrestrial magnetism,
and 0 the moment of this on the magnet, 6 = mlHamd, whence
the strength of the current

^-2^^**°^' <»)

where k is the radius of the circle, n the number of turns, H the
intensity of the horizontal part of the earth's magnetic force as
determined by the usual method, and 0 the angle of deviation of
the magnet suspended in the centre of the circle. As the strength
of the current is proportional to the tangent of the angle 0, an
instrument constructed on this plan is called a tangent galvano-
meter. The instrument called a sine galvanometer may also be

104 PRACTICAL STANDARDS.

Used, provided the coil is circular. The equation is similar to
that just given, substituting sin 9 for tan 0,

To find the dimensions of [G\ the imit electric current, we
must consider that what we observe is the force acting between a
magnetic pole, m, and a current of given length, L, at a given

distance, A, and that this force = ^ ^ . Hence the dimensions

of [0], the unit electric current, are — jp— .

19. Measurement of Electric Currents by their mutual action
on one another. — Hitherto we have spoken of the measurement of
currents as dependent on their action upon magnets; but this
measurement in the same units can as simply be founded on their
mutual action upon one another. Ampere has investigated the
laws of mechanical action between conductors carrying currents.
He has shown that the action of a small closed circuit at a
distance is the same as that of a small magnet, provided the axis
of the magnet be placed normal to the plane of the circuit, and
the moment of the magnet be equal to the product of the current
into the area of the circuit which it traverses.

Thus, let two small circuits, having areas A and Aj, be placed
at a great distance, D, from each other in such a way that their
planes are at right angles to each other, and that the line D is in
the intersection of the planes. Now let currents, C and Ci,
circulate in these conductors; a force will act between them
tending to make their planes parallel, and the direction of the
currents opposite. The moment of this couple will be

(? = d£j^. (10)

Hence the unit electric current conducted round two circuits
of unit area in vertical planes at right angles to each other, one
circuit being at a great distance, D, vertically above the other,

will cause a couple to act between the circuits of a magnitude jg .

The definition of the unit current (identical with the unit founded
on the relations given in § 18) might be founded on this action
quite independently of the idea of magnetism.

20. Weber' 8 Electro-dynamometer. — The measurement described
in the last paragraph is only accurate when D is very great, and

FOR ELECTRICAL MEASUREMENTS 105

therefore the moment to be measured very small. Hence it is
better to make the experimental measurements in another form.
For this purpose, let a length (F) of wire be made into a circular
coil of radius £ ; let a length (li) of wire be made into a coil of
very much smaller radius, A^. Let the second coil be hung in the
centre of the first, the planes being vertical and at the angle 0.
Then, if a current C traverses both coils, the moment of the force
tending to bring them parallel will be

O^^C^^-^sind. (11)

This force may be measured in mechanical units by the angle
through which it turns the suspended coil, the forces called into
play by the mechanical arrangements of suspension being known
bom the construction of the instrument. Weber used a bifilai*
suspension, by which the weight of the smaller coil was used to
resist the moment produced by the action of the currents.

21. Comparison of the Electro-magnetic and Electro-chemical
action of Currents, — Currents of electricity, when passed through
certain compound substances, decompose them; and it is found
that, with any given substance, the weight of the body decom-
posed in a given time is proportional to the strength of the
current as already defined with reference to its electro-magnetic
effect. The voltameter is an apparatus of this kind, in which
water is the substance decomposed. Special precautions have to
be taken, in carrying this method of measurement into effect, to
prevent variations in the resistance of the circuit, and consequently
in the strength of the current. This subject is more fiiUy treated
in Part V, §§ 53, 54.

22. Magnetic Field near a Current — Since a current exerts a
force on the pole of a magnet in its neighbourhood, it may be said
to produce a magnetic field (§ 6), and, by exploring this field with
a magnet, we may draw lines of force and equipotential surfaces
of the same nature as those already described for magnetic fields
caused by the presence of magnets.

When the current is a straight line of indefinite length, like a
telegraph-wire, a magnetic pole in its neighbourhood is urged by
a force tending to turn it round the wire, so that this force is at
any point perpendicular to the plane passing through this point
and the axis of the current

106 PRACnCAL STANDARDS

The equipotential sur&ces are therefore a series of planes
passing through the axis of the current, and inclined at equal
angles to each other. The number of these planes is ^ttC, where
C is the strength of the current.

The lines of magnetic force are circles having their centres in

the axis of the current, and their planes perpendicular to it. The

intensity of the magnetic force at a distance, k^ from the current

is the reciprocal of the distance between two equipotential sur-

2(7
faces, which shows the force to be -77 •

The work done on a unit magnetic pole in going completely
round the current is AnrC, whatever the path which the pole
describes.

23. Mechanical Action of a Magnetic Field on a closed
Conductor conveying a Current. — When there is mechanical action
between a conductor carrying a current and a magnet, the force
€tcting on the conductor must be equal and opposite to that
acting on the magnet. Every part of the conductor is therefore
acted on by a force perpendicular to the plane passing through its
own direction and the lines of magnetic force due to the magnet,
and equal to the product of the length of the conductor into the
strength of the current, the intensity of the magnetic field, and
the sine of the angle between the lines of force and the direction
of the current. This may be more concisely expressed by saying,
that if a conductor carrying a current is moved in a magnetic
field, the work done on the conductor by the electro-magnetic
forces is equal to the product of the strength of the current into
the nwmber of lines of force which it cuts during its motion.

Hence we arrive at the following general law, for determining
the mechanical action on a closed conductor canying a current
and placed in a magnetic field: —

Draw the lines of magnetic force. Count the number which
pass through the area enclosed by the circuit of the conductor,
then any motion which increases this number will be aided by the
electro-magnetic forces ; so that the work done during the motion
will be the product of the strength of the current and the number
of additional lines of force.

For instance, let the lines of force be due to a single magnetic
pole of strength m. These are 4nrm in number, and are in this
case straight lines radiating equally in all directions from the

FOR ELECTRICAL MEASUREMENTS 107

pole. Describe a sphere about the pole, and project the circuit
on its surfiBfcce by lines drawn to the pole. The surface of the
area so described on the sphere will measure the solid angle
subtended by the circuit at the pole. Let this solid angle = a>,
then the number of lines passing through the closed surface will
be m^ ; and if (7 be the strength of the current, the amount of
work done by bringing the magnet and circuit from an infinite
distance apart to their present position will be Cmto. This shows
that the magnetic potential of a closed circuit carrying a unit
current with respect to a unit magnetic pole placed at any
point is equal to the solid angle which the circuit subtends at
that point.

By considering at what points the circuit subtends equal solid
angles, we may form an idea of the surfaces of equal potential.
They form a series of sheets, all intersecting each other in the
circuit itself, which forms the boundary of every sheet. The
number of sheets is ^C, where C is the strength of the current.
The lines of magnetic force intersect these surfaces at right
angles, and therefore form a system of rings encircling every point
of the circuit. When we have studied the general form of the
lines of force, we can form some idea of the electro-magnetic
action of that current, after which the difficulties of numerical
calculation arise entirely from the imperfection of our mathe-
matical skill.

24. Oeneral Law of the Mechanical Action between Electric
Ourrents and other Electric Currents or Magnets. — ^Draw the lines
of magnetic force due to all the currents, magnets, etc., in the
field, supposing the strength of each current or magnet to be
reduced firom its actual value to unity. Call the number of lines
of force due to a circuit or magnet, which pass through another
circuit, the potential coefficient between the one and the other;
This number is to be reckoned positive when the lines of force
pass through the circuit in the same (direction as those due to a
current in that circuit, and negative when they pass in the
opposite direction.

If we now ascertain the change of the potential coefficient due
to any displacement, this increment multiplied by the product of
the strengths of the currents or magnets will be the amount of
work done by the mutual action of these two bodies during the
displacement. The determination of the actual value of the

108 PRACnCAL STANDARDS

potential coefficient of two things, in various cases, is an im-
portant part of mathematics as applied to electricity. (See the
mathematical discussion of the experiments, Appendix D.)

25. Electro-magnMic MeoLaurement of Electric Qua/ntity. — A
conducting body insulated at all points from the neighbouring^
conductors may in various ways be electrified, or made to hold a
quantity of electricity. This quantity (§ 14) is perfectly definite
in any given circumstances ; it cannot be augmented or diminished
so long as the conductor is insulated, and is called the charge of
the conductor. Its magnitude depends on the dimensions and
shape and position of the insulated and the neighbouring con-
ductors, on the insulating material, and finally on the electro-
motive force between the insulated and the neighbouring
conductors at the time when the charge was produced. The
weU-known Leyden jar ia an arrangement by which a considerable
charge can be obtained on a small conductor with moderate
electromotive force between the inner and outer coatings which
constitute respectively the "insulated" and "neighbouring" con-
ductors referred to in general. We need not enter into the
general laws determining the charge, since our object is only to
show how it may be measured when already existing ; but it may
be well to state that the quantity on the charged insulated
conductor necessarily implies an equal and opposite quantity on
the surrounding or neighbouring conductors.

We have already defined the magnitude of a current of
electricity as simply proportional to the quantity of electricity
conveyed in a given time, and we have shown a method of
measuring currents consonant with this definition. The unit
quantity will therefore be that conveyed by the unit current as
above defined in the unit of time. Thus if a unit current is
allowed to flow for a unit of time in a wire connecting the two
coatings of a Leyden phial, the quantity which one coating loses
or which the other gains is the electro-magnetic unit quantity*.
The measurement thus defined of the quantity in a given statical
charge can be made by observing the swing of a galvanometer-
needle produced by allowing the charge to pass through the coil
of the galvanometer in a time extremely short compared with
that occupied by an oscillation of the needle.

* Weber calls thli quantity two units — a fact which must not be lost sight of
in oomparing his results with those of the Committee.

FOR ELECTRICAL MEASUREMENTS 109

Let Q be the whole quantity of electricity in an instantaneous
current, then

Q = 2^8inii, (12)

where Ci » the strength of a current giving a unit deflection (45''
on a tangent or 90'' on a sine galvanometer), t » half the period or
time of a complete oscillation of the needle of the galvanometer
under the influence of terrestrial magnetism alone, and % « the
angle to which the needle is observed to swing from a position
of rest when the discharge takes place; Ci is a constant which
need only be determined once for each instrument, provided the
horizontal force of the earth's magnetism remain unchanged. In
the case of the tangent galvanometer, the formula for obtaining
it has already been given. From equations (9) and (12) we havQ
for a tangent galvanometer

Q^^Htsinii, (13)

where, as before, k » the radius of the coil, and n » the number
of turns made by the wire round the coil.

Tbe quantity in a given charge which can be continually
reproduced under fixed conditions may be measured by allowing a
suoceasion of discharges to pass at regular and very short intervals
through a galvanometer, so as to produce a permanent deflection.
The value of a current producing this deflection can be ascer-
tained; and the quotient of this value by the number of discharges
taking place in a ''second" gives the value of each charge in
electro-magnetic measure.

To find the dimensions of [Q], we simply observe that the unit
of electricity is that which is transferred by the unit current in
the unit of time. Multiplying the dimensions of [C] by [T], we

find the dimensions of [Q] are [L^M*].

26. Electric Capacity of a Conductor. — It is found by experi-
ment that, other circumstances remaining the same, the charge on
an insulated conductor is simply proportional to the electromotive
force between it and the surrounding conductors, or, in other
words, to the difference of potentials (47). The charge that
would be produced by the unit electromotive force is said to
measure the electric capacity of a conductor. Thus, generally,'

110 PKACTICAL STANDARDS

the capacity of a conductor S^^, where Q is the whole quantity

in the charge produced by the electromotive force E. When the
electromotive force producing the charge is capable of maintaining
a current, the impacity of the conductor may be obtained without
a knowledge of the value either of Q or E, provided we have the
means of measuring the resistance of a circuit in electro-magnetic
measure. For let Ri be the resistance of a circuit, in which the
given electromotive force E will produce the unit deflection on a
tangent galvanometer, then, from equations (6) and (12), we have

5.2*^', (14)

where t and i retain the same signification as in equation (13)
(§ 26).

27. Direct Measurement of Electromotive Force. — ^The meaning
of the words ''electromotive force'' has already been explained
(§ 16); this force tends to do work by means of a cunt^nt or
transfer of electricity, and may therefore be said to produce and
maintain the current. In any given combination in which electric
currents flow, the immediate source of the power by which the
work is done is said to produce the electromotive force. The
sources of power producing electromotive force are various. Of
these, chemical action in the voltaic battery, unequal distribution
of temperature in cii-cuits of different conductors, the friction of
different substances, magneto-electric induction, and simple elec-
tric induction are the most &miliar. An electromotive force may
exist between two points of a conductor, or between two points of
an insulator, or between an insulator and a conductor, — in fine,
between any points whatever. This electromotive force may be
capable of maintaining a current for a long time, as in a voltaic
battery, or may instantly cease after producing a current of no
sensible duration, as when two points of the atmosphere at
different potentials (§ 47) are joined by a conductor; but in every
case in which a constant electromotive force E is maintained
between any two points, however situated, the work spent or
gained in transferring a quantity, Q, of electricity from one of
those points to the other will be constant; nor will this work be
affected by the manner or method of the transfer. If the
electricity be slowly conveyed as a static charge on an insulated

FOB ELECTRICAL lAEASUREMENTS 111

ball, the work will be spent or gained in accelerating or retarding
the ball; if the electricity be conveyed rapidly through a con-
ductor of small resistance, or more slowly through a conductor
of great resistance, the work may be spent in heating the
conductor, or it may electrolyze a solution, or be thermo-elec-
trically or mechanically used; but in all cases the change effected^
measured as equivalent to work done, will be the same, and equal
to EQ. Hence the electromotive force from the point A to the
point B is unity, if a unit of mechanical work is gained in the
transfer of a unit of electricity from A to B. This general
definition is due to Professor W. Thomson.

The direct measurement of electromotive force may be made
by the measurement in any given case of the work done by the
transfer of a given quantity of electricity. The ratio between the
numbers measuring the work done and the quantity transferred
would measure the electromotive force. This measurement has
been made by Dr Joule and Professor Thomson, by determining
the heat developed in a wire by a given current measured as
in § 18»

28. Indirect Measurements of Electromotive Force, — The
direct method of measurement is in most cases inconvenient,
and in many impossible ; but the indirect methods are numerous
and easily applied. The relation between the current, C, the
resistance, R, and the electromotive force, E, expressed by Ohm's
law (equation 6), will determine the electromotive force of a
battety whenever R and C are known. A second indirect method
depends on the measurement of the statical force with which two
bodies attract one another when the given electromotive force is
maintained between them. This method is fully treated in
Part IV (43). The phenomenon on which it is based admits of
an easy comparison between various electromotive forces by
electrometers. This method is applicable even to those cases in
which the electromotive force to be measured is incapable of
maintaining a current. The laws of chemical electrolysis and
electro-magnetic induction afford two other indirect methods of
estimating electromotive force in special cases (54 and 31 ^

29. Measurement of Electric Resistance. — We have already
stated that the resistance of a conductor is that property in
virtue of which it limits the amount of work performed by a

* PhiU Mag, vol. u. 4th ser. 1861, p. 65k

112 PRACTICAL STANDARDS

given electromotive force in a given time, and we have shown

that it may be measured by the ratio -^ of the electromotive

force between two ends of a conductor to the current maintained
by it. The unit resistance is therefore that in which the unit
electromotive force produces the unit current, and therefore
performs the unit of work in the unit of time. If in any circuit
we can measure the current and electromotive force, or even the
ratio of these magnitudes, we should, ipso fa>cio, have measured
the resistance of the circuit. The methods by which this ratio
has been measured, founded on the laws of electro-magnetic
induction, are fiilly described in Appendix D. Other methods
may be founded on the measurement of currents and electro-
motive forces described in 18, 19, 20, 27, and 28. Lastly, a
method founded on the gradual loss of charge through very great
resistances will be found in Part lY (45). The equation (25)
there given for electrostatic measure is applicable to electro-
magnetic measure when the capacity and difference of potentials
are expressed in electro-magnetic units.

SO. Electric Resistance in Electr<Mnagnetic Units is measured
by an Absolute Velocity. — The dimensions of [R] are found, by

comparing those of [E] and [C], to be lipU or those of a simple

velocity. This velocity, as was pointed out by Weber, is an
absolute velocity in nature quite independent of the magnitude
of the fundamental units in which it is expressed The following
illustration, due to Professor Thomson, will show how a velocity
may express a resistance, and also how that expression may be
independent of the magnitude of the units of time and space.

Let a wire of any material be bent into an arc of 57^'' with
any radius, k. Let this arc be placed in the magnetic meridian
of any magnetic field, with a magnet of any strength freely
suspended in the centre of the arc. Let two vertical wires or
rails, separated by a distance equal to A;, be attached by a wire
to the ends of the arc ; and let a cross piece slide along these taik
inducing a current in the arc Then it may be phown that the
speed required to produce a deflection of 45"" on the magnet will
measure the resistance of the circuit, which is assumed to be
constant. This speed will be the same whatever be the value
of £, or the intensity of the magnetic field, or the moment of

FOR ELEGTBICAL MEASUREMENTS 113

the magnet. In this form the experiment could not be easily
earried out; but if a length, I, of wire be taken and rolled into a
circular coil at the radius k, and the distance between the vertical

rails be taken equal to j , then if the resistance of the circuit

be the same as in the previous case» the deflection of 45° will be

pnoduced by the same velocity in the cross piece, measuring that

jfc"
resistance ; or, generally, if the distance between the rails be p y ,

then p times the velocity required to produce the unit deflection
(45") will measure the resistance. The truth of this proposition
can easily be established when the laws of magneto-electric induc-
tion have been understood (31).

31. MctgneUhelectric Induction. — Let a conducting circuit be
placed in a magnetic field. Let C be the intensity of any current
in that circuit ; E the magnitude of the electromotive force acting
in the circuit. Let the circuit be so moved that the number of
lines of magnetic force (11) passing through the area which it
encloses is increased by N in the time t, then (23) the electro-
magnetic forces will contribute towards the motion an amount of
work measured by CN, Now Q, the quantity of electricity which
passes, is equal to 0^; so that the work done on the current is
EQ or CEt By the principle of conservation of energy, the work
done by the electro-magnetic forces must be at the expense of
that done by the electromotive forces, or

CN+CEt = 0;

or dividing by Ct, we find that

N
^=-7; (15)

or, in other words, if the number of lines of force passing through
the area enclosed by a circuit be increased, an electromotive force
in the negative direction will act in the circuit measured by the
number of lines of force added per second.

If JJ be the resistance of the circuit, we have, by Ohm's law
(equation 6), E » CR ; and therefore

N^-Et^-RCt^-RQ] (16)

or, in other words, if the number of lines of magnetic force passing
through the area enclosed by the circuit is altered, a current will

B A. 8

114 PRACTICAL STANDARDS

be produced in the circuit in the direction opposite to that of a
current which would have produced lines of force in the direction
of those added, and the quantity of electricity which passes
multiplied by the resistance of the circuit measures the number
of additional lines passing through the area enclosed by the circuit.

The facts of magneto-electric induction were discovered by
Faraday, and described by him in the First Series of his Experi-
mental Researches in Electricity, read to the Royal Society,
24th November, 1831.

He has shown* the relation between the induced current and
the lines of force cut by the circuit treated as a sur£sice or area,
and he has also described the state of a conductor in a field of
force as a state the change of which is a cause of currents. He
calls it the electrotonic state; and, as we have just seen, the
electrotonic state may be inea^sured by the number of lines of
force which pass through the circuit at any time.

The measure of electromotive force used by W. Weber, and
derived by him (independently of the principle of conservation
of energy) from the motion of a conductor in a magnetic field, is
the same as that at which we have arrived; for, from equation
(15), we find that the unit electromotive force will be produced by
motion in a magnetic field when one line of force is added (or
subtracted) per unit of time, and this will occur when in a field
of unit intensity a straight bar of unit length, forming part of a
circuit otherwise at rest, is moved with unit velocity perpen-
dicularly to the lines of force and to its own direction.

To W. Weber, whose numerical determinations of electrical
magnitudes are the starting-point of exact science in electricity,
we owe this, the first definition of the tmit of electromotive force ;
but to Professor Helmholtzf and to Professor W. Thomson J,
working independently of each other, we owe the proof of the
necessary existence of magneto-electric induction and the deter-
mination of electromotive force on strictly mechanical principles.

32. On Material Standards /or the Measurement of Electi-ic
Magnitudes, — The comparison between two different electrical
magnitudes of the same nature, e,g. between two currents or

* Experimental Rfiearcket, 3082, etc.

t Paper read before the Physical Society of Berlin, 1847 (vide Taylor's Seientijie
Memoirs, part ii. Feb. 1858, p. 114).

X Transactions of the British Association, 1848 ; Phil. Mag., Deo. 1851.

FOR ELECTRICAL MEASUREMENTS 115

between two resistances, is in all cases much simpler than the
direct measurements of these magnitudes in terms of time, mass,
and space, as described in the foregoing pages« Much labour is,
therefore, saved by the use of standards of each magnitude ; and
the construction and distribution of those standards form part of
the duties of the Committee*

Electric currents are most simply compared by "electro-
dynamometers" (20) — instruments which, unlike galvanometers,
are practically independent of the intensity of the earth's
magnetism. When an instrument of this kind has been con-
structed, with which the values of the currents corresponding to
each deflection has been measured (19, 20), other instruments
may easily be so compared with this standard, that the relative
valae of the deflections produced by equal currents od the
standard and the copies shall be known. Hence the absolute
value of the current indicated by each deflection of each copy will
be known in absolute measure. In other words, in order to
obtain the electro-magnetic measure of a current in the system
described, each observer in possession of an electro-dynamometer
which has been compared with the standard instrument will
simply multiply by a constant number the deflection produced
by the current on his instrument (or the tangent or sine of
the deflection, according to the particular construction of the
instrument).

Electric quantities may be compared by the swing of the
needle of a galvanometer of any kind. They may be measured
by any one in possession of a standard electro-dynamometer, or
resistance-coil, since the observer will then be in a position
directly to determine (7, in equation (12), or R, in equation (14).

Capacities may be compared by the methods described (26);
and a Leyden jar or condenser (41) of unit capacity, and copies
derived from it, may be prepared and distributed. The owner of
such a condenser, if he can measure electromotive force, can
determine the quantity in his condenser.

The material standard for electromotive force derived from
electro-magnetic phenomena would naturally be a conductor of
known shape and dimensions, moving in a known manner in a
known magnetic field. Such a standard as this would be far too
complex to be practically useful: fortunately a very simple and
piuctical standard or gauge of electromotive force can be based on

8—2

116 PRACTICAL STANDARDS

its statical effects, and will be described in treating of those effects
(Part IV, 43). A practical standard for approximate measure-
ments might be formed by a voltaic couple, the constituent parts
of which were in a standard condition. It is probable that the
Daniell's cell may form a practical standard of reference in this
way, when its value in electro-magnetic measure is known. This
value (centimetre-gramme second) lies between 9 x 1(F and 11 x 1(F
(or 9 X 10* and 11 x 10* metre-gramme second). [Note, 1872. —
Mr Latimer Clark's cell equal to 1*457 x 10^ centimetre-gramme
second series, or 1*457 x 10" metre-gramme second series, is a
better standard of E.M.F. This cell is composed of pure mercury
as the negative element, the mercury being covered by a paste
made by boiling mercurous sulphate in a thoroughly saturated
solution of zinc sulphate, the positive element consisting of pure
zinc resting on the paste. This element must not be used to
produce a current, but forms an excellent standard of E.M.F., when
compared with other cells, by any method which does not involve
the passage of a current through the cell (vide Proc, Roy, Soc.
No. 136, 1872).]

Resistances are compared by comparing currents produced in
the several conductors by one and the .same electromotive force.
The unit resistance, determined as in Appendix D, will be
represented by a material conductor; simple coils of insulated wire
compared with this standard, and issued by the Committee, will
allow any observer to measure any resistance in electro-magnetic
measure.

Part IV. — Measurement of Electric Phenomena by

Statical Effects.

33. Electrostatic Measure of Electric Quantity. — By the appli-
cation of a sufficient electromotive force between two parts of a
conductor which does not form a circuit, it is possible to com-
municate to either part a charge of electricity which may be
maintained in both parts, if properly insulated (14). With the
ordinary electromotive forces due to induction or chemical action,
and the ordinary size of insulated conductors, the charge of
electricity in electro-magnetic measure is exceedingly small; but
when the capacity of the conductor is great, as in the case of long

FOR ELECTRICAL MEASUREMENTS 117

sabmarine cables, the charge may be considerable. By making
use of the electromotive force produced by the friction of unlike
substances, the charge or electrification even of small bodies may
be made to produce visible eflFects. The electricity in a charge is
not essentially in motion, as is the case with the electricity in a
current. In other words, a charge may be permanently main-
tained without the performance of work. Electricity in this
condition is therefore frequently spoken of as statical electricity,
and its effects, to distinguish them from those produced by
currents, may be called statical effects. The peculiar properties
of electrically charged bodies are these : —

1. When one body is charged positively (14), some other body
or bodies must be charged negatively to the same extent.

2. Two bodies repel one another when both are charged
positively, or both negatively, and attract when oppositely charged.

3. These forces are inversely proportional to the square of the
distance of the attracting or repelling charges of electricity.

4. If a body electrified in any given invariable manner be
placed in the neighbourhood of any number of electrified bodies,
it will experience a force which is the resultant of the forces that
would be separately exerted upon it by the different bodies if
they were placed in succession in the positions which they actually
occupy, without any alteration in their electrical conditions.

From these propositions it follows that, at a given distance,
the force, /, with which two small electrified bodies repel one
another is proportional to the product of the charges, q and f,,
upon them. But when the distance varies, this force, /, is in-
versely proportional to the square of the distance, d, between
them; hence

/-f (")

When q and ^i are of dissimilar signs, / becomes negative, ije. there
is an attraction, and not a repulsion. This equation is incompatible
with the electro-magnetic definitions given in Part III, and, if it
be allowed to be fundamental, gives a new definition of the unit
quantity of electricity, as that quantity which, if placed at a unit
distance from another equal quantity of the same kind, repels it
with unit force.

34, Electrostatic System of Units, — This new measurement of
quantity forms the foundation of a distinct system or series of

118 PRACTICAL STANDARDS

units, which may b^ called the electrostatic units, and measure-
ments in these units will in these pages be designated by the
use of small letters; thus, as Q, C, etc. signify the number
of electrostatic units in the same quantities, currents, etc. in
electro-magnetic measure, so g, c, 6, and r, etc. will represent the
electrostatic measure of quantity, current, electromotive force,
resistance, etc.

The relations between current* and quantity, between work,
current, and electromotive force, and between electromotive force,
current, and resistance, remain unchanged by the change from
the electro-magnetic to the electrostatic sjnstem.

35. Ratio between Electrostatic and Electro-nuignetic Measures
of Quantity, — Since the expression forming the second member
of equation (17) represents a force the dimensions of which are

"T* r *^^ dimensions of [q] are — ^ . The dimensions of
the unit of electricity, [Q], in the electro-magnetic system are
[£^3/2] (26). Hence, since in passing from the one system to

the other we must employ the ratio ~ , this ratio will be of the

dimension m h that is to say, it is of the nature of a velocity.

In the present treatise this velocity will be designated by the
letter v.

The first estimate of the relation between quantity of electricity
measured statically and the quantity transferred by a current in a
given time was made by Faraday*. A careful experimental in-
vestigation by MM. Weber and Kohlrauschf not only confirms
the conclusion that the two kinds of measurements are consistent,
but shows that the velocity v is 310,740,000 metres per second —
a velocity not diflfering fi-om the estimated velocity of light more
than the different determinations of the latter quantity differ
from each other, v must always be a constant real velocity in
nature, and should be measured in terms of the system of funda-
mental units adopted in electrical measurements (3 and 55). A
redetermination of v (46) will form part of the present Committee's
business in 1863-64. It will be seen that, by definition, the

* Experinwntal Retearchet, Beries iii. § 361, etc.

t Ahhandlungen der K'dnig, SHchsisehen Oes, Bd. iii. (1857) p. 260; or Poggen-
dorff^s Annalen, Bd. zoiz. p. 10 (Aug. 1856).

FOB ELECTRICAL MEASUREMENTS 119

quantity transmitted by an electro-magnetic unit current in the
unit time is equal to v electrostatic units of quantity. In the
oentimetre-gramme second series the value of v will clearly be
100 times as great as that given above.

36. Electrostatic Measure of Currents. — In any coherent
system, a current is measured by the quantity of electricity
which passes in the unit of time (15); if both current and
quantity are measured in electrostatic units, then

c = f (18)

The dimensions of [c] are therefore —mf \ and in order to

reduce a current firom electro-magnetic to electrostatic measure,
we must multiply G by v, or

c^vC. (19)

37. Eledrostaiic Measure of Electromotive Force. — The
statical measure of an electromotive force is the work which
would be done by electrical forces during the passage of a unit
of electricity from one point to another. The only difference
between this definition and the electro-magnetic definition (16
and 27) consists in the change of the unit of electricity from the
electro-magnetic to the electrostatic.

Hence if q units of electricity are transferred from one place
to another, the electromotive force between those places being e,
the work done during the transfer will be qe ; but we found (27)
that if E and Q be the electro-magnetic measures of the same
quantities, the work done would be expressed by QE\ hence

qe^QE,

but (35) q = vQ,

E
therefore e = - (20)

Thus, to reduce electromotive force fi'om electro- magnetic to
electrostatic measure, we must divide by t*.

L^M^
The dimensions of e are ,» .

120 PRACTICAL STANDARDS

38. Electrostatic Measure of Resistance. — If an electromotive
force, St act on a conductor whose resistance in electrostatic
measure is r, and produce a current, 6, then by Ohm's law

^ = 3 <">

Substituting for e and c their equivalents in electro-magnetic
measure (equations 19 and 20), we have

but(eq. 7) ^~n'

and therefore r^-R (22)

Hence, to reduce a resistance measured in electro-magnetic
units to its electrostatic value, we must divide by r*.

The dimensions of [r] are y , or the reciprocal of a velocity.

39. Electric Resistance in Electrostatic Units is measured by
the Reciprocal of an Absolute Velocity, — We have seen from the
last paragraph that the dimensions of r establish this proposition ;
but the following independent definition, due to Professor W.
Thomson, assists the mind in receiving this conception as a
necessary natural truth. Conceive a sphere of radius k, charged
with a given quantity of electricity, Q. The potential of the

sphere, when at a distance from all other bodies, will be ^ (40, 41,

and 47). Let it now be discharged through a certain resistance, r.
Then if the sphere could collapse with such a velocity that its
potential should remain constant, or, in other words, that the ratio
of the quantity on the sphere to its radius should remain constant,
during the discharge, then the time occupied by its radius in
shrinking the unit of length would measure the resistance of the
discharging conductor in electrostatic measure, or the velocity
with which its radius diminished would measure the conducting
power (50) of the discharging conductor. Thus the conducting
power of a few yards of silk in dry weather might be an inch per
second, in damp weather a yard per second The resistance of
1000 miles of pure copper wire, -j^ inch in diameter, would be

FOR ELECTRICAL MEASUREMENTS 121

about 0*00000141 of a se<K)Qd per metre, or its conducting power
<Mie metre per 0*00000141 of a second, or 709220 metres per
second.

40. ElectrosteUic Measure of the Capacity of a Conductor. —
The electrostatic capacity of a conductor is equal to the quantity
of electricity with which it can be charged by the unit electromotive
force. This definition is identical with that given of capacity
measured in electro-magnetic units (26). Let 8 be the capacity
of a conductor, q the electricity in it, and e the electromotive
ibroe charging it; then

q^se. (23)

From this equation we can see that the dimension of the quantity
« is a length only. It will also be seen that

«=t;*S, (24)

where S is the electro-magnetic measure of the capacity of the
conductor with the electrostatic capacity, «.

The capacity of a spherical conductor in an open space is, in
electrostatic measure, equal to the radius of the sphere — a fact
demonstrable from the fundamental equation (17).

Experimentally to determine 8, the capacity of the conductor
in electrostatic measure, charge it with a quantity, q, of electricity,
and measure in any unit its potential (47), e. Then bring it into
electrical connexion with another conductor whose capacity, ^i, is
known. Measure the potential, ei,ot 8 and «i, after the charge is
divided between them ; then

? = w = (« + «i)ei,

and hence *= «i (25)

e — ei ^ '

In this measurement we do not require to know e and Ci in
absolute measure, since the ratio of these two quantities only is
required. We must, however, know the value of «i; and hence we
must begin either with a spherical conductor in a large open space,
whose capacity is measured by its radius, or with some other form
of absolute condenser alluded to in the following paragraph.

41. Absolute Condenser, Practical Measurement of Quantity,
— As soon as the electromotive force of a source of electricity is
known in electrostatic measure, the quantity which it will produce
in the form of charge on simple figures is known by the laws of
electrical distribution experimentally proved by Coulomb. Such

122 PRACTICAL STANDARDS

simple figures may be termed absoluie condensers. A sphere in an
open space is such a condenser, and its capacity is numerically
equal to its radius. A more convenient form is a sphere of radius x,
suspended in the centre of a hollow sphere, radius y, the latter
being in communication with the earth. The capacity, s, of the
internal sphere is then, by calculation,

^=-^^ (26)

y-x

By a series of condensers of increasing capacity, we may
measure the capacity of any condenser, however large. The com-
parison is made by the method described above (40). Thus, the
practical method of measuring quantity in electrostatic measure is
first to determine the capacity of the conductor containing the
charge, and then to multiply that capacity by the electromotive
force producing the charge (43).

42. Practical Measurement of Currents. — The electrostatic
value of currents can be obtained from equation (21), when e and r
are known, or from equation (19), when v and C are known, or by
comparison with a succession of discharges of known quantities
from an absolute condenser.

43. Practical Measurement of Electromotive Force. — The
relations expressed by eq. (17) and (23) show that in any given
circumstances the force exerted between two bodies due to the
effects of statical electricity will be proportional to the electro-
motive force or difference of potential (47) between them. This
fact allows us to construct gauges of electromotive force, or
instruments so arranged that a given electromotive force between
two parts of the apparatus brings an index into a sighted position.
In order that the gauge should serve to meoMbre the electromotive
force absolutely, it is necessary that two things should be known :
first, the distribution of the electricity over the two attracting or
repelling masses (or, in other words, the capacity of each part) ;
secondly, the absolute force exerted between them. For simple
forms, the distribution or capacity of each part can be calculated
from the fundamental principles (33); the force actually exerted
can be weighed by a balance. By these means Professor W.
Thomson* determined the electromotive force of a Daniells cell to

» Paper read before the Royal Society, February 1860. Vide Proceedings of the
Boyal Society, vol. z. p. 819, and Phil. Mag. vol. zz. 4th ser. (1860), p. 283.

FOB ELECTRICAL MEASUREMENTS 123

be 0'0021 in British electrostatic units, or 00002951 in metrical
units, or 0KK)2951 in centimetrical units. This proposition is
equivalent to saying that two balls of a metre radius, at a distance
d apart, measured in metres, in a large open space, and in con-
nexion with the opposite poles of a DanielFs cell, would attract
one another vrith a force equal

^ 0-0002951 X 00002951 , , , ,. , .
to -Ti absolute metrical units,

0-000000008876
or -7^ gramme weight.

An apparatus by which such a measui^ement as the foregoing
can be carried out is called an absolute electrometer. It will be
observed that, although the definition of electromotive force Is
founded on the idea of work, its practical measurement is effected
by observing a force, inasmuch as when this force exerted between
two conductors of simple shape is known, the work which the
passage of a unit of electricity between them would perform may
be calculated by known laws.

44. Comparison of Electromotive Forces hy their Statical
Effects. — This comparison is simpler than the absolute measure-
ment, inasmuch as it is not necessary, in comparing two forces, to
know the absolute values of either. Instruments by which the
comparison can be made are called electrometers. Their arrange-
ment is of necessity such that the force exerted between two
given parts of the instrument shall be proportional to the
difference of potential between them*. This force may be
variable and measured by the torsion of a wire, as in Thomson's
reflecting electrometer, or it maj^ be constant, and the electro-
motive forces producing it may be compared by measuring the
distance between the two electrified bodies at which these attract
each other with that constant force. The latter arrangement is
adopted in Professor Thomson's portable electrometer, first ex-
hibited at the present meeting of the Association. The indications
of a gauge or electrometer not in itself absolute may be reduced
to absolute measurement by multiplication into a constant co-
efficient.

45. Practical Measurement of Electric Resistance. — The elec-
trostatic resistance of a conductor of great resistance (such as

* A bifilar saspension is now used (1872).

124 PRACTICAL STANDARDS

gutta-percha or india-rubber) might be directly obtained in the
following manner: — Let a body of known capacity, 8 (40), be
charged to a given potential, P (47), and let it be gradually
discharged through the conductor of great resistance, r. Let the
time, t, be noted at the end of which the potential of the body

has fallen to p. The rate of loss of electricity will then be — .

-1 t

Hence p^ Pe^fr and — = log, P . Hence

r=-^; (27)

from which equation r can be deduced, if 8, t, and the ratio — be

P
known, t can be directly observed, 8 can be measured (40), and the

P
ratio - can be measured by an electrometer (44) in constant con-
nexion with the charged body. This ratio can also be measured
by the relative discharges through a galvanometer, first, im-
mediately after the body has been charged to the potential P, and
again when, after having been recharged to the potential P, it has,
after a time f, fallen to potential p. (This latter plan has long
been practically used by Messrs Siemens, although the results
have not been expressed in absolute measure.)

Unfortunately, in those bodies, such as gutta-percha and india-
rubber, the resistance of which is sufficiently great to make t a
measurable time, the phenomenon of absorption due to continued
electrification* so complicates the experiment as to render it
practically unavailable for any exact determination. The apparent
effect of absorption is to cause r, the resistance of the material,
to be a quantity variable with the time t; and the laws of the
variation are very imperfectly known.

46. Experimental Detei^mincUion of the Ratio, v, between
Electro-magnetic and Electrostatic Mea^sures of Quantity. — In
order to obtain the value of v, it is necessary and sufficient that
we should obtain a common electrostatic and electro-magnetic
measure of some one quantity, current, resistance, electromotive
force or capacity. There are thus five known methods by which the
value can be obtained : —

* Vide TraruactioTu of British Atsociation, 1869, p. 248, and Report of the Com-
mittee of Board of Trade on Submarine Cablee, pp. 186 and 464.

FOR ELECTRICAL MEASUREMENTS 125

(1) By a common measure of quantity. Let a condenser of
known capacity, 8, be prepared (40). Let it be charged to a given
potential P (47). Then the quantity in the condenser will be
sP in electrostatic measure. The charge can next be measured
by discharge through a galvanometer (25) in electro-magnetic
measure. The ratio between the two numbers will give the value
of V. The only difficulty in this method consists in the measure-
ment of the potential P, entailing the measurement of an absolute
force between two electrified bodies. This method was proposed
and adopted by Weber*.

(2) By a comparison of the measure of electromotive force.
The electromotive force produced by a battery, in electrostatic
measure, can be directly weighed (43). Its electromotive force,
in electro-magnetic measure, can be obtained from the current it
produces in a given resistance (28). The ratio of the two numbers
will give the value of v. The method has been carried out by
Professor W. Thomson, who was not, however, at the time in pos-
session of the means of determining accurately either the absolute
resistance of his circuit or the absolute value of the current t.

(3) By a common measure of resistance. We know (29 and
45) how to measure resistances in electro-magnetic and electro-
static measure. The ratio between these measures is equal to t;*.
The measure of resistance in electrostatic measure is not as yet
susceptible of great accuracy.

(4) By a comparison of currents. The electro-magnetic value
of a current produced by a rapid succession of discharges from a
condenser of capacity 8 can be measured (18, 19). The electro-
static value of the current will be known if the potential to which
the condenser is charged be known. The ratio of the two numbers
is equal to v.

(5) By a common measure of capacity. The two measure-
ments can be effected by the methods given (26 and 40). The
ratio between the two measurements will give t;". This method
would probably yield very accurate results.

* Pogg. Ann., Aag. 1856, Bd. xoiz. p. 10. Abhandlungen der Kdn, Sachsischen
GfelUchaft, Bd. iii. (L857) p. 266.

t Paper read before the Royal Societji Febrnary 1860. Vide Proceedings of the
Royal Society, vol. z. p. 319.

I 126 PRACTICAL STANDARDS

Part V. — Electrical Measurements derived from the Five
Elementary Measurements; and Conclusion.

47. Electrical Potential. — The word "potential," as applied
by G. Green, to the condition of an electrified body and the space
surrounding it, is now coming into extensive use, but is perhaps
less generally understood than any other electrical term. Electric
potential is defined by Prof. W. Thomson as follows*: —

" The potential, at any point in the neighbourhood of or within
an electrified body, is the quantity of work that would be required
to bring a unit of positive electricity fi-om an infinite distance to
that point, if the given distribution of electricity remained un-
altered."

It will be observed that this definition is exactly analogous to
that given of magnetic potential (10), with the substitution of the
unit quantity of electricity for the unit magnetic pole. (Analogous
definitions might be given of gravitation potential, heat potential ;
and every one of these potentials coexist at every point of space
quite independently one of the other.) In another paperf Professor
Thomson describes electric potential as follows : — " The amount of
work required to move a unit of electricity against electric repulsion
from any one position to any other position is equal to the excess
of the electric potential of the first position above the electric
potential of the second position."

The two definitions given are virtually identical, since the
potential at every point of infinity is zero; and it will be seen
that the difference of potential defined in the second passage
quoted is identical with what we have called the electromotive
force between the two points (16 and 27).

When, instead of a difference of potentials, the potential simply
of a point is spoken of, the difference of potential between the
point and the earth is referred to, or, as we might say, the electro-
motive force between the point and the earth.

The potential at all points close to the surface and in the
interior of any simple metallic body is constant ; that is to say, no
electromotive force can be produced in a simple metallic body by

* Paper read before the British Association, 1S52. Vide Phil, Mag. 1853,
p. 288.

t Paper read before the Boyal Society, February 1860. Vide Proeeedingt of
the Royal Society, vol. z. p. 334.

FOR ELECTRICAL MEASUREMENTS 127

mere electrical distribution ; the potential at the body may there-
fore be called the potential of the body. The potential of a
metallic body varies according to the distribution, dimensions,
poeition, and electrification of all surrounding bodies. It also
depends on the substance forming the dielectric.

In any given circumstances, the potential of the body will be
simply proportional to the quantity of electricity with which it is
charged; but if the circumstances are altered, the potential will
vary although the total amount of the charge may remain
constant.

In a closed circuit in which a current circulates, the potential
of all parts of the circuit is different ; the difference depends on
the resistance of each part and on the electromotive force of the
source of electricity, %.e. on the difference of potentials which it is
capable of causing when its two electrodes are separated by an
insulator or dielectric. The different parts of a conductor moving
in a magnetic field are maintained at different potentials, inas-
much as we have shown that an electromotive force is produced in
this case. The potential of a body moving in an electric field
(t.e. in the neighbourhood of electrified bodies) is constantly
changing, but at any given moment the potential of all the parts
is equal The use of the word "potential" has the following
advantages: it enables us to be more concise than if we were
continually obliged to use the circumlocution, " electromotive force
between the point and the earth " ; and it avoids the conception
of a force capable of generating a current, which almost necessarily,
although falsely, is attached to '' electromotive force."

Equipotential surfaces and lines of force in an electric field
may be constructed for statically electrified bodies ; these surfaces
and lines may be drawn on similar principles and possess analogous
properties to those described in a magnetic field (10). It is hardly
necessary to observe that the magnetic and the electric fields are
totally distinct, and coexist without producing any mutual infiuence
or interference.

The rate of variation of electric potential per unit of length
along a line of force is at any point equal to the electrostatic force
at that point, %.e. to the force which a unit of electricity placed
there would experience. The unit difference of potential is identical
with the unit electromotive force ; and the electrometer spoken of
as measuring electromotive force measures potentials or differences
of potential

128 PRACTICAL STANDARDS

48. Density, Resultant Electric Force, Electric Pressure. — The
three following definitions are taken almost literally from a paper
by Professor W. Thomson*. Our treatise would be incomplete
without reference to these terms, and Professor Thomson's defini*
tions can hardly be improved.

" Electric Density. — This term was introduced by Coulomb to
deilignate the quantity of electricity per unit of area in any part
of the surface of a conductor. He showed how to measure it,
though not in absolute measure, by his proof-plane.

" Resultant Electric Force, — The resultant force in air or other
insulating fluid in the neighbourhood of an electrified body is the
force which a unit of electricity concentrated at that point would
experience if it exercised no influence on the electric distributions
in its neighbourhood. The resultant force at any point in the air
close to the surface of a conductor is perpendicular to the sur&ce,
and equal to iiirp, if p designates the electric density of the surface
in the neighbourhood.

** Electric Pressure from the Surface of a Conductor balanced
by Air, — A thin metallic shell or liquid film (as, for instance, a
soap-bubble), if electrified, experiences a real mechanical force in
a direction perpendicular to the surface outwards, equal in amount
per unit of area to 27r/>*, p denoting, as before, the electric density
at the part of the surface considered. In the case of a soap-bubble
its effect will be to cause a slight enlargement of the bubble on
electrification with either vitreous or resinous electricity, and a
corresponding collapse on being perfectly discharged. In every
case we may consider it as constituting a deduction from the
amount of air-pressure which the body experiences when un-
electrified. The amount of deduction being different in different
parts according to the square of the electric density, its resultant
action on the whole body disturbs its equilibrium, and constitutes
in fact the resultant electric force experienced by the body.*'

49. Teimon. — The use of this word has been intentionally
avoided by us in this treatise, because the term has been some-
what loosely used by various writers, sometimes apparently ex-
pressing what we have called the density, and at others diminution
of air-pressure. By some writers it has been used in the sense of a
magnitude proportional to potential or difference of potentials,
but without the conception of absolute measurement, or without

• Paper read before the Boyal Society, Feb. 1860. Vide Proc, Roy. Soe. vol. x.
p. B19 (1860), and Phil. Mag, vol. xz. ser. 4 (1860), p. 322.

FOR ELECTRICAL MEASUREMENTS 129

reference to the idea of work essential in the conception of
potential.

50. Conducting Power, Specific Resistance, and Specific Con^
ducting Power,

Conducting Power, or Conductivity. — These expressions are
employed to signify the reciprocal of the resistance of any con-
ductor. Thus, if the resistance of a wire be expressed by the
number 2, its conducting power will be 0*5.

Specific Resista/nce referred to unit of Mass. — ^The specific
resistance of a material at a given temperature may be defined as
the resistance of the unit mass formed into a conductor of unit
length and of uniform section. Thus the specific resistance of sk
metal in the metrical system is the resistance of a wire of that
metal one metre long and weighing one gramme. If the centi-
metre is used as the fundamental unit, the specific resistance of a
metal is the resistance of a wire of that metal one centimetre long
and weighing one gramme.

The Specifi>c Conducting Power of a material is the reciprocal of
its specific resistance.

Specific resistance, referred to unit of volume, is the resistance
opposed by the unit cube of the material to the passage of
electricity between two opposed &ces. It may easily be deduced
fiom the specific resistance referred to unit of mass, when the
specific gravity of the material is known.

Specific Conducting Power may also be referred to unit of
volume. It is of course the reciprocal of the specific resistance
referred to the same unit.

It is somewhat more convenient to refer the resistance to the
unit of mass in the case of long uniform conductors, such as metal
wires, of which the size is fi:equently and easily measured by the
weight per foot or metre or centimetre ; and it is, on the other
hand, more convenient to refer to the unit of volume bodies, such
as gutta-percha, glass, etc., which do not generally occur as con-
ducting-rods of uniform section, while their dimensions can always
be measured with at least as much accuracy as their weights.

61. Specific Inductive Capacity*. — Faraday discovered that
the capacity of a conductor does not depend simply on its dimen-
sions or on its position relatively to other conductors, but is
influenced in amount by the nature of the insulator or dielectric

* Experimental lUuarehes, serieB xi.
B A. 9

130 PRACTICAL STANDARDS

separating it from them. The laws of induction are assumed to
be the same in all insulating materials, although the amount
be different. The name '' inductive capacity" is given to that
quality of an insulator by virtue of which it affects the capacity of
the conductor it surrounds; and this quality is measured by
reference to air, which is assumed to possess the unit inductive
capacity. The specific inductive capacity of a material is therefore
equal to the quotient of the capacity of any conductor insulated by
that material from the surrounded conductors, divided by the
capacity of the same conductor in the same position separated
from them by air only. It is not improbable that this view of
induction may be hereafter modified.

52. Heat produced in a Conductor by a Current — ^The work
done in driving a current, C, for a unit of time through a conductor
whose resistance is iJ, by an electromotive force E, is EC = RC*
(§17). This work is lost as electrical energy, and is transformed
into heat. As Dr Joule has ascertained the quantity of mechanical
work equivalent to one unit of heat, we can calculate the quantity '
of heat produced in a conductor in a given time, if we know C and
R in absolute measure. In the series of units founded on the
centimetre, gramme, and second, if we call the total heat B, taking
as unit the quantity required to raise one gramme of water one
degree Centigrade, we have

_ RGH
^ " 4157 X 10* ^^^>

If the metre is used instead of the centimetre the divisor is
4157 ; and in the British system, founded on feet, grains, and
seconds, with a unit of heat equal to the quantity required to raise
one grain one degree Fahrenheit, the divisor is 24*861.

53. Electro-chemical Equivalents. — Dr Faraday has shown*
that when an electric current passes through certain substances
and decomposes them, the quantity of each substance decomposed
is proportional to the quantity of electricity which passes. Hence
we may call that quantity of a substance which is decomposed by
unit current in unit time the electro-chemical equivalent of that
substance.

This equivalent is a certain number of grammes of the sub-
stance. The equivalents of different substances are in the

* Experimental Researches, seriea vii.

FOR ELECTRICAL MEASUREMENTS 131

proportion of their combining numbers; and if all chemical
oompoonds were electrolytes, we should be able to construct
experimentally a table of equivalents in which the weight of each
substance decomposed by a unit of electricity would be given. The
electro-chemical equivalent of water, in electro-magnetic measure,
is about 0'02 in the British, 000092* in the centimetrical system,
and 0^92 in the metrical system. The electro^chemical equiva«
lents of all other electrolytes can be deduced from this measurement
with the aid of their combining numbers.

54. Electromotive Force of Chemical Affinity, — When two
substances having a tendency to combine are brought together
and enter into combination, they enter into a new state, in which
the intrinsic energy of the system is generally less than it was
before, that is, the substances are less able to effect chemical
changes, or to produce heat or mechanical action, than before.

The energy thus lost appears during the combination as heat
or electrical or mechanical action, and can be measured in many
casesf.

The energy given out during the combination of two substances
may, like all other forms of energy, be considered as the product
of two factorsj — the tendency to combine, and the amount of
combination effected. Now the amount of combination may be
measured by the number of electro-chemical equivalents which
enter into combination; so that the tendency to combine may also
be ascertained by dividing the energy given out by the number of
electro-chemical equivalents which enter into combination.

If the whole energy appears in the form of electric currents,
the energy of the current is measured by the product of the
electromotive force and the quantity of electricity which passes.
Now the quantity of electricity which passes is equal to the
number of electro-chemical equivalents which enter on either side
into combination. Hence the total energy given out, divided by
this number, will give the electromotive force of combination.
Thus, if N electro-chemical equivalents enter into combination

* -0009375 by Weber and Kohlraiiach.

t Report Britiih AMsoeiation, 1850, p. 63, and Phil, Mag, vol. zzzn. ser. 3. See
papers by Prof. Andrews, and Favre and Silbermann, ** On the Heat given oat in
Chamieal Action,*' CcmpU$ Aemittf , toIs. xzzti. and zxzyii.

X See Banldney *'Oa the Oenend Law of Transformation of Energy," PkiL
Mag. 18581

9—2

182 PBAGTICAL STANDARDS

under a chemical affinity /, and in doing so give out energy equal
to Wy either as heat or as electrical action, then

But if TTbe given out as electrical action, and causes a quantity
of electricity Q to traverse a conductor under an electromotive
force E, we shall have

By the definition of electro-chemical equivalents, Q^N, there-
fore

I = E;

or the force of chemical affinity may in these cases be measiured
as electromotive force.

This method of ascertaining the electromotive force due to
chemical combination, which gives us a clear insight into the
meaning and the measurement of '' chemical affinity," is due to
Professor W. Thomson*.

The field of investigation presented to us by these considera-
tions is very wide. We have to measure the intrinsic energy of
substances as dependent on volume, temperature, and state of
combination. When this is done, the energy due to any combina-
tion will be found by subtracting the energy of the compound
from that of the components before combination.

As the tendency to increase in volume is measured as pressure,
and as the tendency to part with heat is measured by the tem-
perature, so in chemical dynamics the tendency to combine will be
properly measured by the electromotive force of combination.

55. Tables of Dimermons and other Constants^: —

Fundamental Units.
Length = i. Time^T. Mass = if.

Derived Mechanical Units.
Work = TT = ^. Force = F^ ^. Velocity - F= ^.

* *« On the Mechanical Theory of Blectrolysis," Phil Mag. Deo. 1S51. ,.
t The first Tahle of DimepBions was gi^n hj Fourier, TkSorU de la ChalevTt
p. 157.

FOR ELECTRICAL MEASURElfENTS

133

Derived Magnetical Units.

Strength of the pole of a magnet. ,. m = L^ T"^ M^

Moment of a magnet mZ = Z^ T"^ M^

Intensity of magnetic field H—L"^ T"^ M^

Table of Dimensions,

Name of Quantity

Electrostatic system

Electro-magnetic
system

Number of
electrostatic
units in one
electro-mag-
netic unit

Sym-
bol

Dimensions
of unit

Sym-
bol

Dimensions
of unit

EUdrogtatie Pair
Quaotity of electricity

Electromotive force

■ • *

*
• ■ •

• .•
C
•«.

[xijrir-i]

\J?M 7-2]

identical

Their product : electrostatic
enerKT

Ratio of the first to the second :
capacity of an accumulator ...

Eleetro-mcLffnetie Pair

Electro-magnetic momentum of
a circuit, also strength of
magnetic pole

Strength of electric current, also
magnetic potential

Th^ product, electrokinetic
enerKv ..^...^<>

•

m
c
...

...

'"O^ ..*..•*-.

Batio of the first to the second :
coefficient of electro-magnetic
induction of two circuits

Pair for Conduction and Renstance

Electromotive force

e
e

•••

E
C

•••

U2jf T-»]

identical
t,-2

Steength of electric current

Tbdr product: rate at which
eoeigy is tnmsformed into heat

Batio of the first to the second :
reeistance of a conductor

134

PRACTICAL STANDARDS

All men of science are agreed to use the second of mean solar
time as the unit of time. In all the primary quantities the
dimensions of M are the same, namely ^. The principal differences,
therefore, are in the dimensions of L.

We therefore arrange the different quantities in groups, first,
with respect to the dimensions of Z, and then with respect to M
and T, thus : —

Electrostatic system

Electro- magnetic system , DimeDsion-i ' Group

Quantity of electricity ...
Strength of current

Strength of magnetic pole

Electromotive force

Magnetic intensity

Electric force and electric
induction

Density of electric current
Magnetic induction

Moment of a magnet £JM-T I I.

Strength of magnetic pole
Electromotive force ,

Quantity of electricity ... ' L^¥^
Strength of current I L^M^T'^

Electric force at a point...

Magnetic force and mag-
netic induction

Zij/ir

-2

Electric induction ••

Density of electric current

II.

III.

^~ I IV.

-.1} ^-

The Electrostatic and Electro-magnetic System of Units.

The electrostatic system begins with the definition of the unit
of electricity, as determined by the mechanical force between two
electrified bodies.

The electro-magnetic system begins with the definition of the
strength of a unit magnetic pole, as determined by the mechanical
force between two poles.

The form of the definition is precisely the same in both cases*
Hence the electrostatic unit of electricity is of the same dimensions
as the electro-magnetic unit magnetic pole, and the series of
derived units of the one system form a series having respectively
the same dimensions as another series belonging to the other
system.

FOR ELECTRICAL MEASUREMENTS

135

The most instructive method of exhibiting the relations of
these qu€uitities is to arrange them in pairs, the product of each
pair being either a quantity of mechanical energy, or the work
done in unit of time, or energy existing in unit of volume, or
work done in unit of volume in unit of time. The ratio of the
two quantities is in several cases a quantity of importance in
electrical science.

Let V be the ratio of the electro-magnetic to the electrostatic
unit of quantity (35 and 46) ; then v = 310,740,000 metres per
second approximately, and we have

q=^vQ

c^vC

8 = V^S

Table for the Conversion of British {foot-grain-second) System to
Centtmetrical {centimetre-gramme'Second) System.

(1) for M

(2)
(3)

for A ^, /2, - and V
1 r

for F (alao for foot-grains and
centimetre-grammes)

(4) for W

(5) for ff and electro-chemical
equivalents

(6) for Q, (7, and e

(7) for ^, m, 9, and c

(8) for heat

Number of

oeDtimetrical anits

contained in a

Britiiih unit

0-0647989
30-47945

1-97504
60-198

•0461085
1-40536
42-8346
0-0359994

Number of

British units

contained in a

centimetrioal unit

15-43235
•03280899

•506320
•01661185

21-6880
•711561
•0233456

27-7782

1 volt = 10" absolute units of electromotive force.

1 ohm = 10* centimetres per second.

=3-2809 X lO' feet per second.

= 1 quadrant of the meridian through Paris per

second.

= 3 r074 ohms by Weber and Kohlrausch.

= 28-2 ohms by Thomson.

= 28-8 ohms by Maxwell.

Velocity of light = 298 ohms by Foucault,

136 PRACTICAL STANDARDS

The intensity of gravity at many different stations has been
determined by experiment. Where it has not been so determined,
it may be calculated by the formula

g^G{l- 00025659 cos 2\) |l - (2 - |^')4 ,

where g denotes the intensity of gravity at the station.

0 the intensity of gravity at latitude 45^ at the level of the
sea.

0 = 980-533 centimetres, or 32*1703 feet.

X is the latitude of the station.

The last factor is a correction for the height of the station.

z is the height of the station in centimetres or feet.

r is the mean radius of the earth.

r = 636,619,800 centimetres, or 20,886,852 feet ; p is the mean
density of the earth, about 5*5 times that of water; p is the
mean density of the hill on which the station is placed. If we
suppose this about half the density of the earth as a whole, the
&ctor for correction due to height becomes

1- 1-32 -, nearly.

British System. — Relation between Absolute and other Units.

One absolute unit of I ^°"? = 0-0310666 ^'^^^ *1^ * «^ I in

[ work foot-grams j

London.

In London j'^^'^J^''^ ^.^'^ = 32-1889 absolute units of

(one foot-gram

f force.

( work.

^ I. 1 X -^ i» f force 1 ( unit weight
One absolute unit of-/. , =-^ .. ,_^ -a. i _^i_

( work g [ unit weight x unit length

everywhere.

g in British system = 32-088 (1 + 0-005133 sin« X), where
\ = the latitude of the place at which the observation is made.

Heat — The unit of heat is the quantity required to raise the
temperature of one grain of water at its maximum density l""
Fahrenheit.

Absolute mechanical equivalent of unit of heat = 24861 « 772
foot-grains at Manchester.

FOR BLECTRICAL MEASUREMENTS 137

Thermal equivalent of an absolute unit of work = 0000040224.
Thermal equivalent of a foot-grain at Manchester = 00012953.
Electro-chemical equivalent of water = 0'02, nearly.

CerUimetrical System. — Relation betvteen Absolute and other Units.

One absolute unit of I ^'"'f, = 00010195 ^"•?* f * «""""" }

( work centimetre-gramme )

at Paris.

At Paris j*^«*^;8^^*°f*«™°'™« = 980-868 absolute units of
( or centimetre-gramme

( force,
(work

^n. 1 1 ^ -^ i. ( force 1 unit weight
One absolute unit of-^ i =- -^ • i_^ -x i _xi- •

( work g umt weight x unit length ^

everjrwhere.

g in metrical system = 978024 (1 + 0005133 sin« X), where
X =s the latitude of the place where the experiment is made.

HeaJt. — ^The unit of heat is the quantity required to raise one
gramme of water at its maximum density l"" Centigrade.

Absolute mechanical equivalent of the unit of heat

» 4157*25 X 10^ = 42354*2 centimetre-grammes at Manchester

Thermal equivalent of an absolute unit of work

= 000024054 X 10"*.

Thermal equivalent of a centimetre-gramme at Manchester

= 00000236154.

Electro-chemical equivalent of water = 0*00092, nearly.
1 horse-power = 33,000 foot-pounds per minute.

s 14*732 foot-tons per minute.
= 456,233,300 centimetre-gramme weight per
minute.
„ = 7,603,388*8 centimetre-gramme weight per

second.
= 7,462,455,683 absolute units of work per

second.
= 746 X lO' absolute units of work per second
approximately.

Electromotive force of one Daniells cell, as estimated by
Thomson in 1851,

= 107 X 10* absolute units.
- 1*07 volt.

138

PRACTICAL STANDARDS

And 1 volt through 1 ohm decomposes

10*
000092 X j^ = -000092 gramme

of water per second, and hence decomposes

0000092 X j| = 000332 gramme

of zinc per second = ^j}^ gramme per second very nearly, = 28*8
grammes per day approximately.

Activity = rate of doing work = -p- for a galvanic element.

= 10' for 1 volt through 1 ohm.

Or 1 volt-ohm uses ■gj}^^^ gramme of zinc, and does 10' absolute
units of work per second.

1 horse-power = 746 volt-ohms, and is equivalent to the con-
sumption of -^^jj grammes of zinc per second in a Daniells battery,
or 8952 grammes per hour, or 21^ kilogrammes per day, very
nearly.

Table for the Conversion of British (foot-grain-second) System to

Metrical {metre-grainme-second) System.

Namber of
metrical units
contained in a

British anit

(l)or M. 0-0647989

(2)forZ,p R, ^and V | 0-3047945

(3) for F (also for foot*grains i

and metre-grammes) ' 0*01 97504

(4) for W 0-0060198

(5) for JJand electro-chemical
equivalents 0-461085

(6) for Q, C, andfl ' 0140536

(7) for B,m,g,ajidc 0-0428346

(8) for heat 00359994

Log.

2-8116678
T-4840071

2-2955749
3-7795820

1-6637804
11477874
26317949
2-5562953

1-1884321
0-5159929

1-7044250
2-2204179

0-3362196
0*8522125
1-3682051
1-4437046

Number of
British units
contained in a
metrical unit i

15-43235
3-280899

50-6320
1661186

2-16880
711661
23-3456
27-7782

Metrical System, — Relation between Absolute and other Units.

One absolute unit of I ^""^^ = 010195 ''j^' ""^ * ^^"^"^^ \ at

(work metre-gramme j

Paris.

FOR SLECTBICAL KEASURBHENTS 139

At Paris ( *^^ ^f^^^ **^ * «™™""^ = 9-80868 absolute units of
( or metre-gramme

j force.

I wort

r^ 11. •. i. f force 1 unit weight )

One absolute umt of -^ i =- -x • i.^ -^ i ^l f

\ work g unit weight x umt length j

everywhere.

g in metrical system = 9'78024 (1 +0005133 sin^X), where
X = the latitude of the place where the experiment is made.

HectL — The unit of heat is the quantity required to raise one
gramme of water at its maximum density V Centigrade.

Absolute mechanical equivalent of the unit of heat

= 4157*25 = 423*542 metre-grammes at Manchester.

Thermal equivalent of an absolute unit of work = 0*00024054.
Thermal equivalent of a metre-gramme at Manchester

= 0*00236154.
EUectro-chemical equivalent of water == 00092 nearly.

56. Magnitude of Units and Nomenclature. — In connexion
with the system of measurement explained in this treatise, two
points hitherto unmentioned deserve attention — ^first, the absolute
magnitude of the units, and, secondly, the nomenclature.

The absolute magnitude is in most cases an inconvenient one,
leading to the use either of exceedingly small or exceedingly large
numbers. Thus the units of electro-magnetic resistance and
electromotive force and quantity, and of electrostatic currents, are
inconveniently small ; the unit of electrostatic resistance is incon-
veniently large. Decimal multiples and submultiples of these
units will therefore probably have to be adopted in practice. The
choice of these multiples and submultiples forms part of the
business of the Committee.

The nomenclature hitherto adopted is extremely defective*
In referring to each measurement, we have to say that the number
expresses the value in electrostatic or electro-magnetic absolute
units : if a multiple is to be used, this multiple will also have to
be named ; and further, the standard units of length, mass^ and
time have to be referred to, inasmuch as some writers use the
pound and some the grain, some the metre and some the milli-
metre, as fundamental units. This cumbrous diction, and the risk
of error imported by it, would be avoided if each unit received a

140 PRACTICAL STANDARDS

short distinctive name in the manner proposed by Sir Charles
Bright and Mr Latimer Clark, in a paper read before the British
Association at Manchester, 1861.

Appendix D. — Description of an Experimental Measurement of
Electrical Resistance, made at King's College. By Professor
J. Clerk Maxwell and Messrs Balfour Stewart and
Fleeming Jenkin. (Parts I, III, and IV, by Pi-ofessor
Maxwell. Part II, by Mr Fleeming Jenkin.)

Part I. — General Description of the Method Employed.

In the general Report of the Committee, and in Appendix C,
it has already been shown that the most important aid to the
exact science of electricity would be the determination of the
resistance of a wire in absolute measure, and the duplication of
standards of resistance derived from this wire. This has already
been done by Weber*; but it is desirable that the determination
of a quantity so important should not be left in the hands of a
single person.

Weber has employed two methods.

1st. By suddenly turning a coil of wire about an axis so as to
alter its position relatively to the terrestrial magnetic lines of
force, he produced an electromotive force acting for a short time
in the coil. This coil was connected with another fixed coil having
a magnet suspended in its centre. The current generated by the
electromotive force' passed through both coils and gave the magnet
a sudden impulse, the amount of which was measured by its
extreme deflection.

Thus an electromotive force of short duration produced a
current of short duration. The total amount of electromotive
force depended on the size of the movable coil and on the intensity
of terrestrial magnetism. The total amount of the current is
measured by the impulse given to the magnet, and the mechanical
value of the impulse is measured by the angle through which it
swings. The resistance of the whole circuit, consisting of both
coils, is then ascertained by dividing the electromotive force by
the current.

* Pogg. Ann. Bd. Izzxii. p. 337 (March 1851) ; ElectrUcke Matubestimmungen,
Leipzig, Wiedemann ; Memoirs of the Royal Society of Sciencei of Saxony, vol. i.
p. 197 ; and Phil. Mag. 1861.

FOR ELECTRICAL MEASUREMENTS 141

2ckL Weber's second method consisted in causing a powerful
magnet to oscillate within a coil of wire. By the motion of the
magnet currents are produced in the coil, and these, reacting on
the magnet, retard its motion. The rate of diminution of the
amplitude of the oscillations, when compared with the rate of
diminution when the circuit is broken, affords the means of deter-
mining the resistance of the circuit.

Professor W. Thomson has designed an apparatus by which
the resistance of a coil can be determined in electro-magnetic
measure by the observation of the constant deflection of a magnet,
and his arrangement has been adopted for the experiments made
by the Committee.

The coil of wire is made to revolve about a vertical diameter
with constant velocity. The. motion of the coil among the lines
of force due to the earth's magnetism produces induced currents
in the coil which are alternately in opposite directions with respect
to the coil itself, the direction changing as the plane of the coil
passes through the east and west direction. If we consider the
direction of the current with respect to a fixed line in the east
and west direction, we shall find that the changes in the current
are accompanied with changes in the &ce of the coif presented to
the east, so that the absolute direction of the current, as seen from
the east, remains always the same. If a magnet be suspended in
the centre of the coil, it will be deflected from the north and
south line by the action of these currents, and will be turned in
the same direction as the coil revolves. The force producing this
deflection is continually varying in magnitude and direction ; but
as the periodic time is small, the oscillations of the magnet may
be rendered insensible by increasing the mass of the apparatus
along with which it is suspended. The resistance of the coil may
be found when we know the dimensions of the coil, the velocity
of rotation, and the deflection of the magnet. The intensity of
terrestrial magnetism enters into the measurement of the electro-
motive force, and also into the measurement of the current ; but
the measure of the resistance, which is the ratio of these two
quantities, is quite independent of the value of the magnetic
intensity.

Part IL — Description of the Apparatus.

For convenience of description, the apparatus with which the
experiments were made may be divided into five parts : — (1) the

142 PRACTICAL STANDARDS HITDUCI

driving gear ; (2) the revolving coil ; (3) the governor ; (4) the 8cal<
with its telescope, by which the deflections of the magnet wen ^*^^^"
observed ; (5) the electric balance, by which the resistance of thi
copper coil was compared with a German-silver arbitrary standard

The general arrangement of the first four parts is shown in thi
diagram, fig. 4, Plate 2.

The driving gear consisted of a leaden fly-wheel X, on a shaft j*^
A, turned by hand, and communicating its motion by a band|
66162..., arranged in a way equivalent to Huyghens's gearing, to a
shaft B, a pulley on which drove the revolving coil by a simple/
band aaiO^.... The arrangement of the band 66,61. .. communis /
eating the motion of shafb A to shaft B may be easily understood
from the diagram. CO are two guide-pullejrs running loose on
pins attached to the main framing. DL are two loose pulleys
maintained at a constant distance by the strut E, to which the
weight W is hung.

When the rotation of shafb B is opposed by a sufficient resist-
ance, the effect of turning the fly-wheel in the direction shown by
the arrow is to lift the weight W from the ground, tending to
turn the shaft JB with a definite force, which will be sensibly
constant so long as the weight is kept off the ground and the
band 66163... remains unaltered in length. Wherever, as in the
present experiments, the resistance increases with the speed of
rotation, the speed of the driving-wheel can easily be regulated by
hand, so as to keep the weight from falling so low as to touch the
ground, or rising so high as to foul the gear; and thus, with a little
care, a constant driving force can be applied to the shaft B and to
the machinery connected with it.

The revolving coil formed the most important part of the
apparatus. It is shown one-fifth full size in figs. 1 and 2, PI. 2.

A strong brass frame, HH, was bolted down by three brass
bolts, F F Fy do welled into a heavy stone. It could be accurately
levelled by three stout screws, OOG. The brass rings, //', on
which the insulated copper wire was coiled, were supported on the
frame by a pivot, «/, working in lignum vitse, and by a hollow
bearing, K, working in brass : this bearing worked in a kind of
stuffing-box, k (fig. 3), which, by three screws and a flat spring
washer between it and the firame at J, could be adjusted to fit the
collar e with great nicety, preventing all tendency to bind or shake.
Supported in this way the coil revolved with the utmost freedom
and steadiness.

lirDUCTIOjr APPARATUS,

■rr tfie. dturm iti^tum of EUeuiiob Saatanct, in
ABSOLUTE ELECTRO XAOHETIC UNITS

if.

•^JUj-tt

(

• «»

\ >

v." ■*

.>-■

)

n..

/ \

-->!

'-^^.;

FOR ELECTRICAL MEASUREMENTS 143

The coil of copper wire was necessarily divided into two parts
on the two rings //', to permit the suspension of the magnet S.
The two brass rings were each formed of two distinct halves,
insulated fix)m one €uiother by vulcanite at the flanges//'. This
insulation was necessary to prevent the induction of currents in
the brass rings. These rings, after being bolted together, were
turned with great accuracy by Messrs Elliott Brothers. The
insulated copper wire was wound in one direction on both I'ings ;
the inner end of the second was soldered to the outer end of the
first ; the two extreme ends of the conductor thus formed were
soldered to two copper terminals, h k\ insulated by the vulcanite
piece, X, bolted to the brass rings. Each terminal was provided
with a strong copper binding-screw, and had a mercury-cup drilled
into its upper surface. The two coils could be joined, so as to
form a closed circuit, by a short copper bar between the binding-
screws. The bars, binding-screws, and nuts were amalgamated
to ensure perfect contact. When the copper coils were to be con-
nected with the electric balance, the short copper bar was removed
and the required connexions were made by short copper rods, one-
quarter of an inch in diameter, dipping at one end into the
mercury-cup on the terminals hh\ and at the other end into the
mercury-cups of the electric balance. The absence of all induced
currents influencing the suspended magnet when the circuit was
broken at hh' was repeatedly proved by experiment.

Rotation was communicated to the coils by a catgut band
simply making half a turn round the small V-puUey L The band
could be tightened as required by the jockey-pulley z and weight
w (fig. 4).

A short screw of large diameter, n, gearing into a spur-wheel
of one hundred teeth, o, formed the counter from which the speed
of rotation was obtained, as follows: — A pin, p, on the wheel o
liflted the spring 9 as it passed ; this spring in its rebound struck
the gong M. The blow was of course repeated at every hundred
revolutions, and the time of each blow was observed on a chrono-
meter. The arrangement was equally adapted for rotation in
either direction.

A second V-puUey, r, served for the band c 0, communicating
motion to the governor by which the speed was controlled.

The manner in which the suspended magnet was introduced to
the centre of the coil ia best seen in fig. 3» A brass tripod, N,

144 PRACTICAL STANDARDS .

bolted to the main frame, supported the long brass tube 0, which
passed freely through the hollow bearing at K. A cylindrical
wooden box, P, slipped on to the end of the tube 0. The magnet
hung inside this box, the lower part of which could be removed
to allow the exact position of the magnet to be verified. The
support N also carried a short brass tube 22, on which the glass
case T could be secured by a little sliding tube. The mirror ty
attached to the magnet /S by a rigid brass wire, hung inside this
glass case by a single cocoon-fibre about eight feet long. This
fibre was protected against currents of air by a wooden case (not
shown in the Plate), extending from the point of support down to
the glass case. A little sliding paper prolongation of the wooden
case made it nearly wind-proof by fitting at the bottom against
the main brass fiume. An opening in the case allowed the mirror
to be seen. The fibre at the top was suspended from a torsion-
head, by which it could be turned; it could also be raised and
lowered by a small barrel, and was adjustable in a horizontal plane
by three set screws. The care taken in suspending the magnet
and in protecting it both against currents of air and vibration was
repaid by success, for the image of the scale reflected in the magnet
was as clear and steady when the coil was making 400 revolutions
per minute as when it was at rest.

The governor used was lent by one of the Committee and will
not be described in detail, as an improved governor on the same
principle will be adopted in future experiments, in describing
which an account of its construction will be given. It may be
said, however, that the little instrument actually employed
generally controlled the speed to such uniformity as allowed the
deflections to be observed with as much accuracy as the zero-point.

The scale and telescope hardly require special description;
they were arranged in the usual manner for this kind of experi-
ment, at about three metres fix)m the mirror. The scale was an
engine-divided paper scale nailed to a wooden bar. This plan will
in future experiments be abandoned, as variations in the weather
had a very perceptible influence on the scale.

The annexed (p. 145) diagram shows the electric baiajice by
which the copper coil C was compared with an arbitrary German-
silver standard 8 before and after each induction experiment.
The arrangement is that of the ordinary Wheatstone's balance, as
described in Appendix H of the Report of your CJommittee for

FOR ELECTRICAL MEASUREMENTS

145

1862. A and C represent the arms of the balance ss there
described^ S the German-silver standard, and R the copper coil -to>
be measured. JJi, H Hu MMj, and L Li are four stout copper
bars with mercury-cups at aoiO,..., 66169..., cci, and ddi. Two
short copper rods, F and J^i, can be used to connect a with 6 and 0,
with d. When this is done the arrangement is exactly that of the
simple Wheatstone balance with the keys at K and Ki, and
described in Appendix H of the last Report. A and C were coils
formed of about 300 inches of No. 31* German-silver wire, and
were adjusted to equality with extreme nicety, and each assumed
equal to 100 arbitrary units. If 22 on any occasion had been
exactly equal to S, the galvanometer 0' would have been un*
affected on depressing the keys KKi, when a was joined to 6 and
c to d hy F and jPi, rods of no sensible resistance. This exact
equality between R and 8 could never be obtained, owing to
slight changes in temperature, which affected the two coils very

differently. The object of the modifications introduced was to
allow the ratio between 8 and iZ, differing by a small amount only,
to be measured with great accuracy.

For this purpose a number of German -silver coils were adjusted,
representing 1, 2, 4, 8... 512 in the arbitreury units, equal to the
hundredth part of A or C. These coils were so arranged that any
one or more of them could be introduced between the bars HHi

* Diameter =0*01 inch.
B. A. 10

146 PRACTICAL STAKDARDS

and JJx, A single coil, equal to 1 in the same arbitrary unit,
could be introduced between the bars LL^ and JIf Jfi. In the
diagram this coil is shown in its position and the rod Fx withdrawn.
Similarly F is withdrawn from between H and Q\ and the coil 1
joins Oi and &i in the bars HH^ and «//i. If no other coils were
placed between HH^ and //,, the arms of the balance would now
be 101 and 101 respectively, instead of 100 and 100; but the
ratio would still be that of equality. Let us now suppose that,
when the circuit with the battery is completed, the galvanometer
by its deflection shows that 22 is bigger than By we can reduce the
resistance of the arm between D and Y by various small graduated
and definite amounts by introducing the coils 2, 4, 8, etc. between
HHx and JJ^- Let us first suppose the coil 2 introduced. The
resistance between H and •/. will be the reciprocal of 1*6 or 0*6667;
for where various resistances are added in multiple arc, the re-
sistance of the compound arc is the reciprocal of the sum of their
conducting powers, and the conducting power of a wire is the
reciprocal of its resistance. The ratio between the two arms will
now be 101 : 100*6667. Let us suppose that on completing the
circuit the galvanometer still deflects in the same direction as
before, the arm between D and Y must be still further reduced by
including firesh coils between HHx and «^«A» It is very easy by
trial to find the combination which maintains the galvanometer at
zero when the circuit is completed. Let us suppose that, as in
the diagram, the coils included were 1, 2, 4, 8, and 64. The
reciprocals of these numbers are 1, 0*6, 0*25, 01 25, and 0*015625.
The conducting power between H and J is therefore 1*890625,
the sum of these numbers. The resistance between H and J is
0*52893, the reciprocal of the last number, and the ratio between
the arms will be 101 : 10052893. A little consideration will show
that with the coils named any ratio between 101 to 100*5 and 101
to 101 can be obtained by steps not exceeding 0*00195, the
reciprocal of 512, the largest coil or smallest conducting power
which can be included between the copper bars HHx ^^^ JJi'
By substituting the rod F for the coil 1 between LLx and MMxt
the observer can obtain a fresh series of ratios with the same steps
between 101 to 100 and 100*5 to 100. In this way it will be seen
that unless the coils R and 8 differ by more than one per cent.,
their ratio can be measured in the manner described within 0*002
per cent.

FOB ELECTRICAL MEASUREMENTS 147

It should further be observed that extreme accuracy in the
coik 1, 2, 4, etc. is not necessary, since an error of one per cent, in
the sum of these, as compared with their true relative value to the
coil (7, would only affect the final result O'Ol per cent.

The position of M and 8 in the balance relatively to A and C,
etc. is of course interchangeable.

The diagram is not intended at all to represent the practical
arrangement, but simply to show the connesions. The electric
balance described in Appendix H of last years Report (Plate 1,
figs. 1 to 6, Report 1862) was used with a stout copper rod
between the cups eei, and two additional boards with the copper
bars HHi, c/c/„ LLi, and MMi fitted as in the above diagram.
The coils 1, 2, 4, etc. had amalgamated copper terminals, which
simply dropped into mercury-cups on the copper bars. The ob-
servations could be made very rapidly and accurately, as the
galvanometer was sensitive enough with four DanielFs cells to
indicate the addition or subtraction of the 512 coil with perfect
distinctness. The reduction of the observations to find the ratio
seems somewhat complicated at first, but with the aid of a table
of reciprocals it takes but little time. No improvement seems
necessary in this part of the apparatus. The idea of using large
coils combined with small ones in multiple arc to obtain extremely
minute differences of resistance was suggested to the writer
by Professor W. Thomson, and will be found useful in very
many ways.

Part III. — Mathematical Theory of the Experiment.

A circular coil of copper wire is made to revolve with uniform
velocity about a vertical diameter. A small magnet is suspended
by a silken fibre in the middle of the coil. Its position is observed
when the coil is at rest, and when the coil revolves with velocity
«9 the magnet is deflected through an angle ^. Currents are
induced in the coil by the action of the earth's magnetism, and
these act on the magnet and deflect it from the magnetic
meridian. By observing the deflection and the velocity of rota-
tion, we can determine the resistcmce of the coil in electro-
magnetic units.

In determining the strength of the current we may neglect
the motion of the suspended magnet, as it is found, both by theory
and by experiment, to be insensible. We have therefore, in the

10—2

1:48 PRACTICAL STANDABDS

first place, to determine the electro-magnetic potential of the coil
with respect to the earth's magnetism, with respect to the suspended
magnet, and with respect to itself.

Ist, Let- H be the horizontal component of the earth's

magnetism.
7 the strength of the current in the coil,
0 the total area enclosed by all the windings of the

wire.
0 the angle between the plane of the coil and the

magnetic meridian.

Then the potential of the coil with respect to the earth is

2nd. Let M be the magnetic n^oment of the suspended magnet.
<!> the angle between the axis of the magnet and

the magnetic meridian.
K the magnetic force at the centre of the coil due
to unit current in the wire.

Then the potential of the coil with respect to the magnet is

- MyK sin (0 - <f>).

3rd. Let Ji be the potential of the coil on itself for unit

current.
Then the potential due to a current 7 is

Let P be the electromotive force, and R the resistance, then
the work spent in keeping up the current is Py in unit of time ;

•

or, since P = Ry, the work spent in keeping up the current for a
time St is

If the current is at the same time increased from 7 to 7 + 87,
the work spent in increasing the current will be

Ly By.

If the angular motion of the coil be B0, the work spent in
keeping up the rotation against the electro-magnetic force is

HyO cos 0d0 + MyK cos {0 - 4>) d0.

Since this work is exactly consumed in keeping up or increasing
the current, we must have

HyGco8 0d0 + MyKcoB{0''if>)d0^Rffdt-{'Lydy.

FOR ELECTBICAL MEASUREMENTS H9

Since 0 = (ot and ^ = g), the solution of this equation is

'^t

+ iTif (iZ COS (^ - <^) + ifi, sin ((?- ^))} + Ce'i*,

the last term becoming insensible soon after the beginning of the
experiment.

We can now find the equation of motion of the magnet.

Let ^ be its moment of inertia, MHt the torsion of the fibre
per unit of angular rotation, then

A^^MKyco8(<l>''0)'-MH {sin <I> + T<f>),

Substituting the value of y and separating Xerms in 0, we find

^^.=l^^A(^S(R<^0B4> + La>sm<l>) + KMR}

- ifi? (sin ^ - T^)
+ ^ ;^^ {0^ (i2 cos (2^ - <^) + Zo, sin (2<9 - ^))

+ iTilf (iJ cos 2 (^ - <^) + Z/o) sin 2 (tf - <^))}.

In order that ^ may continue as it does nearly constant, the
part independent of 0 must vanish, or

^—^^^ [OH (R cos <I>+La>sm<l>)'\-KMR}

- JtfJ? (sin ^ + T<^) = 0.
This gives the following quadratic equation for R : —
^ 1 p OKc^ ( ^ . KM\ 1 OKLw' ^ ,

-K"— s -K -: 7 COS 6 + TTTy = - L(o\

2 8m<A+T6V ^ OH) 2- 0

1 +T^ -T

sm 9

The solution of this equation may be expressed to a sufficient
degree of accuracy as follows : —

OKt

2tan<^(l

To determine the quantities occurring in this equation, we
must measure the dimensions of the coil, the strength of the
magnet, and the force of torsion of the fibre.

160 PRACTICAL STANDARDS

1st. Dimensions of the coiL

Let a » mean radius of the coil = 0*1566 metre.

n = number of windings of wire = 307

I = eflFective length of wire = 27ma = 302*063 metres.

b = breadth of section of coil perpendicular

to the plane of the coil «= '0185 metre.

c = depth of section in the plane of the

coil « 0132 „

V ss distance of mean plane of coil firom

axis of motion = '01916 „

= angle subtended at axis by radius of

coil =83T.

COS a « - = •1224&
a

Then G^^'rMa'^l + ^^,) ,

£-= — sin«ajl + ^~(2-15sin»acos«a)

+ o7 "*l(lS sin' a COS* a — 3 sin*a)|- ,

OK =7rnZsm*a U + ^-,+ o — T" sm*acos*a-s -;8ii^ «^•
( 6 a' 8 a* 8 a* J

If the dimensions of the coil are measured in metres, OK will
be expressed in metres.

Let T be the time of 100 revolutions of the coil, expressed in
seconds, then

Ta> = 2007r,

2007r
or « = — «r" •

Let D be the distance of the scale from the mirror, S the
scale-reading measured fix)m the point of the scale which is nearest
to t}ie mirror, then

tan2<^ = ^;

1 ^D

2tan^ 5

jj\^ ISM

FOR SLSCTRICAL MEASUREMENTS 151

To determine MHr, the coefficient of torsion, let the magnet
be turned round so as to twist the fibre nearly 360°. Let the
difference of reading due to the torsion be B\ then

_ y 1

KAf
To determine j^ , let the suspended magnet A be removed,

and let another magnet, which we shall call B, be put in its place.
Let the magnet A be now placed east or west of B, at a distance

equal to the mean distance of the coil, or ^a' + b\ Let the
deflection of B when the north or south end of il is directed to it

be ^, then

KM ,
g-g- = tan fi.

The determination of the quantity L, the electro-magnetic
capacity of the coil, requires a more complex calculation, which
must be explained separately. In the actual experiment the
deviation if> was always small, and therefore tan' ^ was very small,
so that the term depending on L was never important.

We may now write the value of R,

^ 20a7r>J5nZsin»a,, ,. ,

jK =s — |l + corrections).

Li this expression the quantities Dnla are determined before
the experiment is made. The only quantities to be observed are T,
the number of seconds in 100 revolutions, and S, the deviation in
millimetres of the scale.

Part IV. — Details of the Experiments.

In the experiments at £[ing*s College, June 1863,
n, the number of windings, was 307.
I, the effective length of wire, 302063 metres.
8in»o = l --021756.
D, the distance from the mirror to the scale, 2*9853 metres.

DetermincUion of Velocity.

A wheel of 100 teeth turned by an endless screw caused a bell
to be struck every 100 revolutions of the coil. The times of the
bells, as observed with a chronometer, serve to determine T.

152 PRACTICAL STANDARDS

Determination of Deviation,

S is th^ difference between the reading of the scale when the
magnet is acted on by the earth only, and when it is acted on also
by the induced currents in the coil. To determine B, the reading
of the scale is made when the coil is at rest, or when the circuit is
broken. Another reading is taken with the connexion complete
and the coil in motion. If the earth's magnetism remains the
same, the difference of these readings is the true value of B ; but
since the direction of the earth's magnetic action is continually
varying, we must find the difference of declination between the
times of the two readings, and calculate what would have been the
undisturbed reading of the scale at the time when the deviation
was observed.

In our experiments thijg correction was made by comparison
with the photographic registers of magnetic declination made at
Eew at the same time that our experiments were going on.

Corrections.

The corrections being small may be taken separately. Each
has to be multiplied by the factor already considered,

22 = yg {l + il + i? + C + J5 + ^ + ^ + (7 + J5r + etc.}.

A. Correction for the dimensions of the section of the coil.

-4 = ^ -i + Q — r— sm«acos*a-Q -sm»a
6 a' 8 a* 8 a*

= + -000075.

B. Correction for level. Let the axis of rotation be inclined
to the vertical at an angle /3 measured towards the north, and let
the angle of the dipping-needle with the horizontal be /, then
there will be a correction,

S = — tan / sin fi.

In the actual experiment the level was taken with a spirit-level
reading to 12", and found correct to at least that degree of accuracy.

C. Correction for the induction of the suspended magnet on
the coil. The strength of the magnet, as compared with that of
the magnetic field, .was measured by means of a magnetometer
fix)m Kew by the ordinary method. The correction found was

(7 = -f tan fi

= -00780.

FOB ELECTRICAL MEASUREMENTS 163

The small magnet generates induction-currents in the coil
which react on the magnet, and tend to turn it in the direction in
which the coil revolves. If there were no horizontal magnetic force
due to the earth, the coil would drag the magnet round after it.
In the actual case it m'akes the deviation greater than it should be
by -0078.

D. Correction for torsion of the fibre,

^ ^ 4nrD

= - 00132.

This correction depends on the relation between the stiffness
of the fibre and the directive force of the suspended magnet. The
fibre was a single fibre of silk 7 feet long ; the magnet was a steel
sphere -^^ inch diameter, and not magnetized to saturation. The
correction for torsion was therefore much larger than if a stronger
magnet had been used.

R Correction for position of suspended magnet.

Let the centre of the magnet be at a distance ^ above or below
the centre of the coil, 17 north or south of the axis of motion, and
f east or west of the axis, then there will be a correction,

^«+l(l-4cot«a)sin*a|4^-^|-3^]l.

Here a = 156'6 millimetres, and the place of the magnet was so
adjusted that it could not vary one millimetre in any direction
without the error being observed. Hence this correction is
negligible.

F. Correction for irregularity in the magnetic field due to
iron or magnets near the instrument.

Let t be the time of oscillation of a magnet at the centre of
the coil, ti and t, at distances z above and below that point, then

16z* \ t J
This correction may also be neglected.

G. Correction of scale-reading. The quantity observed is
tan 2if>, the quantity to be found is tan ^. The correction to the
value of iZ is

154 PRACTICAL STANDARDS

H. Correction for electro-magnetic capacity of the coil.
Let L be the value of the electro-magnetic capacity, the cor-
rection is

1 » 2Z / 2Z
ID'OK \GK

-)•

In the actual coil L was found by calculation » 397750 metres,

and by a rough experiment = 398500 metres.

Now OK = 660246 metres.

1 S"
The correction is therefore - 1 ;gi (0-596234) = H,

This correction is of the same form with 0, and may be taken
along with it.

The complete expression for R is therefore

R^Y6 ^38145581730 + |j 3055-5.

The nature of the electrical action in the experiment may be
stated as follows : —

Suppose the plane of the coil to coincide with magnetic north
and south, and that the coil is revolving in the direction of the
hands of a watch. Then the north side of the coil is moving from
west to east, and therefore experiences an electromotive force
tending to produce an upward current. The south side of the
coil is moving from east to west, and therefore there is a tendency
to produce a downward current in it. If the circuit is closed there
will be a current upwards on the north side, and downward on the
south side round the coil.

Now this current will tend to turn the north end of the sus-
pended magnet towards the east ; but the earth's magnetic force
tends to turn it towards the north; so that the actual position
assumed by the magnet must depend on the relation between the
strength of the current and the strength of the earth's magnetism.
But the strength of the current depends only on the velocity of
rotation, the resistance of the coil, and the strength of the earth's
magnetism. Hence the position of the magnet will not depend
on the strength of the earth's magnetism, but only on the velocity
and the resistance of the coil.

We must remember that the coil in its revolution comes into
other positions than that which we have mentioned. As the

FOR ELECTBICAL MSASUREMENTS 165

north side moves towards the east, the current continually
diminishes till it ceases when it is due east. The current then
conmiences in the opposite direction with respect to the coil ; but
since the coil itself is now in a reversed position, the effect of the
current on the suspended magnet is still to turn the north end to
the east. The action of the current on the magnet is therefore of
an intermittent nature, and the position of the magnet is not
fixed, but continually oscillating. The extent of these oscillations,
however, is exceedingly smalL In &ct if Z* be the time of vibra-
tion of the magnet firom rest to rest under the action of the earth,
and if ^ be one quarter of the time of revolution of the coil, and if
S be the deviation as read on the scale, then the same amplitude
of these oscillations will be

C = jT, o.

In the actual experiment ^ = about ^^^ and S less than 400

millimetres, so that the whole extent of vibration would be less
than j^ of a millimetre on the scale. This vibration was never
observed and did not interfere with the distinctness of vision.

The only oscillations observed were the free oscillations of the
magnet. They arose fit)m accidental causes at the beginning of
the experiment, and were subject to slight alterations in magnitude
due to changes of speed of rotation, the passage of iron steamers in
the Thames, etc. The time of one vibration was about 9*6 seconds,
and by reading the scale at the extremities of every vibration a
series of readings was obtained, the intervals between which were
proximately equal.

Now since the deviation is proportional to the velocity

and if we take values of 8 at small intervals dt and sum them, we
shall get

js.dt^ cjvdt = Cx,

where x is the whole distance travelled in the time.

Hence all we have to do is to observe the deviation at every
oocillation, and to ascertain the whole number of revolutions
during the time of observation, and the exact beginning and ending
of that time. This was done in the following way.

166 PRACTICAL . STANDARDS

The coil was made to revolve by means of the driving machine,
and its velocity was regulated by the governor. While the required
velocity was being attained, the oscillations of the magnet were
reduced within convenient limits by means of a quieting bar at a
distance. The quieting bar was then put in its proper place and
the observation commenced.

One observer, Ay took the readings of the scale as seen in the
telescope, writing down the deviation at the extremity of every
oscillation and thus obtaining a reading every 9*6 seconds.

Another observer, B, with a chronometer, wrote down the times
of every third stroke of the bell. The times thus found were at
intervals of 300 revolutions. When the observer B noted the time,
the observer A made a mark on his paper, so that after the experi-
ment the readings of deviation could be compared with the readings
of the chronometer taken at the same time.

The mean time of revolution between any two times of ob-
servation could thus be found and compared with the mean
deviation between the same limits of time, and any portion of an
experiment accidentally vitiated could be rejected by itself.

The experiments of each day commenced with a comparison,
by means of an electric balance*, between the resistance of the
experimental coil and that of a German-silver coil (called
"June 4").

Then a series of readings of the scale was taken to determine
the undisturbed position of the magnet. The times of beginning
and ending this series were noted, and called Times of 1st Zero.

Then the coil was made to revolve, and readings of deviation
and of time were taken as already described, and called 1st Spin -H.

Then the direction of rotation was reversed and a second set of
readings obtained, and called 2nd Spin — .

Then the undisturbed position was again observed with a note
of the time. This was called 2nd Zero.

Lastly, the resistance was compared again with the standard
coil. This series of experiments was then repeated if there was
time.

From the values of 1st zero and 2nd zero, together with the
information obtained from the photographic registers at Kew, the
true value of the undisturbed reading during the 1st spin and
2nd spin was obtained. The diflFerence between this and the

* Vide Report, 1862, p. 169, and present Appendix, p. -99.

FOR ELBOTRICAL MEASUBEHBNTS 157

actual reading is the deviation S due to the electric currents.
T was got by the chronqn^eter readings. Now let r be the re-
sistance of the staixdstrd coil at standard temperature, iZ the
resistance of the experimental coil during the experiment, then
by the comparison of resistances we find

where x is the ratio observed by means of the electric balance.

N
But we also know that iZ = =^ + correction, where iV is a known

number given at p. 113. Hence r, the resistance of the standard
coil, may be found in. absolute measure by the formula

the value of xTh should therefore be nearly constant.

Thus, on June 23rd, 1863, the experiments were made as
follows : —

At 12^^ 15°^ the resistance of the copper experimental coil was
compared with that of standard coil " June 4 " taken at 101, and
found to be 10126.

From 12'* 36° to 12»» 46° the undisturbed position of the sus-
pended magnet was observed, and found to be 590*28 scale-divisions
as the mean of all the readings.

The position of the declinometer at Kew at the same time was
7'689 of its own scale-divisions.

From 12*» 47" hVh to l"* 3° 13- the position of the magnet
was again observed while the coil was revolving ; 104 readings of
the scale were taken, of which the mean was 93059. This, when
corrected for scale-error, gives 931*48 as the true reading. The
position of the declinometer at Kew during the same time was
7*679. The resistance, measured after the experiment, was 101*28.

The number of revolutions was 6300 during the time of ob-
servation, so that the time of 100 revolutions was 14''464.

By comparing the Kew apparatus with that at King's College^
it appears that 1*0 of the Kew scale = 19*137 of the King's
College scale. The undisturbed readings at King's College were
found actually to vary very nearly in this proportion to those
at Kew.

Hence it is easy to find the undisturbed reading during any
given experiment by comparison with the Kew numbers.

158 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

Thus, for the first experimeDt on June 23rd we get

Corrected undisturbed reading 591 '54

Deflected reading 931-48

Deflection S = + 339-94

Time of 100 revolutions = T = 14-464

Product T8 = 491690

Resistance at time of experiment a; ... » 101*28

TBx = 4979-76

Three other experiments were made on June 23rd. The result
of the four experiments was as follows ; —

Ist experiment. Positive rotation ... T.S. a? = 4979*75

2nd „ Negative T.B.x =507118

3rd „ Positive r.S.a? = 5093-35

4th „ Negative T.S,x =500766

Mean Positive result 5036'56

Mean Negative result =5039-42

Mean result of June 23rd 5037*98

Mean result of June 19th 5075-77

Mean result of June 16th 504618

Mean of three days 5053*32

It will be observed that the mean results of each day are more
concordant than the individual experiments made on the same
day. The errors, therefore, which we have hitherto been unable
to get rid of are not of a kind which would have the eflFect of
making the result depend on the arrangements adopted on the day
of experiment, but are rather such as would destroy one another in
any long series of experiments.

Dividing N by the number just found, we get for the resistance
called 100 provisionally,

106493470 + 61100 = 10655470,

the second term being the correction for self-induction and for
scale-reading.

Since the coil of German silver, marked June 4th, was called
provisionally 101, we find as the result of the experiments for the
resistance of " June 4 " in absolute measure

107620116 metres per second.

Knowing the absolute resistance of "June 4," we may construct
coils of given resistance by known methods.

THIRD REPORT— BATH, 1864.

In the present Report it is thought unnecessary again to refer
to the objects with which the Committee were appointed, or to
recapitulate the arguments for and against the various systems of
standards which have been from time to time proposed. The
Committee have seen no reason to alter the conclusions previously
adopted, and now propose briefly to state the progress made in the
practical development of those conclusions, which may be found
expressed at length in the Report for 1863.

That Report announced the adoption by the Committee of the
abaolate electro*magnetic system of measurement, based on the
metre, gramme, and second, with certain modifications to facilitate
the practical construction or use of the standards ; and it further
stated that in 1863 the absolute resistance of a certain German*
silver coil had been measured with considerable accuracy.

No standards based on the 1863 determination were officially
issued, inasmuch as it was felt that a second determination was
absolutely required before complete dependence could be placed
either on the method employed or on the results obtained. Some
coils representing 10 of the British-Association units, i,e. 10'
absolute units according to the 1863 determination, were made by
Messrs Elliott Brothers, and a set from 1 to 10,000 was made from
the 1863 determination by Messrs Siemens and Halske of Berlin.
This last set is intended for Col. Douglas, the Superintendent of
the Government telegraph lines in India; and a few of Messrs
Elliotts' coils have been bought by persons who were unwilling
to wait for the final experiments by the Committee. None of
these coils have been in any way certified as correct by the
Committee.

In order thoroughly to test the value of the experiments made
in 1863, it was determined that not only every measurement
should be made afresh, but that every element in the experiment
should be varied. The experiment consisted essentially in causing

160 PRACTICAL STANDARDS

a coil, or rather two coils, of copper wire to revolve or spin at a
certain definite rate, and in observing the deflection of a magnet,
suspended within the coil, by the reflection of a scale in a mirror
attached to the magnet.

The measurements required in the calculation are the
following : —

a. The mean radius of the coils.

n. The number of turns made by the copper conductor forming
the coils.

L The eflfective length of the wire.

b. The breadth of the section of the coil.

c. The depth of the section of the coil.

b\ The distance of the mean plane of the coil from the axis
of rotation.

T. The time of 100 revolutions of the coil,

D. The distance of the scale trom the mirror.

S. The scale^reading during each experiment.

The above measurements are required for what may be called
the simple theory, that is to say, the theory omitting all the
necessary corrections arising from self-induction, torsion of fibre, etc
For these corrections it is further necessary to measure —

1st. The coeflBcient of torsion of the fibre.

2nd. The magnetic moment of the suspended magnet.

3rd. The horizontal component of the earth's magnetism.

4th. The variation of the electrical resistance of the coil
during each experiment and between each experiment.

5th. The variation in the direction of the earth's magnetic
force.

6th. The irregularities resulting from the unavoidable de-
partures from that relative position of the telescope, mirror, scale,
and magnet which would be theoretically most desirable.

In the experiments made at King's College in 1864, every
part of the apparatus, except the distance of the mean planes of
the two coils from the axis of rotation, was altered ; so that every
measurement was not only made afresh, but, where susceptible of
change, was considerably diflferent in magnitude.

Few of the measurements could be made by the means em-
ployed with greater accuracy than one part in 10^000, and some of
them were not determined even with this degree of accuracy. No
very perfect agreement between two entirely distinct series of

FOR ELECTRICAL MEASUREMENTS 161

experiments was therefore to be expected; but the Subcommittee,
conaistmg of Professor Maxwell and Mr Jenkin, who this year have
undertaken the experiments, are fortunately able to report a con-'
oordance between the determinations of 1863 and 1864 which is
most satisfactory.

The difference between a standard constructed from the mean
result of the 1863 experiments and a standard constructed from
the mean result of the 1864 experiments would be only 0*16 per
oent The probable error of the 1863 experiments is 0*24 per cent.
if the mean of each day's experiments be counted as one only ; the
probable error of the 1864 experiments is O'l per cent, if the mean
of each pair of experiments with the coil revolving in two opposite
directions be taken as one experiment.

Taking into account the agreement between the means of the
two years, we may say that the determination of the Subcommittee
does not probably differ from true absolute measurement by 0*08
per cent.

The Committee are of opinion that, in the present state of
electrical science, the result now obtained is satisfactory, and will
justify the immediate construction of final standards of electrical
resistance.

It can hardly be doubted that, with the lapse of time and the
inevitable progress of knowledge, still better determinations will
some day be made ; and that even now, with still greater care and
by still further multiplying the number of experiments, a somcr
what more perfect agreement between the standards and the
theoretical absolute measurement could be ensured.

The Committee had then to consider whether this possibly still
more perfect agreement would be worth the very great time, the
labour, and the money which would have to be bestowed upon it.
It has never been proposed that the British- Association standard
should be considered as representing exactly an absolute measure-
ment ; whatever may be the state of science, any such pretension
could not be well founded, for all that can be done at any time, by
the veiy greatest care, is to reduce the possible error to less than
a certain amount. The amount of probable error in the present
determination is so small as to be insignificant for any of the
present purposes of science, and will always remain insignificant
for any practical applications. For these applications it is chiefly
important that every copy of the standard, whatever that may be,

B. A. 11

162 PRACTICAL STANDARDS

should be accurately made — a condition which is quite unaffected
by the greater or less discrepancy between the standard and true
absolute measurement.

The reproduction of the standard can perhaps be more easily
effected, if ever it be necessary, by a given weight of metal or alloy
than by a firesh absolute determination.

Meanwhile practical standards of resistance are urgently re-
quired, and the Committee are pressed te come te a decision*
Defective systems are daily taking firmer root, and the measure-
ment of currents, quantity, capacity, and electromotive force call
urgently for the attention of your Committee.

Under these circumstances they have decided te rest content
with the results of the experiments now completed, and to com-
mence at once the construction of standard coils.

The details of the experiments on absolute resistance are given
in Appendix A.

It may be useful here to mention that the new unit will be
roughly equal to 0*0736 times Dr Matthiessen's mile of copper
wire, and more exactly 1*0456 times Siemens's unit, according to
standards which have kindly been sent by Dr Siemens to several
members of the Committee and others*.

The questions of chief importance, after the magnitude of the
standard has been chosen and determined, concern the choice of a
suitable form and material for the actual construction of the
standard ; and in this choice the permanence of the standard is
above all essential.

Dr Matthiessen has for two years been endeavouring, at the
request of the Committee, to discover whether the electrical resist-
ance of various metals, under various conditions, can be considered
as constant, or can be proved to alter. His Report for the present
year is given in Appendix B, and will be found to confirm, in a
great measure, the conclusions arrived at in his Report for 1863.

No variation has been observed by him in the electrical re-
sistance of annealed wires of silver, copper, gold, platinum, nor in
the hard-drawn wires of gold, platinum, or of the gold-silver
alloy. But a change has been observed in the hard-drawn wires
of silver and copper — a change most rapid in the first year, but

* Twenty-five units are within one per cent, equal to the mile of No. 16 copper
wire in use by the Electric and International Company. Mr Varley has promised
that for the future exact equality shaU be aimed at.

FOR ELECTRICAL MEASUREMENTS 163

yeiy sensible in the second year ; a somewhat capricious change
has also been observed in certain annealed German-silver wires,
while others have been proved constant. This result has been inde-
pendently observed by other members of the Committee. In the
hard-drawn wires of silver and copper the direction of the change
has been such as to bring the resistance of hard-drawn wires more
nearly to resemble that of annealed wires, diminishing the re<^
sistance ; in other words, it is such a change as would be produced
by partial annealing.

From these experiments it is clearly undesirable that silver or
copper should be used for standards even in their annealed state ;
and the change in these metals further indicates that for standards
of other metals the partially annealed is preferable to the hard-^
drawn condition.

The experiments on these points must be continued for many
years before much reliance can be placed on the results; and
meanwhile equal standards must be constructed of various
materials, and protected in various ways, for reference and
comparison.

The precautions taken to prevent chemical action and me-
chanical injury are given in Appendix B of the Report for 1863.
Coils of wire covered with silk, baked and imbedded in solid
para6Bn, appear, at present, to be the most promising form for the
unit standards. Authentic copies of the standard coils made of
platinum-silver alloy, which appears likely to be permanent, might
be issued at about £2. 10«. each, and coils prepared fix>m these by
electrical instrument-makers could be verified at a moderate rate
at Kew, where the original standards will be deposited. No
officially authentic coil can be issued until the standards them-
selves have been made.

The reproduction of the standard forms the next point for
consideration. Notwithstanding the good results obtained by
Professor Thomson's method of making an absolute measurement,
the Subcommittee do not recommend the adoption of this process
for the reproduction of the standard, which may some day become
necessary, owing to the accidental destruction of, or change in, the
Kew standards. Dr Matthiessen, on the other hand, states, with
confidence, that a standard may be reproduced by means of metal
wires of given weight and length, or by means of mercury, within
about (H)l percent.; the report of his investigation on this subject,

11—2

164

PRACTICAL STANDARDS

made conjointly with Mr C. Hockin, is contained in Appendix C,
and may be summed up as follows. He first draws a distinction
between ordinary care, great care, and absolute care. He considers
that with ordinary care the gold-silver alloy is the most suitable
material (see Report, 1862) for the reproduction, but when great
care is used lead is recommended as the most suitable material ;
but any reproduction by one material should be checked by others,
such as mercury. With absolute care it appears that almost any
material might be used. It must be remembered that Dr
Matthiessen considers that he himself has not taken absolute but
only great care.

The following Table shows the number of wires of each material
tested, their maximum discrepancy, and the probable error in a
standard reproduced by similar experiments: —

Metal

No. of
wires

Mazimam disorepanoy
expressed as a fraction
of the whole oonduot-
ing power

Probable
error

Silver

Copper

3
3
3

4
6
3

0O014

0-0011

0O005

000054

0-00073

0-00151

0-00052
0-00021
0-00011
0O0006
0-00001
0-00009

Gold

Lead

Gold-silver alloy

Mercury

Commercially pure lead differed from the chemically pure lead
by only about 0*04 per cent.

For an account of the care taken by Dr Matthiessen in the
chemical preparations of the metals he used, and in their sub-
sequent treatment and electrical comparison, we must refer to
Dr Matthiessen's own Report, Appendix C.

With reference to mercury, great difficulty exists in making the
experiments ; and it is much to be regretted that Dr Matthiessen's
experiments, very accordant in themselves, do not give results-
agreeing with Dr Siemens s experiments. The discrepancy will be
best explained by the following Table, giving the value of a
column of mercury at 0°C. one metre long, and having a cross
section equal to one square millimetre, according to various ex.^
periments, and with the specific gravity used respectively by
Dr Siemens and Dr Matthiessen.

• . .1

[To face page 165]

Description

Name

Absolute j X 10*^ electromag- 1

second f

netic units (new determination) J

Absolute -J X 10^ electromag- 1

second >

netic units (old determination) J

Twenty-five feet of a certain copper )
wire, weighing 345 grains ]

metre

Absolute J X 10^ electromag- 1

second ^* {

netic units determined by Weber

(1862)

One metre of pure mercury, one)
square millimetre section, at 0** C. )

One metre of pure mercury, one^
square millimetre section, at O"" C. )

One metre of pure mercury, one)
square millimetre section, at 0*" C. (

British- Association unit B.A. Unit, or Oh mad

Absolute , xlO'

second

Thomson's unit

Jacobi

Weber*s absolute
metre

second

Siemens's 1 864 issue ,

Siemens (Berlin),

Siemens (London) ...

One kilometre of iron wire, four
milli metises in diameter (ten^pe-
rature not known)

Oiie kilometre of iron wire, four
millimetres in diameter (tempe
rature not known)

One kilometre of iron wire, fourl
millimetres in diameter (tempe- /
rature not known) J

One English standard mile of pure]
annealed oo[)per wire y\ inch in >
diameter at 15' "5 C. j

One English standard mile of one |
special copper wire ,^^ inch in .
diameter J

One German mile = 8238 yards of\
iron wire j^ inch in diameter >
(temperature not known*) J

Diguey

Bn*quet

Swiss

Matthiessen.

Varley

German mile

Absolut
foot

second

l-000«»

1 osoi*

2*0881'

301

r>fc

3138*

I
3156?

3-194

3*82]^

30-40

32-03

34-21

44-5'

.84-01

188-4

Mfttih

0-0!

O-Oli

0O(

o-c;

O'O';

oi)-;

CH3'

0-6^

0-71

0*7(i

l-OO

1-88

4-2:2

P>ove.

FOB ELECTRICAL MEASUREMENTS 165

Definition J?*°« ^

B.A. amis

1. Mercury unit according to Siemens's standard issued

in 1864. Sp. gr. mercury assumed at 13*557 ... 0*9564

2. Mercury unit according to Siemens's experiments

made for 1864 standard, but assuming sp. gr.
mercury at 13*595* 09534

3. Mercury unit according to Dr Matthressen's experi-

ments. Sp. gr. mercury assumed at 13*557 0*9646

4. Mercury unit according to Dr Matthiessen's experi-

ments. Sp. gr. mercury assumed at 13*595 0*9619

5. Mercury unit according to one set of coils exhibited

in 1862 by Dr Siemens (Berlin) 0*9625

6. Mercury unit according to a second set of coils ex-

hibited in 1862 by Dr Siemens (London) * 0*9742 *

Dr Matthiessen considers No. 4 the true value, while Dr
Siemens supports No. 1. The Committee do not desire to express
any opinion on this subject, but only to draw attention to the
great discrepancies which follow the apparently simple definition
of the mercury unit (first proposed by Mari6 Davy). Even now
it cannot be said that a trustworthy standard, answering to the
definition, exists.

The Committee have little to report concerning the standard
instruments for the measurement of currents, quantity, capacity,
or electromotive force. The drawings for a standard galvanometer
and electro-d3mamometer have been begun. An electro-dynamo-
meter, suitable for general use, has been constructed by f^rofessor
W. Thomson, and experiments are being made with it.

Professor Thomson has also had some fine apparatus made for
the measurement of electrostatic phenomena and their comparison
with electro-magnetic measurements; but it will be best to describe
the instruments when the experiments have been completed.

Dr Joule has made some preliminary experiments with the
view to redetermine the mechanical equivalent of the unit of heat
by electrical means.

Thus, although the Committee have not accomplished all that
they hoped, they feel that such progress is being made as will
justify their reappointment.

* Thifl is the mean of the Tallies giyen by Eopp, Begnanit, and Balfour Stewart.
The diserepanoy between the two Tallies is far greater than oonld be dne to any
eonftision as to the reference of the specifio gravity to water at U** G. and at maximum
deDnty.

166 PRACTTICAL STANDARDS

They have received assurances that the British-Association
system of units will be readily adopted in this kingdom, in India,
Australia, and Qermany. They believe that it will be accepted in
America and in many other parts of the world.

From France no response has yet been obtained.

The Committee wish to express their sincere regret at the death
of one of their members, Dr Esselbach. He had made valuable
experiments on the electromotive force of various chemical com-
binations, and had promised to communicate them to the Com-
mittee; but their record is now probably lost.

Before concluding, the Committee have to thank Mr Charles
Hockin for the efficient assistance he has afforded, both in the
determination of the resistance unit and in Dr Matthiessen's
researches.

Appendix A. — Description of a farther Experimental measure-
ment of Electrical Resistance made at King's College. By
Prof J. C. Maxwell and Mr Fleeminq Jenkin, with the
assistance of Mr Charles Hockin.

The method employed in these experiments has been fully
described in Appendix D to the Report of 1863. In the new
experiments the elements of the calculation were varied as much
as possible; fresh wire was wound on the experimental coils;
observations were made with velocities differing widely from one
another. Fresh measurements were made of all the corrections
required, and greater precautions were taken to avoid local dis-
turbances.

w, the number of windings, was 313.

I, the effective length of the wire ... 31 1*1 18 metres.

-, the mean circumference 0*993987 metre.

a, the mean radius 0*158194 „

6, the breadth of each coil 0*1841 „

26, the distance from centre to centre

of each coil 003851

c, the depth of the layers 001 608 „

The weight of the wire and silk 110 oz. 8 dwt.

logsin»a = l-9624965.

D, the distance from the mirror to the scale; 2212 millims. in

some experiments, 2116 millims. in others.

FOR ELECTBICAL MEASUREMENTS

167

The following Table gives the result of the experiments, and
the comparison with those of 1863.

Time of 100

rerolations,

in seconds

17-64

17-68

77-62

7617

63-97

64-63

41-76

41-79

64-07

63-78

17-697

17-783

17-81

17-78

17-01

16-89

21-36

21-38

21*362

21-643

11-247

16-737

Values found for

coil in terms of

lO' for each

experiment

4-7201
4-6914
4*8848
4-4871
4-6607
4-6666
4-6279
4-6276
4-6496
4-6146
4-6108
4-7313
4-6452
4-7489
4-7667
4-6187
4-6834
4-6727
4*6526
4-7134
4-8668
4-6305

Value of B.A. unit in

i. • t^M metre

terms of 10' 3- ,

seconds'

as calculated from

each experiment

1-0121
0-9836
1-0468
0-9613
0-9986
0-9998
0-9916
0-9936
0-9961
0-9886
0-9878
r0136
0-9962

-0174
•0191
-9896
-0034
•0011
0-9968
1-0096
1-0424
0-9707

1
1
1
1
1

Value from
mean of each

pair of
experiments

0-9978
1-0040
0-9992
0-9926
0*9924
1-0007
1*0063
1*0043
1-0022
1*0040
0*9981

Percentage
error from
mean value

-0-22
+0*40
-0*08
-0-76
-0*76
+0*07
+0-63
+0-43
+0*22
+0*40
-0-19

Probable error of R (1864) = 0-1 per cent.
Probable error of R (1863) = 024 per cent.
Difference in two values 1864 and 1863 = 0'16 per cent.
Probable error of two experiments = 0*08 per cent.

In constructing the standard coil, in consideration of the much
greater range of velocities used in 1864, the 1864 mean value was
allowed to have five times the weight of the mean value obtained
in 1863.

Appendix B. — On the Electrical Permanency of Metals and Alloys.

By A. Matthiessen, F.R.S.

In Appendix A of the Report of your Committee of last year,
I gave the results of some experiments made to test the electrical
permanency of some metals and alloys. On August 5 of this year
I re-tested them, and give the results in the following Table^

168

PRACTICAL STANDARDS

taking the conducting power of No. 15 = lOO'OO, as was done in
last year's Report.

May 9,
1862

1. Silver: hard-drawn

2. Silver: annealed

3. Silver: hard-drawn

4. Silver: annealed

5. Copper: hard-drawn

6. Copper: annealed

7. Copper: hard-drawn

8. Conner: annealed

9. Qoid: hard-drawn

10. Qold: annealed

11. Gold: hard-drawn

12. Gk)ld: annealed

13. Platinum: hard-drawn...

14. Platinum: hard-drawn...
16. Gold-silver alloy : hard-
drawn

16. Gold-silver alloy : hard-

drawn •»

17. German silver: annealed

18. German silver: annealed

19. German silver : annealed

100-00
100-00
100-00
10000
10000
100-00
10000
100-00
100-00
100-00
100-00
100-00
lOOOO
100-00

lOOW

10000
100-00
lOOOO
100-00

20-2
20-2
20-2
20-2
20-1
20-1
20-0
20O
20-0
20-0
20-0
20-0
20-0
20-0

20-0

19-9
20-3
20-3

Jane 14,
1863

103-915
99-947
102-807
100-031
100-248
100-016
100-149

100-046
100-062
99-869
99-877
99-951
99-999

100-000

99*963
100-162
100-146
100-217

20-0
20-1
20-2
20-0
20-2
20-0
19-8

20*2
200
20-2
20-3
20-2
20-2

20-2

20-3
20-0
20O
20-2

Aug. 5,
1864

104-397
100-013
103-665
100-048
100-276
100-010
100-200

100-000
99-960
99-937
99-960
99-989

100-008

100-000

99-996
100-136
100-162
100-193

20-2
20-1
20-1
20-0
20-0
20-1
20-2

20-2
20-2
20-0
20-0
20-2
20-1

20-2

20-0
20-0
20O
20-2

From the above it will be seen that the following wires have
not sensibly altered in their conducting power during the space of
two years : —

No.

May 9,
1862

Jane 14,
1863

August 5,
1864

Maximum

difference

corresponds to

100-00

99-911

99-977

0-25

100-00

99-959

99-976

0-10

100-00

99*979

lOOOlO

0'05

100-00

100117

100-072

0-30

10.

100-00

100-062

100-032

0-20

♦11.

100-00

99*941

99-937

0-15

♦12.

100-00

99-985

99-960

0-10

13.

100-00

100-023

100-061

0-15

14.

100-00

100-071

100-044

0-20

15.

100-00

100-000

16.

100-00

99-963

99-996

* Without taking into consideration the corrections dae to temperature, I placed
in last yearns Report these two wires with those whose conducting powers ha<l
changed.

FOB ELECTRICAL MEASUREMENTS 169

All the values have been reduced to the first observed tem-
perature, assuming that all pure metals vary in conducting power
alike with temperature. The correction made was the addition
or subtraction of 0'036 for each 0***1, which number corresponds to
the correction of conducting power for temperature at 20**. No
correction has been made in the cases of Noa 15 and 16, for it is
so small that it may be neglected, being about 0*006 for each O"*'!.

As stated in last year's Report, the differences may be con-
sidered due to temperature ; for, as there explained, a difference
in the temperature of the wire and the bath might well exist,
and we find in most cases a difference in the conducting power
corresponding to 0°'l to 0***2.

It is interesting to find that hard-drawn silver and copper wires
become partially annealed by age, at least the increment in the
conducting power would indicate such to be the case. In the case
of silver, a decided increment will be observed*

No. 8, copper, annealed, has altered so much, that there can be
no doubt that it was badly soldered.

With regard to the alteration observed with the German-
silver wires, it may here again be stated that it is not to be
assumed that all wires of this alloy will alter in like manner. An
example of this has lately come to my notice. About two years
ago I made a coil of the gold-silver alloy, which was compared
with one of Prof. Thomson's German-silver coils ; and having them
still in my possession, they have now been re-compared, with the
following results : —

July 8th, 1862. Resistance of Thomson's coil being 1 at 18'''4,
that of the gold-silver coil was *88445 at IS''*^

August 6, 1864. Resistance of Thomson's coil being 1 at 18°*4,
that of the gold-silver coil was -88447 at 18**-4.

It is worthy of remark that the first comparison was made by
Dr C. Vogt, the last by Mr C. Hockin, and with entirely different
apparatus, showing that different observers with different apparatus
obtain absolutely the same results when they take great care in
making the observations.

The above proves that the conducting power of all specimens
of German-silver wire does not alter by age. Further experiments
are being made on this subject, and in the course of a year or so
we shall be able to say how far German^silver may be trusted for
making resistance coils.

170 PRACTICAL STANDARDS

Appendix C. — On the Reproduction of Electrical Standards by
Chemical Meams, By A. Matthiessen, F.R.S., and C. Hockin,
Fellow of St John's College, Cambridge.

Having been requested by your Committee to make some
experiments with the view of discovering the best method of
reproducing a unit of electrical resistance by chemical means, we
have carried out the research of which we now propose to give
the results.

The experiments have been made with unusual care. It is
important to point out the degree of precaution that has been
taken to insure trustworthy results. The care taken in these
experiments may be called great care as opposed to ordinary care
on the one hand and thorough care on the other. By ordinary
care is meant the care usually taken in scientific research, where
no extraordinary precautions are had recourse to. The sort of
accuracy obtained when a unit is reproduced with ordinary care
may be seen by reference to former results. For instance, in the
determination of the conducting power of mercury, described in
Phil, Trans,, results were obtained diflfering in some cases by
1*6 per cent. The same degree of accuracy was obtained in the
determination of the mercury unit by Dr W. Siemens, described
in Phil. Mag.

On the other hand, in the experiments to be described, and
in those made by Mr Sabine, the results differ by only a few
hundredths per cent.

The results of the determinations of the conducting power of
the gold-silver alloy, described in the Phil. Mag. Feb. 1861, differ
from each other by 1*5 per cent.; the values now found for the
same quantity differ by only seven-hundredths per cent. No doubt
if greater care had been taken and more perfect instruments used,
still better results would have been obtained.

Perhaps the great difference between what is above called great
care and ordinary care lies in the time employed. The experi-
menter using great care has to neglect almost all consideration of
time, and repeat his experiments at reasonable intervals, in all
cases in which it is possible that by lapse of time such error as at
first there is no means of detecting may increase and so become

FOR ELECTRICAL MEASUREMENTS 171

apparent. The meaning of absolute care is clear. When absolute
care is taken no precautions are omitted, the best instruments
obtained, and every care taken in the manipulation.

The apparatus used in the following research will first be
described, the results obtained will be then given, and finally some
remarks made on them.

Description of Apparatus.

Battery, — The battery employed was a single Bunsen's cell.
The wires connecting it with the bridge ran parallel to each
other the whole of their length, so that no attraction was exercised
on the magnet of the galvanometer by the current traversing
them*.

Balance. — For measuring the resistance of the wires a Wheat-
stone's balance, as modified by EirchhoJBT, was employed. A plan
of it is given in Plate 3 (fig. 1).

L and 22 are two resistance coils acting as the arms of the
balance. They are joined by the wire AA\ along which the
block B connected with one end of the galvanometer coil can be
moved.

The wire ^^1' of the instrument was made of an alloy con-
taining 85 per cent, of platinum and 15 per cent, of iridium.
The advantages of employing this alloy are that it does not
readily oxidize, that it does not change much in conducting power
with an alteration of temperature, and that it does not alloy with
mercury.

S isBL standard coil immersed in an oil-bath.

OP is the wire to be me^isured or compared with the standard
8, and is immersed in a large trough of water.

(? is an ordinary galvanometer by which approximate results
are first observed.

6i is a very sensitive Thomson's reflecting galvanometer, by
which the final observations are made.

Jf], Mf etc., mi, tn, etc. are mercury cups used to connect the
several parts of the circuit by thick copper rods and bars, plainly
shown on the drawing. The arrangement shown was found con-
venient, as it admitted of adjustment to various positions and
dimensions of the conductors to be compared. The position of B

* The battel^ oirooit was generally broken, and was closed by pressing down a
treadle, placed under the table, with the foot. The terminals were of platinnm.

172 PBACTICAL STANDARDS

on the wire A A' could be observed by a boxwood scale divided into
millimetres and a pointer on the block.

^ is a key used to complete the battery circuit, and worked
by a treadle firom below. An enlarged section of the block B is
given in fig. 3. a is a wooden handle by which the rod 6, with
the platinum point (2, can be depressed so as to come in contact
with the wire of the bridge. When the pressure of the hand is
removed the spring e lifbs the handle and breaks the contact.
The galvanometer wire is screwed in between the metal plates f
and g. A pad of gutta-percha between the knob h and the handle
prevented any sensible thermal current. To the top of the block
tiras fixed a piece of brass with a slit in it to serve as a pointer. A
lens also was fastened to the handle to read fi'actions of a milli-
metre on the scale. The body of the block was of lead, with a
slab of ebonite at the top and bottom. The block ran on a tram-
way parallel to the scale and wire of the balance.

A section of one of the mercury cups is given at fig. 2. At
the bottom of the cylindrical cup hmvo is placed an amalgamated
copper plate, and mercury is poured into the cup; the plate is
held down by the wooden cylinder j9, and this is kept in its place
by the pin rs. This plug fits the cup closely, and is pierced with
two or more holes for the terminals to pass through. The cupe
were propped up with wedges, when placed under the fixed
terminals of the balance, that these might press firmly against
the metal bottoms of the cups.

Eku^h of the coils R and L had a resistance of about 20 metres
of the wire of the instrument. Careful measures were made of the
resistance of the wire of the bridge at different points in order to
find if there were any very &ulty points in it ; this was done by
putting the coils R and L in their places, and increasing the resist-
ance of one of them by means of a short piece of wire. The effect of
this wire was to shift the zero-point. Two coils, differing about one-
tenth per cent., were then placed in the centre of the instrument
and the reading taken; these coils were then reversed and the
reading again taken.

Suppose %l the resistance of the circuit from the point £ to ^
when the short wire is removed, z the change in the zero-point
caused by the insertion of the short wire above mentioned, and m
the difference of a pair of readings, resistances being expressed in
millimetres of the wire AA\ and lengths expressed in millimetres

t^t'

rx_?j'

FOB ELECTRICAL MEASUREMENTS

173

of the scale ; then the resistance of a millimetre of the wire of the
instrument about the zero-point is

a-b 2

(^ + ^)-

a+ 6*a?'

r is the ratio of the two centre coils.

The value of this expression was found for different points
from one end to the other of the wire, and did not vary more than
two or three-tenths of a millimetre, an error not considerable
enough to affect the results obtained with the instrument.

The value of the coil R was thus found. It was placed in the
mercury cups mi', m,', and the cups m,, wi, were joined by a stout
copper bar. Two coils, the ratio of the resistance of which was
known, were placed in the two centre cups and the reading taken.

Let T be the ratio of two centre coils, x the reading of scale,

which was divided from A' to Ay R+r the resistance of the circuit
from B^ to the point of wire opposite that end of the scale nearest
to R, viz. A', I the corresponding quantity for the other side of the
instrument.
Then clearly

iZ-hr-fa; a
l + lOOO-x'^b'

2J+r«g(Z + 1000-a:)-a?.

The readings are given in the following Table : —

Ratio of I

Reading

Value of B + r

Value of K+r

24:1

120-6

20987 + 24Z

21216

26:

166-5

20964 4- 26^

21210

29:

269

20930 +29^

21206

34:

376

20876 + 34Z

21197

409

20867 + 36/

21208

37;

425-26

20841+37/

21192

464-26

20830+39/

21201

493*26

20790 + 42/

21188

647-26

20732 + 47/

21177

613

20672 + 66/

21193

646-26

20626+60/

21194

688

20628+68/

21173

720-76

20483+76/

21203

174 PRACTICAL STANDARDS

Zero-point was at 516.

Resistance of half length of circuit is 21712 millimetres of
wire.

All these values are within necessary errors of observation. The
first few values are most to be relied on, as the values of r + jR
depend nearly directly on 1000 — r.

So many measurements were made in order to find whether the
wire tapered towards either end. The similarity of the values
found for i2 + r shows this better, perhaps, than the direct method
before described.

A set of similar measurements were made with the coil L in
the left-hand mercury cups, and equally good results obtained.

The galvanometer employed was one of Thomson's reflecting
galvanometers, made by Messrs Elliott Brothers. A short coil
was employed. The instrument was placed in a deal box, blackened
inside, with large apertures to observe through. The spot of light
could thus be clearly seen, and the divisions of the scale were
sufficiently illuminated to enable the observer to see immediately
in which direction the spot of light moved. The instrument was
sufficiently delicate to show 0*001 per cent, diflference in the ratio
of any two nearly equal conductors compared, corresponding to
1^ millim. on scale of bridge.

An ordinary galvanometer was also at hand to find about the
place of reading on the scale.

The balance employed for weighing was by Liebrich of Giessen,
and would weigh to -j^^ of a milligramme with accuracy. The
weights were adjusted by Oertling, and again tested by weighing
them against the largest weight (50 grms.). Mr Balfour Stewart
was kind enough to test this weight, and found its value to be
exactly 50*000 grms. All weighings made in this research were
double weighings.

The measurements of lengths of wires tested were made with
a beam-compass. It was furnished with a vernier carrying a
telescope. The instrument was fixed horizontally before a window,
the ends being clamped to shelves in the wall on either side of the
window.

The telescope pointed downwards, and the wires to be measured
were laid on a board fixed below the instrument.

With this apparatus measurements could be made with the
greatest certainty to ^ of a millimetre, the telescope being suffi-
ciently powerful to show much smaller lengths than this..

FOR ELECTRICAL MEASUREMENTS 175

We are indebted to Mr B. Stewart for measuring the values of
the divisions of the instrument.

Thermometers, — Two thermometers were employed. They
were made by Messrs Negretti and Zambra. One was divided to
^ of a degree Centigrade, the other to single degrees. The large
thermometer was found to be correct by the Eew standard. The
zero-points of the thermometers were carefully taken.

Trough, — The wires, the resistances of which were to be deter-
mined, were placed in a glass tube immersed in a trough of water.

The trough was 1*5 m. long by 0*15 m. square section. A stream
of water flowed through it, coming in by the tube V (fig. 1) and
escaping by the waste-pipe W. This arrangement was adopted
because it was found that naphtha or oil soon acted on the wires
and altered their resistance, so that they could not be immediately
exposed to the action of a liquid. The details of the arrangement
will be understood by reference to fig. 4.

The wire to be tested, ah, was soldered at its ends to copper
bars, as ac. On to each of these bars was slipped a piece of glass
tubing, as ef. These tubes were fastened to the copper bars by
india-rubber tubing. The wire, with its connexions, was then
placed in the large glass tube AB, The piece of tubing ef
was then festened to the bent tube CEDF by india-rubber
tubing.

The ends of the terminals ac were beaten out flat and
amalgamated. The bent tubes were nearly filled with mercuryi
and the terminal o was connected with the mercury cups m^\ m^'
of the instrument by copper rods amalgamated at each end.

The resistances of the wires were compared with those of coils
of German-silver, well varnished, immersed in a cup of oil. The
temperature of the oil was determined by the small thermometer
before described.

Method of observing, — ^The wires were placed in the trough
and the connexions made. The water was then turned on and
i^Uowed to flow for about fifteen minutes. The large thermometer
was placed in the trough, and the temperature was read off by
means of a lens placed so as to avoid all error of parallax. The
small galvanometer was then connected with the electric balance,
and the approximate reading found.

The large galvanometer was next connected, and the block
handle pressed down until any thermal current that existed had
ceased to cause the needle of the galvanometer to oscillate. The

176 PR^CnCAL STANDARDS

battery contact was then made for an instant with the foot. The
slight kick given by the spot of light at once showed which way
the block had to be moved, without its being necessary to keep the
battery on long enough to heat the conductors sensibly.

The observing-room was kept at a very equable temperature
by a screen before the window, also the wire of the balance was
protected by a piece of boarding from the heat radiating from the
observer's body.

After every observation the temperature of the coil and the
water in the trough was read off, and if any difference was observed
between these readings and those first taken, the observation was
rejected and another one taken.

Four observations were made on each wire at intervab of from
twenty to forty minutes.

Before noting down the scale-reading all the connectors were
moved, and if no change in resistance was observed the connexions
were presumed to be good.

All results are given in terms of weight and length, as it is
impossible to measure the diameter of a small wire with the
accuracy with which the weight can be found; moreover, the
cross sectiQ^ of a wire is not generally a circle, and the mean
diameter varies slightly from point to point however carefully it
may be drawn.

A great oversight was made in not observing the specific
gravity of each wire, so that the results of the experiments now
made could be compared with former ones. This omission was
first iQade because it was thought that the results of former ex-*
periments could be used; but after several measurements had
been made it was found that the values of the specific gravity of
wires of the same metal, given by different observers, varied so
much that it was impossible to find the resistance of a wire of a
metal of which the length and sectional area are known, fit)m
the resistance of a wire of which the length and weight are
known, without taking the specific gravity of the wire actually
experimented on.

Silver,

Three silver wires were compared.

No. I. firom commercially pure nitrate of silver.
No. II. fix)m French coin.
No. III. fix)m English coin.

FOR ELECTRICAL MEASUREMENTS

177

The silver was first dissolved in nitric acid and then diluted
with water and precipitated by hydrochloric acid. The chloride
was then well washed, and afterwards fused with pure carbonate
of sodium. The resulting button of silver was fused a second
time with borax and a little nitrate of potassium ; lastly, before
casting, it was fused with a piece of charcoal floating on the top.
The mould was about 36 millimetres long by 4^ millimetres
diameter. The drawing of the wire was conducted with the ut-
most care. The wire was annealed only twice during the process.

In drawing all wires the end first entering the hole was
reversed at each successive drawing, after it had been drawn down
to about one-half its required diameter. The wires were twice
drawn through each of the smallest holes, the ends being reversed
as before.

To measure the harder wires they were straightened by rolling
them between two smooth boards, and then passed through a
thermometer tube of such a length that the ends just projected
fi-om the tube, the long ones being cut into two or three lengths
for the purpose. It was found that the wire could be pulled out
of the tube and reinserted many times without altering the length
by half one-tenth of a millimetre. Some care was necessary in
soldering the wires to their connexions. A small lump of hot
solder was placed in the terminal, and the end of the wire steadily
and slowly pushed into it until it set. Thus the boundary between
the wire and solder was well defined, and the wire could be cut off
at exactly the required point. The wires were weighed and
measured after the resistance had been taken.

The care taken in drawing the silver wires accounts for the
close agreement of the results. Anqther wire was drawn as rapidly
as possible through the latter holes to harden it, and a difference
of 3^ per cent, was found in its conducting power.

The results are given in the following Table : —

Wire No. I.

Temperature of ooil

2il
21-2
21*4
21*3
Length 1*5906 m.

a A.

Beading of bridge-scale { Temperatare of wire

88S
8S8
890
891
2^ro-point at 514*26.

21-3
21*3
21-4
21*4
Weight 2*9208 grammes.

178

PBACnCAL STANDARDS

No. II.

Temperatnre of ooil

Reading of bridge-aeale

Temperature of wire

18-8

194

19-3

19-0

199

19-4

19-3

204

19-5

19-4

206

19-6

Length 1-6749 m.

2^ro-poiDt at 514*25.

No. IIL

Weight 3-4419 grammoR.

18-6

840

18-2

18-8

855

18-8

19-3

870

19-2

19-8

880

19-5

Length 1*3692 m.

2^ero-point at 513-7.

Weight 2-1572 grammes.

Resistance of metre-gramme wire No. I. 1*0000

No. IL 0-9991
No. III. 0-9986

Copper.

Three copper wires were tried. The copper employed was
electrotype copper, and it was drawn without previous fusion.
The copper of wires Nos. I. and II. was prepared by Messrs De la
Rue & Co., that of No. III. wire as follows : — Sulphate of copper
was made by dissolving electrotype copper in pure sulphuric acid,
and twice recrystallizing : the copper was obtained from the sul-
phate thus prepared by electrolysis; it was precipitated on a greased
platinum pole, the other pole being of electrotype copper.

Wire No. I.

Reading of bridge-soale

Temperature of coil

23-4

23-6

23-7

23-8

Length 1-9324 m.

20-1
20-5
20-8
20-8
Length 1181-05 m.

21-6
21-8
21-8
22-0
Length 1*6187 m.

244
246
248
250
2^ro-point at 514

No. II.

198
217
221
223
Zero-point at 514.

No. III.

565
570
573
572-5
Zero-point at 514.

Temperature of wire

2! -2
21*3
21-3
21-4
Weight 3*9867 grammes.

19-9
20-2
20-4
20-45
Weight 1*4908 gramme.

20*8
21
21
21
Weight 2*7151 grammes.

FOB ELECTRICAL MEASUBEKENTS

179

Resistance of metre-gramme of wire Na L I'OOOO

No. IL 10005
No. IIL 10011

Gold.

Three gold wires were tried.

No. I. from Australian gold.
No. II. from English coin.
No. IIL from English coin.

The metal was first dissolved in nitro-hydrochloric acid, the
excess of acid was then evaporated off, and the salt largely
diluted with water to precipitate the chloride of silver. After
filtering the gold was precipitated by sulphurous acid, the pre*
cipitate collected in a small beaker, and washed four times with
hydrochloric and nitric acid alternately. After drying it was frised
with borax and nitrate of potassium and cast. It was again frised,
and finally cast in the mould.

Temperatare of ooil

20-2

20-4

20-8

Length 0*8854 m.

21-6
21-6
21-6
21-8
Length 0*9998 m.

19-8
201
20*5
20*8
Length 1*0211 m.

Wire No. I.

Beading of bridge-scale

849
849*8
851*6
852*5
Zero-point at 515*2.

No. IL

634*5
638
638
638
Zero-point at 515*2.

No. III.

782
788
784*6
797
Zero-point at 515.

Temperatare of wire

18*8
18*8
18*9
18*9
Weight 2*2200 grammes.

20*2
20*3
20*3
20*3
Weight 2*9021 grammes.

19*2
19*4
19*6
19*8
Weight 2*9753 grammes.

Resistance of metre-gramme of wire No. I. 1*0000

No. IL 0-9998
No. IIL 0*9995

12—2

>»

180

PBAGTICAL STANDARDS

Lead.

With lead very" good results were obtained. Five wires were
determined. The wires were pressed at a gentle heat, the press
being carefully bored and clecuied beforehand. As the wire came
from the press it was received on a smooth board. It was then at
once soldered on to the connexions and placed in the trough. The
solder employed was Wood's cadmium alloy. After being cut
from the connectors the wire was straightened by rolling between
two boards with great care; it was then placed on the board
beneath the beam-compass, adjusted to the groove below the line
of motion of the cross wires of the telescope, and carefully
measured and then weighed.

Wire No. I. was cut from a bar of commercially pure lead>
{)repared by Mr Baker of Sheffield.

Wire No. II. made from lead obtained by heating the acetate
thrice recrystallized. This specimen was kindly prepared by
Mr Mathewa

Wire No. III. from the acetate of lead of commerce twice
ciystallized.

Wire No. IV. from the acetate of lead of commerce three times
crystallized.

Wire No. V. from the seventh recrystallization of acetate of

«M^A. ■■■■■■■"»«^ ^» v^*«

Wire No. L

Jujvrjk*.

Temperature of coil

Beading of bridge-scale

Temperatare of wire

18*1
18-2
18*3
18-4
Length 0*4907 m.

355
362
362
362
Zero-point at 514.

• No. II.

17-5
17*8
17*6
17*6
Weight 2-0689 grammes.

16*4
16-4
18-0
181
Length 0*5100 m.

855
867
869 .
669
Zero-point at 514*5.

No. ni.

17*1

17*5 '
17*6
17*6
Weight 2*1320 grammes.

17-0
17-0
17*2
17*4 *
Length 0*4910 m.

746
748
748
748
Zero-point at 516.

161
16*2
16*2
16*3
Weight 1-9883 gramme.

FOB ELECTRICAL MEA80BBHBNTS

I8t

NalV.

Temperature of ooil -

Beading of bridge-soale

h Temperatore of wire -

17-2
17*6
17*7
17*8
Length 488*2 m.

525
529
530
535
Zero-point at 515*2.

15*3
15*3
15-3
15-4
Weight 1-9991 gramme.

No, V.

1 . - -

ia-8

18-8

19-1

19*6

Ijength 0*4915 m.

62a

628
. 634
640
Zero-point at 515-5.

« ■ •

! 17*8
17*8

la-o .

• 18-2 ' - -'-
Weight 2-0253 grammes.

Resistance of

metre-gramme of wire

No. I. 1-00000

II. 100045

m. 100029 .

IV. 1-00054

V. 100026

Gold-silver alloy.

No. I. Part of the alloy formerly prepared for the experi-
ments described in Phil, Mag. Feb. ISiSl, and there described as
wire No. I.

No. II. Part of No. VII. there described.

No. m. Part of No. VIII. there described.

No. IV. From' the first three* alloyB mixed a;nd refused and
drawn.

No. V. Alloy reprepared from the pure metals.

Wire No. I.

Temperatare of coil

Beading of bridge-eoale

Temperature of wire

17*8
18*4
19-8
20*0
Length 0*5374 m.

816-0
821*6
831*5
845-0
Zero-point at 517*6^

17*8
18-2
20-2
21*2
Weight 1*8607 gramme."

182

PRACnOAt STANDARDS

Tempenhtnxe of ooil

18*4

18-8

19*6

20-0

Length 0*4203 m.

18-8
19-0
191
19-3
Length 0*5472 m.

18-8
19-0
19-2
19*2
Length 0*6333 m.

No. IL

of bridge-floale

481-0
482*4
486-0
497-6
Zero-point at 517*3.

No. ni.

19-2

594-0

20-0

600*2

18-6

595*4

19-0

596-0

Tiftngth 0*3709 m.

Zero-point at 517*4

No. IV.

870
870
869
869
2iero-point at 514*5.

No. V.

542-0
541*6
541*4
541-8
Zero-point at 515,

Temperature of wire

18*6
18*8
19*4
21*3
Weight 1*2082 gramme*

17-8
19-2
18*2
18-5
Weight 0*9052 gramma

17*8
17-9
18-0
18-0
Weight 1*9199 gramme.

17-6
17*7
17*6
17*7
Weight 2*6497 grammes.

Results for Oold-silver alloy.

Resistance of metre-gramme of wire No. I. I'OOOOO

No. II. 0-99963
No. III. 100017
No. IV. 1-00036
No. V. 0-99996

»}

»»

Mercury.

Three tubes were filled with mercury and their resistance
taken.

Tubes Nos. I. and IL with distilled mercury treated with nitric
and sulphuric acid.

Tube No. III. with mercury distilled ifrom a specimen which
contained a small quantity of gold.

FOB ELECTRICAL MEASUBEMENTS

The lengths of the column are given below in their order.

183

TubeL

TnbeU.

Tnbe m.

mm.

383

291

245

384

288

242

390

289

240

386

287

240

389

288

242

384^

288

243

381

291

243

377

290

244

384^

292

246

392

288

246

399

288

248

405

289

248

407

288

252

407

288

253

406

288

254

413

290|

254

418

292

257

424

293

228

416*5

295

260

416

297

262

414

265

405

267

405

1 ^'
log-g^

1 ^'
log^

=1-9996018.

=1-9998710.

= 1-9995614.

Several other kinds of mercury were tried in one and the same
tube, and the resistances found to be the same within two or three
hundredths per cent.

Sixteen tubes were obtained, picked from a great number, and
of these the three best ones were taken. No. I. was not so good a
tube as the others, as the outside was uneven, rendering it im-
possible to calibrate it with very great accuracy.

To calibrate a tube it was taken and carefully cleaned with
pure nitric acid, and then with a solution of caustic potash. It
was then well rinsed with distilled water, and dried by passing a
current of hot air through a chloride-of-calcium bulb and then
through the tube. A small column of mercury was put in the
tube, and the length of column measured by the beam-compass.
The column was shifted along the tube by sucking up or blowing
through an india-rubber tube with a chloride-of-calcium tube

184 PRACTICAL STANDARDS

•

inserted between it and the tube to be calibrated. By this arrange-
ment the column could be adjusted with the greatest nicety to
the place in the. tube required. The lengths of the column were
taken at equal intervals from one end of the tube to the other.
The formula for correction used is given below.

Let C be conducting power of a tube of uniform bore and of
length and capacity equal to that of tube considered ; C observed
conducting power. Then

C'^C

1 '

where n is the number of measurements made, X the length of the
column of mercury in any position, the summation extending to
all the readings taken.

The ends of the tubes were ground by putting some emery
powder and naphtha on a slate table, holding the tube vertically
upright with the left hand, and with the right hand rubbing the end
of the tube in contact with the table round the circumference of a
small circle. Thus the end of the tube was made slightly convex,
the opening being at the apex of the convexity. To measure the
tubes they were placed under the beam-compass, and a stout pin
inserted partially into each end.

From the shape of the ends of the tube, the point where the
pin emerged from the tube could be exactly seen and the measure-
ment made with certainty. Many measurements were made
turning the tube round its axis through a small angle before each
measurement, and the mean of the lengths found taken for the
true length. To find the weight of the tube full of mercury it was
carefully cleaned, filled with mercury, and placed in a long narrow
trough full of pure mercury. The tube was held down by iron
weights, a thermometer inserted in the trough, and the apparatus
allowed to stand imtil the temperature was constant. After the
true temperature had been obtained the tube was taken out of the
trough and the contents weighed.

This was managed in the following manner. One operator
took hold of the tube by pressing a finger against each end and
lifting from the trough ; the little globules adhering to the out-
side of the tube were then rapidly removed by two assistants
^ith brushes.

FOR ELECTRICAL HEASUREMENTS

185

The mercury was then allowed to flow slowly out into a small
porcelain crucible and weighed. In this way pretty consistent
results were obtained if the tubes were cleaned before each
filling.

To determine the resistance of the tubes they were placed in
the water trough, with bent pieces of tubing fastened on to the
ends with india-rubber tubing and reaching above the sur&ce of
the water.

The terminals were of copper, well amalgamated. They dipped
into the bent tubes and came flat against the ends of the tubes
the resistance of which was to be determined. In the calculation
of the weight of mercury at 0** in the tube from the observed
weight, Regnault's value for the expansion was used.

Clonnexions of amalgamated platinum were first used, but did
not give good results. It was found that the amalgamation was
imperfect. The mercury adhering to the platiigijim. was rubbed off
against the ends of the tube, and the resistance varied with the
height of the mercury in the bent tubes. The platinum was
amalgamated by dipping it into a mixture of mercury and sodium
amalgam. The sodium was then oxidized and dissolved off by
dipping the platinum in a little dish of water and hydrochloric
acid. The terminal was then drawn through a dish of clean
mercury, so that the water floated off. The platinum was then for
the time beautifully amalgamated ; but the mercury soon drained
off when the plate was exposed to the air, and could be easily rubbed
off even when the platinum was immersed in mercury.

Tube No. I.

Temperature of ooil

181
18-4
18-4
18-8
o grms.

Wt, temp. 19-6 24-7021

„ 21-0 24-6930

„ 20-8 24-6960

Wte. reduced to 21" :—

24-6958
24-6930
24-6940

Beading of
bridge- scale

346-0
349-2
349-2
348-0
Zero-point 515

Temperatare of wire

18-2
18-6
18*8
18-0
Length 09365 m.

186

PRACTICAL STANDARDS

Temperature of coil

19-6

19-9

'20-0

20-0

Tube No. 11.

Reading of
bridge-Bcale

188-0
186-8
186-5
186-6

Wt at 21-2
21-6

gnns.
12-3140
12*3132

Temperature of wire

19-0

19-05
19-1
19-1
Length 0*6563 m.

19-0
19-0
19-1
19-15

Tube No. m.

633-5
633-1
631-7
631-3

Wt. at 22-2
22-2

»••

grms.

8-2894

8-2836

18-9
18-8
18-8
18-8
Length 0-5497 m.

Jtestdts,

Resistance of tube No. 1. 1-00000

No. II 0-99849
No. III. 100000

An approximate table is subjoined of the resistances of a
metre-gramme of the different metals in terms of the B. A. unit>
1864 :—

Copper 01469

Silver 01682

Gold 0-4150

Gold-silver alloy 1*668

Lead 2-257

Mercury 1306

From the foregoing results we may draw these conclusions : —
That ¥rith great care a unit may be reproduced with great
accuracy by any of the metals or alloys above mentioned.

Of those tested it appears that lead is the most preferable on
account of its easy purification, and because the presence of im-
purities, amounting to several per cent., produce no very dispro*
portionate effect on its conducting power. For instance —

>9
>l

FOR ELECrrBICAL MEASUREMENTS 187

(yonducting power of lead is 7*77

Of lead with 12"7 per cent, volume of tin is 813

10'6 „ „ cadmium ... 8*38

2-3 „ „ bismuth 70

3"8 „ „ antimony ... 7*1

2-3 „ „ silver 79

With the other metals and alloys tested a much greater
difference is found in the conducting power when such impurity
exists.

A few examples ¥rill show this.

The conducting power of pure copper is 100

Of copper with 1*6 per cent, in volume of silver 65

Of silver with 1-2 „ „ gold 59

Pure silver being taken as 100.

Of gold with 1*2 per cent, of silver 73*8

The conducting power of pure gold being 78.

If the conducting power of mercury is 10'9

That of mercury with 1*18 per cent, volume of lead is . .. 11'5

1-8 „ „ tin 11-8

18 „ „ zinc 12-4

0-7 „ „ gold 11-6

12 „ „ silver ... 11*6

The manipulation with lead is rendered easier by its high
resistance.

Mercury is easily purified, and good results are always
obtained with it. It would, however, in reproducing a unit, be
necessary to distil the mercury, because traces of such impurities
as silver and gold, which may easily get into it when in use in a
laboratory, cannot be removed by treatment with nitric acid.
The chief labour is in selecting and calibrating the tubes, and this
is very great.

The results obtained with the gold-silver alloy, even when
prepared by different persons, supposing great care used, give very
accordant results ; and for the easiness with which it can be made
it may be recommended for producing a unit.

Electrotype copper would appear a good substance. The
agreement of results obtained with commercial electrotype copper
with those obtained with copper prepared from pure salts shows
this.

188 PRACTICAL STANDARDS

The maximum difference of the conducting powers of electro-
type copper, as observed with ordinary care, is 1*6 per cent. Copper
is not, however, to be preferred, as great care and some practice is
necessary to draw a good wire.

The purification and drawing of pure gold and silver would, in
the hands of anyone but a chemist, lead to no good results, in all
probability. These facts being considered, we should prefer lead
for the reproduction of a unit. No doubt it would be well to use
two or three substances to check the results. For these auxiliary
substances mercury and the gold-silver alloy maybe recommended.
The choice between these two will depend on the appliances of the
individual observer. When thorough care is taken all the above
means ai*e equally good.

On forming an opinion on the difficulty of reproducing a unit
by chemical means it must be remembered that if anything like
accuracy is wished for; not only expensive and delicate apparatus
is required, but also very much time must be spent, and a great
deal of experience in the manipulation is required. The ex-
periments here described extended over about six months. Any
person wishing to reproduce a unit should bear these considera-
tions in mind, especially as it is the intention of your Clommittee
to cause coils to be issued representing a known resistance. That
copies of ia given resistance can be made to a much greater accuracy
than that to be obtained by chemical or other known means of
reproduction, and that coils can be compared by different observers
with different apparatus to almost any degree of accuracy (although
this fact has been brought into question by a former experimenter),
is proved by the following &cts.

The two units which have come into our hands, made by
Messrs Siemens and Halske from copies of the coil used last year
by your Committee for the determination of the absolute unit,
were compared against the standard coil and found to agree mth
it within two-hundredths per cent. Again, copies of Weber's unit,
one made of the gold-silver alloy, the other of German-silver, were
compared at the interval of two years by different observers with
different apparatus, and the results found to agree to one-half a
hundredth per cent.

It is from the fact that copies can be produced with almost
absolute accuracy, with a minimum of cost and labour as compared
with chemical or mechanical means of reproduction, that we seem

FOR ELECTRICAL MEASUREMENTS 189

quite justified in recommending all who wish to obtain a standard
to procure a copy of the British- Association unit, or any other in
general use. As copies of the British- Association unit are being
sold at a reasonable price by several of the leading instrument
makers, which, we are given to understand, will agree together
very closely, we confidently recommend the adoption of this unit.

And, in conclusion, we still adhere to the opinion, given in
Appendix C of the Report of 1862, that the best means of repro-
ducing a unit, for those who have not the opportunity of procuring
a copy, and who cannot afford the time and expense necessary to
reproduce the unit with great care, is to procure a given length
and weight of the gold-silver alloy, such as shall have been found
equal to the unit adopted (the quantity required being very nearly
0*5995 of a metre of a wire, one metre in length of which would
weigh a gramme) for the British-Association unit.

FOURTH REPORT— BIRMINGHAM, 1865.

The Committee have the pleasure of reporting that the object
for which they were first appointed has now been accomplished.

The unit of electrical resistance has been chosen and deter-
mined by fresh experiments ; the standards have been prepared,
and copies of these standards have been made with the same care
as was employed in adjusting the standards themselves; seven-
teen of these copies have been given away, and sixteen have
been sold.

The chief work of the Committee this year has been done by
Dr A. Matthiessen. Last year's Report announced the completion
of the experiments determining the resistance in absolute mecisure
of a certain coil of German-silver wire. Taking this coil as the
basis, Dr Matthiessen, assisted by Mr C. Hockin, prepared ten
standards, each expressing the British-Association unit of electrical
resistance ; two of these standards are coils of platinum wire, two
are of platinum-silver alloy, two are coils of wire drawn from a
gold-silver alloy, two are coils of wire drawn from a platinum-
iridium alloy, and the remaining two are tubes of mercury.

The wires employed in the coils are from 0*5 millim. to 0*8
millim. diameter, and range from one to two metres in length.
They are insulated with white silk, and are wound round a long
hollow bobbin of brass. The wires are imbedded in solid paraffin,
and enclosed in a thin brass case, which allows the coils to be
plunged in a bath of water by which their temperature may be
conveniently regulated and observed. Two short copper terminals
project from the case and are forked at their ends, so that they
may be connected with Wheatstone's balance in the manner re-
commended by Professor W. Thomson, avoiding the error due to
the possible resistance of connexions. The mercuiy standards
consist of two glass tubes about three-quarters of a metre in
length.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 191

These ten standards are equal to one another and to the British*
Association unit, at some temperature stated on the coil or tube,
and lying between 14*'*5 and 1(>**"5C.

None of them, when correct, differ more than 0*03 per cent,
from their value at Id'^'S C.

In the choice of the material of which the standards are con-
structed, the Committee have been much assisted by the experi-
ments on permanency made by Dr Matthiessen.

Silver and copper were found to alter in their resistance simply
by age. German-silver was also found to alter in some cases.

These materials had therefore to be rejected. Gold appears
constant; but owing to its low specific resistance a considerable
length would have been required, unless a wire had been adopted
of very small diameter. This was not thought desirable, for
several reasons: any slight decay or injury in the sur&ce of a
small wire would cause much greater alteration in the resistance
than the same injury to a large wire ; a small wire would be more
liable to mechanical injury, and would be much more rapidly
heated by the passage of currents. The Committee having rejected
small wires for these reasons, thought it unnecessary to incur the
expense of a large and thick gold wire. The great change of
resistance caused by a change of temperature furnished another
reason for rejecting gold and other pure metals. One pair of
standcuxls, however, was made of platinum*, which appeared the
most suitable of all the pure metals. Platinum and the three
alloys named appear all to be very constant — that is to say, their
resistance is not altered by age, or even by being subjected to
considerable heat and recooled.

These materials also possess considerable mechanical strength ;
they are not easily injured by chemical action, they have con-
siderable specific resistance, and the resistance, in the case of the
three alloys, changes little with a change of temperature.

It is of course impossible to say with certainty that their
resistance will not vary with time ; but it is most unlikely that the
resistance of all will vary in the same ratio. If, therefore, as is
hoped, the eight coils made of such different materials retain their
relative values, some confidence may be felt in the permanence of
the unit.

Some additional security is given by the power of reproducing

* See Report for 1906.

192 PRACTICAL STANDARDS

^he unit, if lost, by chemical means, or by fresh experiments on
absolute electro-magnetic measure, although neither of these
means at present appear to give such perfect accuracy as would
be secured by the permanency of a material standard. Fresh
absolute experiments of the kind described in previous Reports
would hardly reproduce the same value much within one part in
-a thousand; and Dr Matthiessen, as appears from last year's
Report, is not very sanguine of obtaining a better result than this
by chemical means. Thus a difference exists in Dr Siemens s and
Dr Matthiessen's reproduction of a unit by means of mercury, as
pointed out in last year's Report. It is of course probable that
differences of this kind will in time disappear; and Dr Siemens
fairly points out that the discrepancy mentioned in last year's
Report, between coils made from a very old and those made from
a new determination of the mercury unit, affords no criterion of
the accuracy with which mercury can now be used as a means of
reproduction. Dr Siemens was the first person who produced
numerous sets of coils accurately adjusted ; and although unable
to recommend the adoption of his unit of resistance, the Com-
mittee once more take an opportunity of expressing their sense of
the high value of Dr Siemens's researches on the reproduction of
units by means of mercury. Dr Siemens is confident that a unit
can be and has been reproduced by means of mercury with an
accuracy of 0*05 per cent.; but, meanwhile, the chief security for the
permanency of the unit consists in the preservation of standards
constructed in various ways aiKj of various materials.

The mercury tubes furnish an additional security. A molecular
change may occur in the wires, that is to say, they may become of
harder or softer temper; they may be injured chemically in course
of time by some action on their surface ; it is just possible that
the repeated passage of currents may alter them in some way»
although we have no reason as yet to expect such an alteration.

Mercury is free from all these objections. Its temper cannot
vary, and as it would be purified afresh on each occasion, it will
be chemically uninjured.

On the other hand, some fresh dangers may occur in its use.
The tubes themselves may alter in time, or the mercury may not
always be absolutely pure. Absolute security cannot be had ; but
the choice of a variety of materials will probably prevent any
serious alteration from occurring without detection.

FOR ELECTKICAL MEASUREMENTS 193

The copies which have been issued are similar in form to the
standard coils; but the terminals are simple thick copper rods,
intended to be dipped in mercury cups. The security given by
this mode of connexion is sufficient for all ordinary purposes, and
it was feared that the use of the double terminals might not be
everywhere understood. The platinum-silver alloy has been used
in all the copies. Wire made of this alloy is veiy strong and
ductile. It can, for instance, be drawn down to a diameter of
0*0002 inch. Its resistance is not permanently altered even by a
great change of temperature, and even annealing hardly affects
it. Moreover, the change in its resistance due to a variation of
1** Centigrade is at ordinary temperature only 0*032 per cent.,
being less than that of any other alloy tested. It is also a com-
mercial alloy, which has been long used by dentists; and Dr
Matthiessen points out, as a curious coincidence, that many com-
mercial alloys coincide with proportions indicating peculiar
electrical properties. Vide Appendix A.

The copies of the standard have been supplied for £2. 10«. in
boxes, with small mercury cups for the connexion, and with a
printed direction for use inside the box, stating the temperature at
which that particular coil is equal to 1 B.A. unit.

A satisfactory proof of the accuracy with which these coils
have been prepared was given by four independent observations,
by practical electricians not belonging to the Committee, of the
relative value of four distinct B.A. coils and four independent
standards issued by *Dr Siemens.

These four observations gave 10456, 10455, 10456, and 10457
as the measures of Siemens's standard, in terms of the B.A. units,
proving the accuracy both of Dr Siemens's work and that of the
Committee.

Twenty coils were to be distributed gratis, and seventeen have
actually been given away to the following recipients: —

The Directors of Public Telegraphs in

France. Spain. Prussia.

Austria. Italy. Sweden and Norway,

Belgium. Portugal. Russia.

India. Victoria.

Queensland. New South Wales.

Also to Professor Eirchhoff, Dr Joule, Professor Neumann, and
Professor Weber.

B. A. 13

194 PRACTICAL STANDARDS

Three remain for distribution. Sixteen have been sold. Dr
Faraday, on behalf of the Royal Institution, was the first
purchaser.

In distributing the coils, it was thought best not to give them
to institutions, where they would probably have laid on a shelf
useless and unknown, but rather to distribute them widely, where
they might become available to practical electricians.

The new unit has been actually employed to express the tests
of the Atlantic Telegraph Cable. Mr Varley promises that the
unit shall in future be the basis of the coils used by the Electric
and International Company.

Sir Charles Bright promises that the unit shall be exclusively
used by the British and Irish Magnetic Telegraph Company.

A standard has been supplied to the Royal Engineers at their
request. The head of the Telegraph Department in India has
introduced the unit, and there is little doubt that the British
Colonies generally will adopt it

More time will certainly be required to introduce it on the
Continent. The French Government has taken no steps to insure
its introduction ; but M. Blavier, the official editor of the Annales
Teligraphiques, has promised his cordial support to the Committee.
The Austrian Government has promised to use the coils experi-
mentally, and the German gentlemen to whom coils were given
have promised to give their best assistance.

Coils have also been bought by the managers of two large
telegraphic establishments in Switzerland, at Neuch&tel and
Zurich. There is therefore reason to hope that the unit may come
into extensive use.

When standard galvanometers, Leyden jars, and electrometers
are issued, all forming part of one coherent and necessary system,
it is probable that the B.A. unit will be found so much more useful
than any other as to supplant them entirely. Until these further
issues take place, it will only be adopted either by men who
can understand the advantage given by it in calculation, or by
electricians who feel confidence in the recommendations of your
Committee.

With a view to experiments which will allow of these further
issues of electrical units, a large electrodjmamometer has been
designed and is nearly complete. Graduated Leyden jars, with
air as the only dielectric, have also been designed and are nearly

FOB ELECTRICAL MEASUREMENTS 195

ready for use. An apparatus for the determination of the
quantity called v in Appendix C of the 1863 Report is in the
same condition. Prof. W. Thomson has for some time had ready
apparatus for absohite measurements of electrical effects, but his
connexion with the Atlantic Cable has suspended his work. Dr
Joule promises fresh measurements of the mechanical coefficient
of heat, and has only been delayed by the want of experiments
which other members of the Committee must previously complete.
In conclusion, the Committee are at last able to report one
positive result, but they feel that much more remains to be done.

Appendix A. — On the Construction of the Copies of the B,A. vnit
By A. Matthiessen, F.R.S., and Mr Charles Hockin.

The standard coil used in the experiments at King's College,
described in the Report of your Committee for 1864, was put into
our hands about last Christmas, in order that unit coils represent-
ing a resistance equal to ten million metres per second in Weber's
olectro-magnetic system might be made from it.

Since that time several unit coils have been made and issued.

We propose to state the method by which these coils were
made, and the reasons for choosing the particular alloy which has
been adopted for the conductor. The alloy referred to is composed
of 66 per cent, of silver and 33 of platinum.

This alloy possesses many properties which fit it for the use to
which it has been put.

As to its electrical properties : —

I. It alters less in electrical resistance with changes of tem-
perature than any other known alloy.

The importance of this point needs hardly to be enforced on
Anyone who has used resistance-coils.

The increment in the resistance of the alloy due to a change
of temperature from 0° to 100° C. is only 32 per cent.

n. The conducting power of the alloy is very low, and is about ,
one-half that of German-silver.

III. The conducting power of the alloy is not altered by
baking, that is by exposing it to a temperature of about 100° C.
for several days.

This is a property of great importance, for it has been observed
that those conductors which do not alter by baking, do not alter

13—2

196 PRACTICAL STANDARDS

by age either. The experiments by which this has been established
have been published in former Reports.

IV. The conducting power of a wire of the alloy is little
altered by annealing.

Further, the alloy does not oxidize by exposure to the air ; it
does not readily alloy with mercury; it makes a suflSciently pliable
wire, and can be drawn to a very great degree of fineness.
Dentists have made considerable use of it in consequence of its
good chemical and mechanical properties*. Of this alloy twenty
unit coils have been made and sent to several leading electricians
at home and abroad. The form of bobbin adopted for putting up
the wire, and shown in Plate 4, fig. 1, has been found very con-
venient, as it can be immersed in water during an observation.
The wire is twice coated with silk, and protected by being im-
bedded in solid paraffin.

Besides the coils already mentioned, ten unit coils have been
made, which will be deposited at the Kew Observatory.

Anyone possessing a copy of the B.A. unit may have it com-
pared at any future time against one of these coils for a small
pa3nnent.

Of the coils to be sent to Kew, two are of the platinum-silver
alloy, two of the gold-silver alloy, two of a platinum-iridium alloy,
and two of commercially pure platinum. Two mercury units have
also been prepared.

With so many coils for reference, made of such different
metals, it appears quite improbable that the unit now proposed
should be lost.

Along with the above-mentioned coils will be preserved the
standard coil used in the experiments first referred to, the coil used
in the similar experiments made by your Committee in 1863, and
several copies of these coils.

Of the coil called "June 4th" in the Report of your Committee
for 1863, two German-silver copies have been made. Of the other
coil used in 1864, two German-silver, two gold-silver, and one
platinum-silver copy have been made.

These coils have twice been recompared together at intervals
of three montlis, and will be again compared ; and if they are still

* Messrs Johnson and Matthey inform us that this alloy has been in use for
nearly twenty years.

Srituh Attodation Report Flate 4,

tjua iMv.

E XZ.XK^TZOX.

_ _ . - _ *"

r I

. \ . ' A » \

JV,'-

Tfft-^l

•I

1^'

t"i

1. - rn

(

-;- '■•f.'

vl<u'\1.A^

iVv*\ >

I 1

i '

^?:i^"i»>i ;

.-r '

"V>

Si/t

^£^in^viA.^ v^"«

■i

jf^-^^^

«kirfM

V.Vn !«^

FOR ELECTRICAL MEASUREMENTS 197

found not to have altered, they will be deposited at the Kew
Observatory for reference, their values being engraved on them.

The method adopted to obtain the unit from the standard
which had at a certain temperature a resistance of 4*6677 B.A.
units was this: —

Coils were made with the following approximate resistances,
viz.: —

Two coils nearly equal to ^ unit, called ^a and ^6.

1 unit, „ la „ 16.
One coil „ 2 units, „ 2.

„ „ 2i units, „ 2^.

The electrical balance used was that described in a paper on
the reproduction of a unit by chemical means, in the Report of
your Committee for 1864.

With this instrument two conductors, differing in resistance by
not more than 3 per cent., could be directly compared, and the
ratio found depended on to 0'0025 per cent,

Mumerous comparisons were made by means of this balance
between the following sets of coils, viz.: —

^a was compared with ^Ik

ia + i6

la.

lb.

la + 16

2 + io

2*.

2 + 2i> „

standard.

By taking the mean of several very concordant observations,
the value of the coil 1 a was found in terms of the standard, and
therefore of the unit, to a great degree of accuracy ; and from this
coil the first platinum-silver unit was constructed.

All the coils to be issued are recompared some weeks after they
are made, and rejected if they are found to have altered in
resistance by O'Ol per cent.

All the coils sent out are correct at the temperature written on
them to within 0*01 per cent., and this temperature lies between
14'5 and 16*6 in all cases.

FIFTH REPORT— DUNDEE, 1867.

The Committee have much pleasure in reporting that during
the past year considerable progress has been made, and that the
principal instruments required by the Committee for experiments
have been completed and are in use.

Experiments have been conducted by Dr Joule, having for their
object the determination of the mechanical equivalent of heat, by
observing the heat generated in part of a voltaic circuit, the re-
sistance of which was measured in absolute units by means of the
standard of resistance issued by the Committee.

Last year preliminary experiments of this kind had been made
by Dr Joule, and the agreement which he then reported between
his mechanical equivalent obtained by frictional experiments and
that obtained by the electrical method was so close as to lead to a
suspicion that it was partly fortuitous.

The experiments, which have this year been conducted with
every possible care, give 783 as the value derived from the
B.A. standard of resistance, while 772. is the well-known number
derived from friction.

The details of the experiments are contained in an Appendix
which accompanies this Report. Dr Joule states his opinion that
the electrical method has been carried out with greater accuracy
than the frictional method, assuming the B.A. standard to be an
exact decimal multiple of the absolute unit. The following
extract from Dr Joule^s Report will show the laborious nature of
the experiments. He says, "The last and most perfect series of
experiments comprise thirty for the thermal eflFect of currents in
the spiral, thirty for the eflFect of radiation, etc., and thirty for the
horizontal intensity of the earth's magnetism." Dr Joule expresses
himself willing to make a new determination by friction. Mean-
while the experiments already completed remove all fear of any
serious error, either in the number hitherto used as "Joule's
equivalent" or in the B.A. standard — a fear which hitherto, re-
membering the very discrepant results obtained by others, has
been very naturally entertained even by the Sub-committee from
whose experiments the standard was constructed*

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 199

In connexion with the measurement of resistances, Mr C. W.
Siemens has invented a simple and excellent contrivance, by
which the measurement of resistances can be made by persons
wholly unaccustomed to electrical experiments. They have only,
after the necessary connexions are made, to turn a screw till a
needle stands opposite a fiducial mark, when the resistance required
may be read directly on a scale with considerable accuracy. Mr
Siemens proposes to apply this invention to pyrometers, where
the resistance read will indicate the temperature, and the only
electrical connexions required will be the joining of the battery
wires to two terminals. Other applications of this invention will
doubtless arise, and extend the practical application of electrical
measurements. A full description of the instrument is contained
in the Appendix. Mr Siemens reports very favourably of this
instrument, which possesses considerable advantage in cheapness
and portability. Mr Siemens has constructed the instrument, and
made the experiments entirely at his own expense.

An instrument similar in object, and suggested by the above,
is also described by Mr Jenkin in an Appendix.

Mr Hockin has tested the constancy of the standard resistance-
units, with satisfactory results., except in the case of one mercury
tube. The exact result of Mr Hockin*s comparisons are appended.
He suggests that lead-glass was used for the mercury tube, and
that the glass may consequently have been injured by the nitric
acid used to clean it.

Mr Hockin has also made interesting experiments on the
construction of large resistances by the use of selenium. He finds
that resistances of one million units and upwards can be made of
this material, and that these artificial resistances maintain a
sensibly constant resistance at high temperatures, such as 100° C.
It is hoped that these very high artificial resistances will be found
useful in practice and much superior to those hitherto constructed
of gutta-percha or other insulators, which were of comparatively
little use in accurate work, owing to absorption, change of resis*
tance with temperature, and inconstancy when kept for any
considerable time. These valuable experiments have not caused
any expense to the Association.

The determination of a unit of capacity has occupied Dr
Matthiessen, Mr Hockin, Mr Foster and Mr Jenkin during the
last two years.

200 PRACTICAL STANDARDS

Very considerable difficulties have been encountered, and are
not yet wholly overcome. The methods by which both the elec-
trostatic and electromagnetic units can be determined, and
multiples or sub-multiples prepared, are sufficiently simple in
theory; but they assume that the condensers or Leyden jars
compared have really a definite capacity, and that with a given
electromotive force between the induction surfaces a definite
quantity of electricity will be contained in the jar or condenser.
This is very fiur from true with condensers of ordinary form.
Whether the dielectric separating the plates be glass, mica, gutta-
percha, paraffin, ebonite, or any other known solid insulator, an
absorption of electricity takes place; the longer the plates are
charged, the more electricity the condenser will contain, and, con-
versely, it will continue to discharge itself for a very long period
after the inner and outer armatures have been joined. With some
of the best insulators the effect will continue for hours, if not for
days. Condensers made with these solid dielectrics have there-
fore no definite measurable capacity. This capacity will differ
according to the time during which they have been charged ; and
it may also vary with extreme variation in the electromotive forces
employed, although this latter change has not been detected when
the differences of potential are such as between one Darnell's cell
and two hundred cells.

Only gaseous dielectrics appear free fix>m this embarrassing
peculiarity, called absorption, polarization, or residual charge.
One object of the Sub-committee has therefore been to construct
condensers in which air alone separates the induction-plates. But
new difficulties arose in canying this idea into practice. Some
support for each plate was necessary, and then leakage occurred
from one plate to another over the surface of any small insulating
supports employed, such as glass balls or vulcanite stems. It was
possible, by great care in drying the air, occasionally to make con-
densers of this type, which would remain insulated for a short
time, or even for some months; but long experience has shown
that an artificially dried atmosphere cannot be conveniently
maintained in any instrument which is not hermetically sealed.

Dust also accumulated between the plates of the trial con-
densers; this altered their capacity and increased the leakage
from plate to plate. Even a single filament of dust, by spring-
ing up and down between the two electrified surfaces, would

FOB ELECTRICAL MEASUREMENTS 201

oocasionally bring them to the same potential with great rapidity,
neutralizing the charge ; moreover a condenser of this type could
not be taken to pieces and cleaned, for no mechanical contrivances
could insure that the parts after cleaning would return to their
original position so exactly as to constitute a condenser of the
same capacity before and after the cleaning. It is therefore clear
that an air-condenser can only be constructed in an hermetically
sealed case, containing an artificially dried atmosphere ; and even
with these conditions, excluding the graduated and adjustable
condensers which were first tried, the air-condenser is not easily
constructed. For large capacities, which are alone useful in con-
nexion with practical telegraphy, the plates require to be so
numerous and large as to make the expense great and the bulk
veiy inconvenient.

It is hoped by the use of tin plates soldered to metal rods, and
supported on insulated stems inside a soldered metal case, that
these objections may be partly avoided ; but meanwhile practical
men have introduced condensers of a more convenient form,
overlooking the disadvantage which they all possess of ill-defined
capacity.

These condensers consist of sheets of tinfoil separated by
paraffin and paper, a preparation of gutta-percha, or mica — three
plans adopted by Mr Varley, Mr Willoughby Smith, and Mr
Latimer Clark respectively.

Condensers of this type have been made approximately equal
to a knot of some submarine cable; and the rough units thus
introduced are gradually creeping into use, although all elec-
tricians have been anxious that the Committee should issue a
more scientific standard. Under these circumstances, Mr Jenkin
has adjusted a mica-condenser, approximately equal to lO""
absolute electromagnetic units. The capacity of this condenser is
assumed as that which it possesses after electrification for one
minute, and is measured by the discharge through a galvanometer,
in the manner usually practised when testing the charge of a sub-
marine cable. The formula for obtaining the measurement in
absolute units from the throw of the needle is very simple, re-
quiring only observations of the time of oscillation, of a resistance
in absolute measure, and of a deflection of the galvanometer-
needle. All of these observations can readily be made, so that
their accumulated error cannot exceed one per cent.; and for the

202 PRACTICAL STANDARDS

present purpose this accuracy is sufficient, inasmuch as, when
using the condenser, small variations inevitably occur, arising from
the residual discharge. While, therefore, the new provisional unit
of capacity has no claim to a high scientific accuracy, it will supply
a practical want and introduce a unit based on the principles
adopted by the Committee, in place of the random measures
supplied by a knot of Persian-Oulf or Atlantic cable.

No decision has yet been arrived at whether the new unit shall
be issued by the Committee or on Afr Jenkin's own responsibility,
nor has the price been fixed.

The experiments by which it has been obtained are given in
an Appendix.

The practical applications of the standard of capacity are
important. It will allow the capacity of submarine cables to be
universally expressed in comparable figures, and may lead to im-
provement by the diminution of the specific inductive capacity of
the insulator, precisely as the introduction of units of resistance
has assisted the improvement in insulation and conductivity.

The electromagnetic capacity standard will also, by comparison
with the electrostatic standard about to be made, furnish one mode
of determining the constant called v in previous Reports, a number
of much importance in the theory of electricity.

The next unit or standard for consideration is that of the
diCFerence of potentials or electromotive force in absolute measure,
concerning which the experiments have been wholly in Sir William
Thomson's hands. He reports that he has at last succeeded in
constructing a series of electrometers capable of measuring dif-
ferences of potential ranging from ^jj of a Danieirs cell up to
100,000 cells, and that these measurements can all be reduced to
absolute units by comparison with one instrument of the series.

This class of instruments has been created by Sir William
Thomson, who year by year has produced electrometers each sur-
passing its predecessor, both in accuracy and delicacy; but
although those who have had practical experience of the admirable
results obtained by these have for the last two or three years
believed that the limit of excellence has been reached. Sir William
Thomson has not ceased to invent better and simpler forms,
until the instruments now supplied surpass every expectation of
practical electricians and furnish, indeed, a new engine for elec-
trical research.

FOR ELECTRICAL MEASUREMENTS 203

The chief difficulties encountered have been the insulation of
the Leyden jar, which has formed an essential part of all the
contrivances, its maintenance at a constant potential, and the
reduction to absolute measurement. In the present instrument
absolutely perfect insulation is no longer required ; for by a new
device for converting mechanical force into statical electricity
(first constructed by Mr Varley in 1869) Sir William Thomson is
able at any moment to replenish the jar by a few turns of a
handle, and, by a gauge electrometer, he can insure that the same
charge is constantly miaintained in the instrument. The difficulty
of the reduction to absolute units consists in the difficulty of
comparing the extremely small forces produced by electrostatic
attraction with the force of gravitation, and in the accurate
measurement of the extremely small distances which separate the
attracting surfaces. Sir William Thomson reports that these
difficulties have been overcome in his opinion, and that he will be
shortly in a position to construct and issue a simple pattern of an
absolute electrometer or gauge of potential which will serve as a
standard for general use.

Further experiments and tests are, however, required before
this can be done, as any precipitation would only injure the
interests- of the Committee. It is right here to mention that the
above experiments have been carried out almost entirely at the
expense of Sir William Thomson.

The replenisher, which is founded on the principle of the
electrophorus, may very possibly supersede the old form of elec-
trical machine entirely; it has some analogy with the electro-
magnetic machines lately invented by Mr C. W. Siemens and
Professor Wheatstone, by which intense dynamic effects are evolved
from the smallest initial trace of magnetism by the conversion of
mechanical force into electric currents, and was, indeed, suggested
by this invention to Sir William Thomson, who reinvented the
plan patented by Mr Varley*.

A modification of the same contrivance will allow the com-
parison of extremely minute quantities of electricity, such, indeed,
as might be accumulated on a pin's head ; by a series of rapid
inductions a charge is accumulated on the electrode of an electro-
meter, which may be made equal in potential to that on the pin's
head, but infinitely exceeding it in quantity ; the effect of this

* A Bimilar plan was proposed by Mr Nicholson in 17S5 {vide Phil, Trafu,).

204 PRACTICAL STANDARDS

charge in the electrometer can then be observed without difficulty,
and any increase or diminution in the quantity of electricity on
the pin's head or proof plane can be detected and the rate of loss
or increase observed The potentials to which various small
bodies are charged can also be observed by the same method, the
advanti^e of which consists in the fact that the original charge on
the body tested is undisturbed by the test, whereas by any of the
older tests the charge was altered by being touched by a proof
plane or by the electrode of the electrometer.

A similar plan has already been proposed by Mr Varley and
Sir William Thomson, with a water-dropping arrangement, but
the mechanical contrivance is in all ways preferable. No expense
has been incurred by the Committee for these instruments or
experiments.

Passing to the unit of current, the Committee regret that no
experiments have yet been made with the large absolute electro-
dynamometer constructed with the funds granted by the Royal
Society. Much difficulty has been experienced in finding a suffi-
ciently solid foundation in London, and probably the instruments
must be moved into the country for accurate use.

A portable electrodynamometer has been constructed which
will be suitable for distribution as a standard instrument* It can
be compared with the large absolute instrument, and can also be
compared directly with the most sensitive astatic galvanometers
yet made, as has been already proved by experiment. These in-
struments cannot be distributed until further experiments on their
constancy have been made.

Sir William Thomson, at his own expense, has also constructed
an electrodynamometer for absolute measure. His results will
check those obtained in London, and the portable standard will
also be tested by being sent backwards and forwards between
Glasgow and London, to be compared alternately with the absolute
instruments.

The determination of v, the ratio between the electrostatic
and electromagnetic units, is also an object pursued by the
Committee. Sir William Thomson has made preliminary experi-
ments, and has obtained numbers for this constant by the aid of
the absolute electrodynamometer and the absolute electrometer
already named. The number he has obtained differs so consider-
ably from that hitherto received that he prefers to extend his

FOR ELECTRICAL MEASaRBMENTS 205

experiments before publication. The same remark applies to the
measurement of the electromotive force of a Daniell's cell made
by the absolute electrometer.

It is hoped that the present Report contains satisfactory evi-
dence that valuable work is being done by the Committee, and
that the sums of money liberally granted by the Association have
been expended on proper objects.

It will be seen that these grants have stimulated further
expenditure on the part of more than one member; and thanks
are also due to the Electric and International Telegraph Company,
for the liberality with which they have lent large batteries, there-
by saving much expense. The Committee are willing to be
reappointed, and require no grant of money for the ensuing
year.

Appendix.
I. On a "Resistance-Measurer" By C. W. Siemens, F.R.S.

For the measurement of small resistances the method formerly
employed was that of the tangent galvanometer, which ^method is
still valuable in the determination of resistances which are in-
separable from a diflference of electric potential, such, for instance,
as a galvanic element.

In measuring wire-resistance more accurate and convenient
methods have been devised, amongst which that of the common
differential galvanometer and that known as Wheatstone's balance
hold the most prominent places.

But both these systems have disadvantages which render them
insufficient in a great many cases. For instance, in the first
method a well-adjusted variable resistance coil is necessary, which,
if the method is intended to be applicable between wide limits,
will have impracticable large dimensions. The bridge method,
though very beautiful, requires three adjusted coils, and frequently
gives rise to calculations which render it unavailable for unskilled
operators. The sine method, which is the most suitable for
measuring great resistances, requires even a superior amount of
skill and mathematical knowledge on the part of the operator.

Many years' experience of these methods made me feel the

206

PRACTICAL STANDARDS

want of an instrument which would, by its simplicity of construc-
tion and ease of manipulation, be capable of employment by an
unskilled operator with a degree of exactness equal to that of the
bridge method.

The conditions upon which such an instrument could be
successful appeared to be the following: —

1. The employment of a zero method, by which the galvano-
meter-needle should always be brought to the direction of the
magnetic meridian or the same given point upon the scale, and
therefore be independent of the unknown function of the angle of
deflection.

2. The readings to be made upon a simple lineal measure
divided into equal parts signifying equal units of resistance.

3. The employment of a single and unalterable comparison-
resistance.

The apparatus constructed to fulfil these conditions is repre-
sented by the following diagram: —

Two equal and parallel helices, h and h, are fixed upon the
common slide 88 , which moves in the direction of its length
between guide rollers. This motion is effected by the end ^,
armed by a facing of agate, which presses against the face of the
metal curve cc'4 The latter is fixed upon a slide moving in a
groove in the rule dd\ at right angles in the direction dd\ by

FOR ELECTRICAL MEASUREMENTS 207

means of a milled hecid % on the axis of which is a pinion gearing
into a rack underneath the straight edge of the curve cc\ The
rule dd' is graduated in equal parts ; and opposite to the divisions
is a nonius up the straight edge and the curve, to divide each
degree into ten parts. Whenever the milled head %, therefore, is
turned, the position of the curve is altered ; and as the point s of
the bobbin-slide is pressed against it by means of a spring, the
bobbin follows it in all its movements.

The wires of the two bobbins are connected together, in the
common point a, with the pole of a galvanic battery e, the other
pole being connected with two resistances ?•, x, and through these
with the remaining end of the galvanometer-helices. The re-
sistance r is made constant, and adjusted so that when a = 0 the
index of the curve stands exactly opposite the zero of the
graduated scale dd', the unknown resistance being represented
by X,

It is evident that, the resistance in the bobbins being equal, as
also their dimensions and initial magnetic effects upon the needle
suspended between them, if we make the resistance x equal to r,
the current in the two branches will be equal, and the magnet-
needle therefore balanced between them only when the helices are
equally distant from it. Should, however, either of these resist-
ances preponderate, the strength of current in that branch will be
lessened ; and in order to reestablish the balance it will be necessary
to shift the bobbins, approaching the one in which the weaker
current is circulating towards the suspended magnet.

The instrument is erected upon a horizontal metal table
standing upon three levelling-screws. The bobbin, the suspended
magnet, and dial plate for observing the zero of the pointer are
contained in a glass case, supported by four brass pillars. The
instrument is supplied with terminals for the battery-connexions,
and a current-breaker for interrupting the battery-circuit. Op-
posite to these are four terminal screws for receiving the ends of
the resistances r and x, with contact-plugs between them, in order
to quickly establish a short circuit in case the operator should be
in doubt towards which side he has to move the adjusting-
curve. Two constant resistances accompany the apparatus, i- —
that which is used during the measurement, and a, a resistance of
known value, which is introduced between the terminals x in order
to enable the operator for his own security to make a control

208 PRACTICAL STANDARDS

measurement by which he may verify the accuracy of the instru-
meDt at any time. Another purpose of this resistance is to
facilitate the readjustment of the zero-point, in case the galvano-
meter should at any time be cleaned or a new silk-fibre put in.

In constructing the sliding-curve of this instrument, it might
be determined by calculation from the formula given by Weber for
the deflection produced by a circular current of known magnitude
upon a magnetic point, and from the given distance of the coils
from each other. I prefer, however, in practice to determine the
curve of each separate apparatus empirically, because it is not
possible to coil a helix mathematically true, or to set it when
coiled absolutely at right angles to the plane of its horizontal
motion.

In the determination of each curve I use a delicately adjusted
rheostat or scale of resistances in the circuit of x, giving it varying
values corresponding to the equal divisions of the engraved scale,
and constructing the curve according to the position which it is
found necessary to give to the point b' in order to arrive at the
magnetic balance. With each instrument it would be possible to
have two values of r — one expressed in mercury and the other in
B.A. units ; and in order to measure at pleasure in either of these
units, it would only be necessary to insert the one or other between
the terminal screws for r.

The instrument has been found to be very convenient for the
measurement of the wire-resistances of overland lines, or for the
reading of resistance-thermometers ; it reduces the operation and
the observation of the zero position of a needle, and the reading
upon a graduated scale, which can be performed by a person of
ordinary intelligence without experience in electrical measure-
ment. In accuracy and range it equals the bridge method, while
as regards portability and cheapness of apparatus the advantages
are decidedly in its favour*.

* I haTe lately oonstrncted the same instrament on this principle with a circnlar
instead of a straight sliding-pieoe, whioh gives the advantage of a longer graduated
scale in the form of a circle. The circular sliding-curve is adjusted by radial set
screws in a solid ring working in a V-groove round the galvauometer.

FOR BLECtRIOAL MEASUREMENTS

209-

U. On a Modification of Siemens's Resistance- Measurer,

By Flebming Jenkin, F.R.S.

The following method of measuring resistances was suggested .
to Mr Jenkin by the above invention of Mr Siemens : —

Let two tangent galvanometer-coils of equal magnetic moment
be fixed together at right angles, with a short magnet hung in
their centre, having a long light index pointing at a fiducial mark
when the needle is in the magnetic meridian. Let the battery
and coils be so joined that the current shall divide in the ratio of
the resistances in the two coils, and shall pass in such a direction
as to tend to turn the needle in opposite directions*

The dotted lines show the positioii of the ooils when the corrent is passing*

Let one coil with a resistance R at the beginning of the
experiment stand in the magnetic meridian, and the other coil
with a resistance i2i in a plane perpendicular to the meridian;
and when the current is passing in such a direction that R tends
to turn NS in the direction of the arrow, let the coils be turned
till the needle is again brought to the fiducial point and the coil
Ri makes an angle <f> with the magnetic meridian, then we have
R =s tan <f>Ri; for the force exerted by the coil Ri to deflect the
needle in the direction of the arrow will then equal m sin ^ ; the

& A.

210 PRACTICAL STANDARDS

force exerted by the coil Ri to deflect the needle in the opposite
direction will be vh cos <f> ; and we have m sin <f>^mi cos <f>, or

- - s» tan <t>, where m and iiii are the couples experienced by the

magnet under the action of the two coils; but as we have
supposed these coils to have equal magnetic moments with equal

currents, — =»r> , therefore i2 = tan6i2i. -B ^nd Ri need not
m Ml

be the resistances of the galvanometer-coils only, but may consist
of two parts, 0-hr and (?i H- ri, where 0 and Oi are the resistances
of the galvanometer-coils, but r and ri are added resistancea Thus
when O, Gi and a* are known, Vi can be obtained by a simple ob-
servation.

If 6 + r be one, one hundred, or one thousand units, the
resistance of Vi will be equal to the tangent of <f>, or to one
hundred or one thousand times that tangent respectively minus
in each case a constant = &i.

If the range of the instrument were not required to be very
great, the coils would be turned by the pushing of a straight slide,
equal divisions on which would correspond to equal increments of
the tangent of <f>, and the scale would be numbered, so that the
resistance r^ should be read off directly, as in Mr Siemens's instru-
ment.

The tangent coils should be made of German-silver wire,
and might be an.'anged as practised by Helmholtz and Oaugain.
Theoretically, the range of each instrument would be infinite,
1.6. any instrument would be capable of measuring an infinitely
small or infinitely large resistance; but clearly the resistance of
G + r should be so arranged in each case that the angle observed
was not very different fi:om 45''. The range of the instrument
may be further increased by the use of elements.

FOR ELECTBICAL MEASUBEMENTS

211

in. Comparison of B,A. Units to be deposited at Kew

Observatory*. By C. HocKiN,

The following Table shows the value of the different copies of
the B. A. units that have been made for preservation at Kew : —

Material of coil

No. of
eoil

Date of obserraiion

Temperatures

at wbioh coil

has a resistaDce

= 1(F-

Obsenrer

Platinum-iridium alloy ...

Platinum-iridium alloy...
Qold-silver alloy

10
58

II.
IIL

January 4, 1865
June 6, 1865
February 10, 1867
January 4, 1865
June 6, 1865
February 10, 1867
January 5, 1865
February 10, 1867
April 10, 1865
June 6, 1865
February 10, 1867
January 7, 1865
August 18, 1866
February 10, 1867
January 7, 1865
August 18, 1866
February 10, 1867

(February 15, 1865
March 9, 1865
February 10, 1867

(February 2, 1865
July 18, 1866
February 11, 1867
February 3, 1865
August 18, 1866

(February 11, 1867
February 11, 1867

U'6 C.

160

16-0

15-3

15-8

15-6

15-3

157

15-7

15-5

15-7

15-2

160

16-7

14-8

17-9

C. H.
A. M.
C. H.

c. a

A.M.
C. R
A.M.
C. H.
A. M.
A.M.
C. H.
C. H.
A. M*
C. H.
C. H.
A.M.
C. H.
C. H.
A.M.
C. H.
A. M.
A. M.
C. H.
A.M.
AM.
C. H.
C. H.

Gold-silver alloy

Platinum .,

•
Platinum

Platinum-silver alloy

t Mercury

Mercury

Mercury ^

* Farther refeieDoes to these eoilB are made in the lleports for 1883 and 1908.

t The alteration of this coil, observed on Febmaty 11, 1867, is due, no doubt,
to a defect observed in the glass tnbe.

The tube was of lead-glass. Perhaps the strong nitric acid used to dean the
tube attacked the glass. A new mercury unit (No. III.) was made in consequence
of this defect.

The apparent alteration in the platinum-iridiam coils firom the first value found,
I believe to be owing to a clerical error. No alteration has been observed in them
since the second observation made by Dr Matthiessen in Jane 1865.

The values given in the above Table are deduced from the Qerman-silver coil
called B, used in your Committee's experiments in 1864. This coil was foand (by
comparison with copies made, in 1864, of gold-sUver, German-silver, and platinum*
silver) not to have altered. The coil B was also compared with the coil (Jane 4)
nsed in 1863, and the ratio of the two coils was foand not to have altered.

14—2

212 FBACTIGAL STANDABDS

IV. Experiments on Capacity, By Flebminq Jenkin, F.R.S.

The capacity of a condenser made of mica and tinfoil was
adjusted so as to be approximately eiqual to lO""^^ electromagnetic
absolute imits, according to the following experiments. The
capacity of any condenser can be directly measured in absolute
o^e^bsure by the following formula, appljdng to the eflfect of a
single discharge from the condenser through a galvanometer: —

{vide Report, 1863, Appendix C, p. 110), where R^ is the
resistance of a circuit in which the electromotive force used to
cheurge the condenser would produce the unit deflection, while
i is the angle to which the needle is observed to swing from a
position of rest, and t is half the period or time.. of a complete
oscillation of the needle of the galvanometer under the influence
of terrestrial magnetism alone.

This formula, which is analogous to that for any ballistic
pendulum acted upon by a known impulse;' supposes that the
Jirhole impulse is given in a time very short as compared with t*
and it also supposes that the deflection i is unimpeded by friction.

I employed a Thomson's astatic reflectiqg galvanometer with
double coils pf German-silver wire. The oscillations, with the
usual mirror and magnet, subside so rapidly that t cannot be
measured with accuracy, and i is very sensibly affected- by the
resistance of the air ; to obviate this I attached a brass ball to
the lower magnet of the galvanometer, weighing 55 grains*

A single floss-silk fibre can just support this weight, under
which it continues to stretch sensibly for about three days. In
order that the discharge from the condenser, electrified by fix)m
20 to 30 cells, should have force to move this heavy ball through
a sensible angle, the galvanometer was made highly astatic ; and
then I found that with even a single cocoon fibre the needle did
not return to zero within three or four divisions of the scale for
some minutes, exhibiting a kind of viscosity. The floss-silk fibre,
though much weaker, gave a very constant zero. The value of t
with the weighted needle seldom differed much from 20 seconds,
and the times could be observed for 10 or 11 minutes, during

•

* The ball, two magnets, mirror, and connecting bar, forming the whole
saspended system, weighed 67^ grains.

FOR SLECTRICM. MEASUREMENTS SIS

^hich time t was found to remain sensibly constant. As theiB
was no difficulty in observing the times of oscillation within one
second, it may be said that the observed value of t was correct
within one part in 500. Greater accuracy was not required, aJ3
the possible error from other sources considerably exceeds this.
Twenty DanielFs cells were used to charge the condenser, and the
discharge observed was about 180 divisions ; but the observations
were recorded within a quarter of a division : as this is done by
estimating the position of the reflected spot stationary, between
the two black lines of the scale for an almost insensible time, it
would not be right to assume that the deflection t is observed
with greater accuracy than one part in 400. When the spot of
light returned after making one complete oscillation, the diminu-
tion in the deflection was from 10 to 12 divisions; one-quarter of
this amount was therefore added as correction in each case to
the deflection observed. The resistance of the whole circuit was
.composed of the battery resistance, that of German-silver re-
sistance-coils, and of the German-silver coils in the galvanometer;
no considerable variation could therefore occur except in the
battery, which formed only a small portion of the total resistance.
The coils (adjusted by Mr Hockin) are probably correct within
one part in a thousand, and the measurement of the galvanometer-
coils is equally well known.

From what has been said, it might be expected that the
capacity of any condenser could be obtained with an accuracy of
one part in 400 or 500 at least; but successive discharges were
occasionally found to differ by as much as two divisions, though
this amount of discrepancy was rare. It was due partly to the
residual effect of former charges in the condenser (though great
core was taken to avoid this), partly, it is believed, to slight
changes in the electromotive force of the battery (which .was not
in very good order, the discharges being generally less toward the
.end of a set of experiments), and partly to slight motion of the
needle at the moment of taking the discharge. This last source
of error made it impossible to make the observations in London ;
even in the country the needle was seldom, if ever, absolutely still,
though the oscillations were generally less than one division. The
•variation of the electromotive force and resistance of the batteiy
when taking a permanent deflection was another source of error.
Owing to the great inertia of the swinging parts, no observation

214 PRACTICAL STANDARDS

could be taken until the current had been flowing for at least a
minute, and often more; and, especially when small resistances
were used, the deflections visibly diminished with time. Owing
to all these causes, I do not depend on the results obtained as
certainly accurate within less than one per cent. This is the less
to be regretted, as the capacity of a mica condenser is veiy ill
defined within wide limits, owing to absorption.

The condenser used consisted of 38 plates of mica, about
0-003 in. thick, and having a circular piece of tinfoil 3 in. in
diameter cemented to each side of the mica, with a piece of each
tinfoil projecting beyond the mica so as to join all the upper
tinfoils and all the lower tinfoib together, and form the inner and
outer armature of the condensers. This plan has for some time
been practised by Mr Latimer Clark and makes a very constant
and well-insulated condenser, extremely easy to adjust roughly by
altering the number of the mica plates, and for small corrections
by cutting away portions of the tinfoil from the top plate. Mica,
like all other soUd dielectrics with which I am acquainted, ap-
parently absorbs electricity to a very large extent, and continues
to do so for a long time, discharging it at first rapidly, but at the
last very slowly indeed, so that a complete discharge is not effected
for hours. The total capacity of the condenser varies therefore as
the time varies during which it is charged, and the apparent
discharge varies with the time during which we measure it; for
instance, if we merely observe the discharge due to a momentary
contact, we shall obtain a different result firom that given when
we maintain the contact all the time the needle is swinging ; the
result will also vary in the latter case with the time of oscillation
of the galvanometer-needle. If the needle oscillates slowly, it will
be acted upon by a greater quantity of electricity than if oscillating
rapidly. Thus, in one experiment, the deflection, when the dis-
charging contact was permanently maintained, was 166 divisions ;
when a momentary contact was made by a blow it was only 156.
When the contact was made for about 1*7 second the deflection
was 161, and when the contact was maintained for 3*4 seconds the
deflection was 164 ; the maximum deflection of 166 was reached
after 5 seconds : these experiments show that when the needle had
travelled two-thirds of its maximum distance, the current being
discharged exercised a very sensible influence on the deflection.
The ballistic formula is therefore not strictly applicable to a case

FOR ELECTRICAL MEASUREMENTS

215

of this kind, and a different result would be obtained with a
galvanometer oscillating either more or less quickly than the one
I used. It seemed therefore unnecessary to take great precautions
or to aim at any high degree of accuracy; and my object has
simply been to provide a unit for cable-testing which shall be ap-
proximately equal to the ideal standard chosen by the Committee,
and which can be used with at least as great accuracy as those
copies of knots of Atlantic or Persian-Gulf cables hitherto used.
The value of iZj, in the formula given at the commencement,
was found by two methods, which we will call the indirect and
direct method. In the indirect method three sets of resistance-
coils (a, b, c) were arranged as in fig. 1, with a battery, B, a

Fig. 1.

Fig. 2.

galvanometer, &, and a shunt, Z, equal in resistance to -g^ of the
galvanometer-coils. The resistance c was made equal to 1000 units,
and the resistances a and b adjusted until a convenient deflection
was obtained on the galvanometer; the resistance a was next
changed to Oi, and b was then altered to &i, so as to give the same
deflection as before on the galvanometer 0. Then calling d the
deflection observed, 0 the resistance of the galvanometer, we have

Ri^ nd

(b + c-{--0\(bi + c +

<^-^) TTTTc —

a formula for which the resistance of the battery need not be
calculated (n » 1000).

^1^ PRACTICAL STANDARDS

The second or direct method of obtaining Ri was, first, to
calculate the resistance of the battery B by the following formula
(fig. 2): — h and / are variable resistances ; g the resistance of the
shunted galvanometer, = 47'2 in my experiments ; break the
circuit at /, and adjust h till a convenient reading is obtained ;
then join /, as shown in the sketch, and adjust / and h until the
same deflection is obtained as before; then, calling hi the last
resistance at h, we have

Secondly, a direct deflection d was obtained with a resistance
k in circuit ; then Ri = nd(k -\- B + g).

The following is a record of the experiments made in chrono-
logical order : —

September 22. Discharge, — Values of i after charging for one
minute with 20 cells : —

Mean.

167

166

J66

166

Adding 2*5 to compensate for friction of air, i = 1685 ; and the
angle being very small, sin ^i«= 84*25.

Test for insulation ; discharge after one minute's insulation 154.

Times, — First four oscillations, the spot crossed the central
point in the same direction at

0' 35'',

o'55", ri4r, 1'

33"*;

last four oscillations,

y 13",

9' 32", 10'

10", 10' 29".

Total number of oscillations 31. Mean value of 2t =

19"15.

Valiw of Ri. Indirect method :■

—

ai.

h, &|.

Ri , Ohms.

1....8000

10000

1000 649

1000

276i

617 X 10»

.2.... 6000

8000

1000 675

1000

354^

617 „

3.. ..8000

10000

1000 647

1000

274i

512 „

4..*.6000

8000

1000 674

1000

355J^

6-18 „

Mean value of Ri in absolute measure = 5'16 x 10^. Value of
S = 99-53 X 10-".

* [Note added in 1911. Apparently the 29th vibration was missed and the times
given for the "last four oscillations*' are the times of the 27tb, 28tb, 80th, and Slst.
If so the first pair of times are those of 27 oscillations and give a mean period of
11;"' 17 and the last pairs are those of 28 oscillations and give a mean period of
19"-18. Mean of all=19"-16.]

FOR BLECTBICAL MEASUREMENTS 2L7

47 ^l Mean value of 5=-488.

Vaiue of Ri. Direct method. Battery resistance : —

. /. ht hi, . g, B,

1.... 2 18700 30 47 4841

2....10 18000 300

Deflection with variable resistance in circuit : —

d. k, B. g. fi. Ri, Ohms.

1....226i 22000 488 47 1000 610xl0»
2....310i 16000 488 47 1000 513xl0»

Mean value of i2i = 5*125 x 10" absolute units. Value of S from
values of t and i as above = 100'21 x lO"".

September 24. Dist^rge, — sin ^t« 84-75. JBi from indirect
method : —

a. Oi, h, ^ h^, ... ^v ^» ^u Ohms.

1....6000 8000 ioOO 575 JLOOp 364 6-18xl0»

2....8000 10000 ioOO 648* IOOO 275 516xl0»

Mean value of i2i in absolute measure = 5*17 x 10*. Assuming t as
on September 22, S =99*92 x 10-'«.

The box holding the condenser was now filled. np with an
insulating composition.

October 13. Discharge. — 184 divisions, 12 divisions lost on
return, sin ^i = 93*5. Discharge after one minute's insulation
181 divisions.

Time. — First four oscillations,

o'30'', o'sr, r ir, r3r;

last four oscillations,

10' 4'', 1^23", 10'43i", 11' 5"-
Total number of oscillations 31. Mean value of 2t = 20'H7.
JBi by indirect method : —

a. ai, b, &i« c. d.

Ri'

8000 10000 1000 .646 1000- -- 332

\ 6-19 X 10».

Value offlf=98*42x 10-".

Ri by direct method. Battery resistance : —

f. A. hi, g.

10 17400 700. 47

223^

Direct deflection : —

■ • •

d. It. B, g, R.

• • •

"1....270^ 22000 223^ 47 1000

"6-OlxlO^

2....331 18000 223^ 47 1000

6-06 X 10»

TVfean value of iJj = 603 x 10»« absolute

units. Value of

i8f« 101*03x10-".

218 PRACTICAL STANDARDS

October 16. Discharge : —

L 2. 8. 4. 6, 6.

186 185 184^ 184 184^ 184{

Mean 184-6, adding 3 for air, sin ^i « 93-8.

Times.— Fiist four, 0' 23'', (T 42 J", missed, 1' 24" ; last three,
r 55'', 8' 16", 8' 35"; 24 oscillations in all Mean value of
2««20"-66*

Independent series of observations divided into triplets : —

first two, 0' 22i", 1'24", last two, 9' 37^", 10' 39";

80 oscillations in alL Mean value of 2t » 20"'55.

Value o/Ri. Direct method. Battery resistance : —

1 223

2 216

Mean 219

Direct deflection : —

d, k. B» g. n* JRi, Ohms.

1.,..278 22000 219 47 1000 619 xlO^
2....321^ 19000 219 47 1000 619 xlO^

Mean value of Bj in absolute units = 6*19 x 10". Value of
i8f« 99-2x10-".

October 17. Discharge: —

1. 2. 8. 4. Mean.
179 180 179 180 1796

8in^i»91^.
Times: —

0'55", l'56i", 10' 7i", ir8i".

Total number of oscillations 30. Mean value of 2t = 20"'46.
VaiiLe of Ri. Direct method. Battery resistance : —

1 210

2 223

Mean 2165

Direct deflection :—

d. k. S, g, n. i?,, Ohms.

1....268 22000 216^ 47 1000 5-97 x 10»
2.**«329 18000 216J 47 1000 6-01xl0»

Mean value of -Bi « 5*99 x 10*» absolute units. Value of S « 99-26.

♦ [Note added in 1911. It appears that 7' 66"— (/ 23", and 8' 16"--0' 42^ were
each the period of 22 oeeiUationB, while 8' 86"—!' 24" was the period of 21 osdl-
lationa. These giye as the mean for one oscillation 20"*56.]

FOR ELECTRICAL MEASUREMENTS 219

The seven values obtained for S give a mean value of
0*9965 X 10~" as the capacity of the mica-plate condenser when
charged for one minute, and measured by a discharge through
a galvanometer, on the needle of which it acts for about 5 seconds.
If we reject the two observations made on Oct. 15 (which were,
indeed, only preliminary, and made with less care than all the
others) we find the average to be 0'9962 x 10~", and the approxi-
mation between this mean and any single results is 0*42 per cent.
It is therefore probable that a unit copied firom this preliminary
standard will not be one per cent, wrong.

A tenfold multiple (10~~*' absolute measure) of the condenser
measured is a convenient magnitude as a practical unit of capacity
for telegraphy ; thus the capacity of the Atlantic cable per knot
thus measured is 0*3535. Assuming that the practical unit of
electromotive force will be chosen as that multiple which is most
nearly equal to Daniell's cell, ue, 10* electromagnetic units, then
the capacity of the proposed practical unit is such that it contains
with the unit E.M.F. the same quantity of electricity as would be
passed in one second through a circuit of the resistance of one
Megohm. Thus 10* E.M.F., acting on a circuit of 10'^ will pass
in one second 10~^ absolute units of quantity; and, similarly,
10" E.M.F. will charge a condenser of absolute capacity equal to
10~" with 10~* absolute units of quantity. This practical series
of units is that which, in the opinion of Mr Latimer Clark and
myself, is best adapted for practical use in telegraphy. Mr Clark
calls the unit of quantity thus defined (10~*) one Farad, and
similarly says that the unit of capacity has a capacity of one
Farad, it being understood that this is the capacity when charged
with unit electromotive force (10").

V. Report on Electrometers and Electrostatic Measurements.

By Sir Wm. Thomson, F.R.S.

§ 1. An electrometer is an instrument for measuring differences
of electric potential between two conductors through effects of
electrostatic force, and is distinguished from the galvanometer,
which, of whatever species, measures differences of electric poten*
tials through electromagnetic effects of electric currents produced
by them. When an electrometer merely indicates the existence

^220 JTRACTICAL STANDABDS

of electric potential, without measunng its amount, it is commonly
called an electroscope; "but the name electrometer is properly
applied when greater or less degrees of difference are indicated
on any scale of reckoning, if approximately constant, even during
a single series of experiments. The first step towards accurate
electrometry in every case is to deduce from the scale-readings
numbers which shall be in simple proportion to the difference of
potentials to be determined. The next and last step is to assign
the corresponding values in absolute electrostatic measure. Thus,
when for any electrometer the first step has been taken, it remains
only to determine the single constant coefficient by which the
numbers deduced firom its indications as simply proportional to
differences of potential must be multiplied to give differences of
potential in absolute electrostatic measure. This coefficient will
be called, for brevity, the absolute coefficient of the instrument in
question.

§ 2. Thus, for example, the gold-leaf electrometer indicates
differences of potential between the gold leaves and the solid walls
enclosing the air-space in which they move. If this solid be of
other than sufficiently perfect conducting material, of wood and
glass, or of metal and glass, for instance, as in the instrument
ordinarily made, it is quite imperfect and indefinite in its indica-
tions, and is not worthy of being even called an electroscope, as
it may exhibit a divergence when the difference of potentials
which the operator desires to discover is absolutely zero. It is
interesting to remark that Faraday first remedied this defect by
coating the interior of the glass case with tinfoil cut away to
leave apertures proper and sufficient to allow indications to be
seen, but not enough to cause these indications to differ sensibly
from what they would be if the conducting envelope were com-
pletely closed around it ; and that not till a long time after did
any other naturalist, mathematician, or instrument-maker seem to
have noticed the defect, or even to have unconsciously remedied it.

§ 3. Electrometers may be classified in genera and species
according to the shape and kinematic relations of their, parts ;
but as in plants and animab a perfect continuity of intermediate
species has been imagined between the rudimentary plant and the
most perfect animal, so in electrometers we may actually construct
species having intermediate qualities continuous between the
^most widely different genera. But, notwithstanding, some such

FOR ELECTRICAL JffEASUREMENTS 221

classification as the following is convenient with reference to the
several instruments commonly in use and now to be described : —

I. Repulsion electrometers.

Pair of diverging straws as used by Beocaria, Volta,
and others, last century.

Pair of diverging gold leaves (Bennet).

Peltier's electrometer.

Delmann's electrometer.

Old-station electrometer, described in lecture to the
Royal Institution, May 1860 ; also in NichoFs Cyclo-
pcedia, article "Electricity, Atmospheric" (edition
1860), and in Dr Everett's paper of 1867, "On
Atmospheric Electricity" {Philosophical Transdc-
tions).

n. Symmetrical electrometers.

Bohnenberger^s electrometer.
Divided-ring electrometers.

in.* Attracted disk electrometers.

Absolute electrometer.
Long-range electrometer.
Portable electrometer.
Spring-standard electrometer.

§ 4. Class I. is sufficiently illustrated by the examples referred
to; and it is not necessary to explain any of these instruments
minutely at present, as they are, for the present at all events,
superseded by the divided-ring electrometer and electrometers of
the third class.

There are at present only two known species of the second
class; but it is intended to include all electrometers in which a
symmetrical field of electric force is constituted by two sym-
metrical fixed conductors at different electric potentials, and in
which the indication of the force is produced by means of an,
electrified body movable symmetrically in either direction firom
a middle position in this field. This definition is obviously ful-
filled by Bohnenberger's well-known instrument*.

* A single gold leaf hanging between the upper ends of two eqnal and similar
dry piles standing vertioally on a horizontal plate of metali one with its positive and
the other with its negative pole np.

222 PRACTICAL STANDARDS

§ 5. My first published description of a divided-ring electro-
meter is to be found in the Memoirs of the Roman Academy
of Sciences* about 1856; but since that time I have made great
improvements in the instrument — first, by applying a light mirror
to indicate deflections of the moving body ; next, by substituting
for two half rings four quadrants, and consequently for an electrified
body projecting on one side only of the axis, an electrified body
projecting symmetrically on the two sides and movable round an
axis ; and, lastly, by various mechanical improvements and by the
addition of a simple gauge to test the electrification of the movable
body, and a replenisher to raise this electrification to any desired
degree.

§ 6. In the accompanying drawings, Plate 5, fig. 1 represents
the fi-ont elevation of the instrument, of which the chief bulk
consists of a jar of white glass (flint) supported on three legs by
a brass mounting cemented round the outside of its mouth, which
is closed by a flat cover of stout sheet brass and a lantem>shaped
cover standing over a wide aperture in its centre. For brevity, in
what follows, these three parts will be called the jar, the main
cover, and the lantern.

Fig. 5 represents the quadrants as seen from above ; they are
seen in elevation at a and b, fig. 1, and in section at c and d, fig. 2.
They consist of four quarters of a fiat circular box of brass, with
circular apertures in the centres of its top and bottom. Their
position in the instrument is shown in figs. 1, 2, and 6. Each of
the four quadrants is supported on a glass stem passing down-
wards through a slot in the main cover of the jar, from a brass
mounting on the outside of it, and admits of being drawn out-
wards for. a space of about f of an inch (1 centim.) firom the
positions they occupy when the instrument is in use, which are
approximately those shown in the drawings. Three of them are
secured in their proper positions by nuts (e, e, e) on the outside of
the chief flat lid of the jar shown in fig. 4. The upper end of the
stem, carrying the fourth, is attached to a brass piece (/) resting
on three short legs on the upperside of the main cover, two of
these legs being guided by a straight V-groove at g to give them
fi~eedom to move in a straight line inwards or outwards, and to
prevent any other motion. This brass piece is pressed outwards
and downwards by a properly arranged spring (A), and is kept

* Accademia Poutificia dei Nuovi Linoei.

FOR ELECTRICAL MEASUREMENTS 223

from slidiDg out by a micrometer-screw (t) turning in a fixed nut.
This simple kinematic arrangement gives great steadiness to the
fourth quadrant when the screw is turned inwards or outwards
and then left in any position ; and at the same time produces but
little firiction against the sliding in either direction. The opposite
quadrants are connected in two pairs by wires, as shown in fig. 5 ;
and two stout vertical wires (/, m), called the chief electrodes,
passing through holes in the roof of the lantern, are firmly sup.
ported by long perforated vulcanite columns passing through these
holes, which serve to connect the pairs of quadrants with the
external conductors whose difference of potentials is to be tested.
Springs (n, o) at the lower ends of these columns, shown in figs. 1
and 2, maintain metallic contact between the chief electrodes and
the uppersides of two contiguous quadrants (a and b) when the
lantern is set down in its proper position, but allow the lantern
to be removed, carrying the chief electrodes with it, and to be
replaced at pleasure without disturbing the quadrants. The
lantern also carries an insulated charging-rod (j>), or temporary
electrode, for charging the inner coating of the jar (§11) to a
small degree, to be increased by the replenisher (§ 12), or, it may
be, for making special experiments in which the potential of the
interior coating of the jar is to be measured by a separate electro-
meter, or kept at any stated amount from that of the outer
coating. When not in use this temporary electrode is secured in
a position in which it is disconnected firom the inner coating.

§ 7. The main cover supports a glass column (q, fig. 2) pro-
jecting vertically upwards through its central aperture, to the
upper end of which is attached a brass piece (r), which bears
above it a fixed attracting disk («), to be described later (§ 13) ;
and projecting down fi-om it a fixed plate bearing the silk-fibre
suspension of the mirror (t), needle (u\ etc., seen in figs. 1 and 2,
and fixed guard tubes (v, w), to be described presently.

§ 8. The movable conductor of the instrument consists of a
stiff platinum wire (x), about 8 centimetres (3| inches) long, with
the needle rigidly attached in a perpendicular plane to it, and
connected with sulphuric acid in the bottom of the jar by a fine
platinum wire hung down from its lower end and kept stretched
by a platinum weight under the level of the liquid. The upper
end of the stiff platinum wire is supported by a single silk-fibre so
that it hangs down vertically. The mirror is attached to it just

S24: PRACTICAL 8TANDABr:DS^ '^

below its upper end. Thus the mirror, the needle, and the stiflT
platinum stem constitute a rigid body having very perfect freedom
to move round a vertical axis (the line of the bearing fibre), and
yet practically prevented from any other motion in the regular j

use of the instrument by the weight of its own mass and that of
the. loose piece of platinum hanging from it below the surface of
the liquid in the jar. A very small magnet is attached to the
needle, which, by strong magnets fixed outside the jar, is directed
to one position, about which it oscillates after it is turned through
any angle round the vertical axis and then left to itself. The
external magnets are so placed that when there is magnetic
equilibrium i)he needle is in the symmetrical position shown in
figs. 5 and 6 with reference to the quadrants*.

§ 9. The needle (u) is of very thin sheet aluminium cut to
the shape seen in figs. 5 and 6, the very thinnest sheet aluminium
that gives the iiequisite stiffiiess being chosen. If the four
quadrants are in a perfectly symmetrical position round it, and
if they are kept at one electric potential by a metalliop^arc con-
necting the chief electrodes outside, the needle may be istrongly
electrified without being disturbed from its position of magnetic
equilibrium ; but if it is electrified, and if the external electrodes
be disconnected and any difference of potentials established be-
tween them, the needle will clearly experience a couple turning
it round its vertical axis, its two ends being driven from the
positive quadrants towards the negative if it is itself positively
electrified. It is kept positive rather than negative in the
ordinary use of the instrument, because I find that when a
conductor with sharp edges or points is surrounded by another
presenting everywhere a smooth sur&ce, a much greater difference
of potentials can be established between them, without producing
disruptive discharge, if the points <and edges are positive than if
they are negative.

§ 10. The mirror (t) serves to indicate, by reflecting a ray of
light fix>m a lamp, small angular motions of the needle round the
vertical axis. It is a very light, concave, silvered glass mirror,
being only 8 millimetres (^ of an inch) in diameter, and 22 milli-
grammes (J grain) weight. I had for many years experienced
great difficulty in getting suitable mirrors for my form of mirror

* Recently I have made experiments on a bi filar suspension with a view to
superseding the magnetic adjustment, which promise weU.

FOB ELECTRICAL MEASUREMENTS

225

. I

I >

I 1

\.V*v

S24: iPRAOTlCAi; 8TANDABr:DS^

FOR ELECTRICAL MEASUREMENTS 225

galvanometer ; but they are now supplied in very great perfection
by Mr Becker, of Messrs Elliott Brothers, London. The focus for
parallel rays is about 50 centimetres (20 inches) from the mirror
and thus the rays of the lamp placed at a distance of 1 metre
(or 40 inches) are brought to a focus at the same distance. The
lamp is usually placed close behind the vertical screen a little
below or above the normal line of the mirror, and the image is
thrown on a graduated scale extending horizontally above or below
the aperture in the screen through which the lamp sends its light.
When the mirror is at its magnetic zero position the lamp is so
placed that its image is, as nearly as may be, in a vertical plane
with itself, and not more than an inch above or below its level;
so that there is as little obliquity as possible in the reflection,
and the line traversed by the image on the screen during the
deflection is, as nearly as may be, straight. The distance of the
lamp and screen from the mirror is adjusted so as to give as
perfect an image as possible of a fine wire which is stretched
vertically in the plane of the screen across the aperture through
which the lamp shines on the mirror; and with Mr Becker's
mirrors I find it easy to read the horizontal motions of the dark
image to an accuracy of the tenth of a millimetre. In the ordinary
use of the instrument a white paper screen, printed from a copper
plate, is employed, and the readings are commonly taken to about
a quarter of a scale-division ; but with a little practice they may,
when so much accuracy ia desired, be read with considerable
accuracy to the tenth of a scale-division. Formerly a slit in front
of the lamp was used; but the wire giving a dark line in the
middle of the image of the flame is a very great improvement,
first introduced by Dr Everett, in consequence of a suggestion
made by Professor P. G. Tait, in his experiments on the elasticity
of solids made in the Natural-Philosophy Laboratory of Glasgow
University *.

§ 11. The charge of the needle remains sensibly constant from
hour to hour, and even from day to day, in virtue of the arrangement
according to which it is kept in communication with sulphuric acid
in the bottom of the jar, the outside of the jar being coated with

* A Dnimmond light placed about 70 centimetres from the mirror gives tax
image, on a screen about 3 metres distance, brilliant enough for lecture-illustrations,
and with sufficient definition to allow accurate readings of the positions on a scale
marked by the image of a fine vertical wire in front of the light.

B. A. 15

226 PRACTICAL STANDARDS

tinfoil and connected with the earth, so that it is in reality a Leyden
jar. The whole outside of the jar, even where not coated with
tinfoil, is in the ordinary use of the instniment, especially in our
moist climate, kept virtually at one potential through conduction
along its surface. This potential is generally, by connecting wires
or metal pieces, kept the same as that of the brass legs and frame-
work of the instrument.. To prevent disturbance in case of strongly
electrified bodies being brought into the neighbourhood of the
instrument, a wire is either wrapped round the jar from top to
bottom, or a cage or network of wire, or any convenient metal
case, is placed round it; but this ought to be easily removed or
opened at any time to admit of the interior being seen. When
the instrument is left to itself from day to day in ordinary use,
the needle, connected with the inner coating of the jar as just
described, loses, of course, unless replenished, something of its
charge; but not in general more than ^ per cent, per day when
the jar is of flint glass made in Glasgow. On trying similar jars
of green glass I found that they lost their charge more rapidly
per hour than the white glass jars per month. I have occasionally,
but very rarely, found white glass jars to be as defective as those
green ones ; and it is possible that the defect I found in the green
jars was an accident to the jars tested, and not an essential property
of that kind of glass.

§ 12. I have recently made the very useful addition of a
replenisher to restore electricity to the jar from time to time when
required. It consists of (1) a turning vertical shaft of vulcanite
bearing two metal pieces called carriers (6,6, figs. 17 and 18, Plate 5);
(2) two springs (d, d, figs. 16 and 18), connected by a metallic arc,
making contact on the carriers once every half turn of the shaft,
and therefore called connectors ; and (3) two inductors (a, a) with
receiving springs (c, c) attached to them, which make contact
on the carriers once every half turn, shortly before the connecting
contacts are made. The inductors (a, a, figs. 16 and 18) are pieces
of sheet metal bent into circular cylindrical shapes of about 120^
each ; they are placed so as to deviate in the manner shown in
the drawing from parts of a cylindrical surface coaxal with the
turning-shaft, leaving gaps of about 60'' on each side. The diameter
of this cylindrical surface is about 15 millimetres (about ^ an inch).
The carriers (6, b, figs. 17 and 18) are also of sheet metal bent to
cylindrical surfaces, but not exactly circular cylinders, and are so

FOB ELECTRICAL MEASUREMENTS 227

placed on the bearing vulcanite shaft that each is nibbed by
the contact springs over a very short space, about 1 millimetre
beyond its foremost edge, when turned in the proper direction for
replenishing. The receiving springs (c, c, figs. 17 and 18) make
their contacts with each carrier immediately after it has got fairly
under cover, as it were, of the inductor. Each carrier subtends
an angle of about 60° at the axis of the turning-shaft. The
connecting contacts are completed just before the carriers commence
emerging from being under cover of the inductors. The carriers
may be said to be under cover of the inductors when they are
within an angle of 120'' on each side of the axis subtended by the
inductors. One of the inductors is in metallic communication
with the outside coating of the jar, the other with the inside.
Figs. 16, 17, and 18 illustrate sufficiently the shape of carriers and
the succession of the contacts. The arrow-head indicates the
direction to turn for replenishing. When it is desired to diminish
the charge, the replenisher is turned backwards. A small charge
having been given to the jar from an independent source, the
replenisher when turned forwards increases the difference of
potentials between the two inductors and the two coatings of the
jar connected with them by a constant percentage per half turn,
unless it is raised to so high a degree as to break down the air-
insulation by disruptive discharge. The electric action is explained
simply thus: — ^The carriers, when connected by the connecting
springs, receive opposite charges of induction, of which they deposit
large proportions the next time they touch the receiving springs.
Thus, for example, if the jar be charged positively, the carrier
emerging from the inductor connected with the inner coating
carries a negative charge round to the receiving spring connected
with the outside coating, while the other ciirrier, emerging from
the inductor connected with the outside coating, carries a positive
charge round to the receiving spring connected with the inside
coating. If the carriers are not sufficiently well under cover of the
inductors during both the receiving contacts and the connecting
contacts to render the charges which they acquire by induction
during the connecting contacts greater than that which they
carry away with them from the receiving contacts, the rotation,
even in the proper direction for replenishing, does not increase,
but, on the contrary, diminishes the charge of the jar. The
deviations of the inductors from the circular cylinder referred to

15—2

228 PRACTICAL STANDARDS

above have been adopted to give greater security against this
failure. A steel pivot fixed to the top of the vulcanite shaft, and
passing through the main cover, carries a small milled head
(y, fig. 1) above, on the outside, which is spun rapidly round in
either direction by pressing the finger on it; and thus in less
than a minute a small charge in the jar may be doubled. The
diminution of the charge, when the instrument is left to itself
for twenty-four hours, is sometimes imperceptible ; but when any
loss is discovered to have taken place, even if to the extent of
10 per cent., a few moments' use of the replenisher suffices to
restore it, and to adjust it with minute accuracy to the required
degree by aid of the gauge to be described presently. The principle
of the "replenisher" is identical with that of the "doubler" of
Bennet. In the essentials of its construction it is the same as
Varley*s improved form of Nicholson's "revolving doubler."

§ 13. The gauge consists of an electrometer of Class III. The
movable attracted disk is a square portion of a piece of very thin
sheet aluminium of the shape shown at a in fig. 4. It is supported
on a stretched platinum wire passing through two holes in the
sheet and over a very small projecting ridge of bent sheet aluminium
placed in the manner shown in the magnified drawing, fig. 3. The
ends of this wire are passed through holes in curved springs, shown
in fig. 4, and are bent round them so as to give a secure attachment
without solder and without touching the straight stretched part
of the wire. The ends of the platinum wire (/8, /8) are attached
by cement to the springs, merely to prevent them fi-om becoming
loose, care being taken that the cement does not prevent metallic
contact between some part of the aluminium wire and one or both
of the brass springs. I have constantly found fine platinum wire
rendered brittle by ordinary solder applied to it. The use of these
springs is to keep the platinum wire stretched, with an approximately
constant tension, from year to year and at various temperatures^
Their fixed ends are attached to round pins, which are held with
their axes in a line with the fibre by friction, in bearings forming
parts of two adjustable brass pieces (7, 7) indicated in fig. 4; these
pieces are adjusted once for all to stretch the wire with sufficient
force, and to keep the square attracted disk in its proper position.
The round pins bearing the stretching-springs are turned through
very small angles by pressing on the projecting springs with the
finger. They are set so as to give a proper amount of torsion

FOR ELECTRICAL MEASUREMENTS 229

tending to tilt the attracted disk (a) upwards, and the long end.
of the aluminium lever (S), of which it forms a part, downwards.
The downward motion of the long end is limited by a properly
placed stop. Another stop (c) above limits the upward motion,
which takes place under the influence of electrification in the use
of the instrument. A very fine opaque black hair (that of a small
black-and-tan terrier I have found much superior to any hitherto
tried) is stretched across the forked portion of the sheet aluminium
in which the long arm of the lever terminates. Looked at hori-
zontally from the outside of the instrument it is seen, as shown in
fig. 7, Plate 5, against a white background, marked with two very
fine black circles. These sight-plates in the instruments, as now
made by Mr White, are of the same material as the ordinary enamel
watch-dials with black figures on a white ground. The white space
between the two circles should be a very little less than the breadth
of the hair. The sight-plate is set to be as near the hair as it can
be without impeding its motion in any part of its range ; and it is
slightly convex forwards, and is so placed that the hair is nearer
to it when in the middle between the black circles than when in
any other part of its range. It is thus made very easy, even without
optical aid, to avoid any considerable error of parallax in estimating
the position of the hair relatively to the two black circles. By a
simple plano-convex lens (^, fig. 2), with the convex side turned
inwards, it is easy, in the ordinary use of the instrument, to dis-
tinguish a motion up or down of the hair amounting to j^ of an
inch. With a little care I have ascertained, Dr Joule assisting,
that a motion of no more than ^^ of an inch from one definite
central position can be securely tested without the aid of other
magnifying-power than that given by the simple lens. The lens
daring use is in a fixed position relatively to the framework bearing
the needle, but it may be drawn out or pushed in to suit the focus
of each observer. To give great magnification, it ought to be
drawn out so far that the hair and sight-plate behind may be but
little nearer to the lens than its principal focus, and the observer's
eye ought to be at a very considerable distance fi'om the instrument,
no less than 20 centimetres (8 inches), to get a good magnification;
and a short-sighted person should use his ordinary concave eye-lens
close to his eye. The reason for turning the convexity of the small
plano-convex lens inwards is, that if the eye of the observer is too
high or too low, the hair seems to him curved upwards or downwards.

230 PRACTICAL STANDARDS

and he is fchus guided to keep his eye on a level sufficiently constant
to do away with all sensible effects of parallax on the position of
the hair relatively to the black circles. The framework carrying
the stretched platinum wire and movable attracted disk is above
the brass roof of the lantern, in which a square aperture is cut to
allow the square portion constituting the short arm of the aluminium
balance to be attracted downwards by the fixed attracting disk
(§ *^)y to l>e presently described. A side view of the attracting
plate, the brass roof of the lantern, the aluminium balance, the
sight-plate, the hair, and the plano-convex lens is shown in section
(fig. 2), also a glass upper roof to protect the gauge and the interior
of the instrument below fi-om dust and disturbance by currents of
air, to which, without this upper roof, it would be exposed, through
the small vacant space round the movable aluminium square. The
fixed attracting disk is borne by a vertical screw screwing into the
upper brass mounting (z, fig. 2) (§ 7), connected with the inner
coating of the Leyden jar through the guard tubes, etc., and is
secured in* any position by the "jam nut," shown in the drawings
at z, fig. 2. This disk (s) is circular, and about 38 millimetres
(1^ inch) diameter, and it is placed horizontally with its centre
under the centre of the square aperture in the roof of the lantern.
Its distance from the lower surface of the roof and of the movable
attracted disk may be firom 2^ to 5 millimetres (from -j^ to ^ of an
inch), and is to be adjusted, along with the amount of torsion
in the platinum wire bearing the aluminium balance-arm, so as
to give the proper sensibility to the gauge. The sensibility is
increased by diminishing the distance fi*om the attracting to the
attracted plate and increasing the amount of torsion. Or, again,
the degree of the potential indicated by it when the hair is in the
sighted position is increased by increasing the distance between
the plates, or by diminishing the amount of torsion. If the
electrification of the needle is too great, its proper position of
equilibrium becomes unstable; or before this there is sometimes
a liability to discharge by a spark across some of the air-spaces.
The instrument works extremely well with the needle charged
but little less than to give rise to one or both of these faults,
and I adjust the gauge accordingly.

§ 14. The strength of the fixed steel-directing magnets is
to be adjusted to give the desired amount of deflection with any
stated difference of potentials maintained between the two chief

FOR ELECTRICAL MEASUREMENTS 231

electrodes, when the jar is charged to the degree which brings
the hair of the gauge to its sighted position. In the instraments
already made, the deflection* by a single cell of Daniell's amounts
to about 100 scale-divisions (of ^ of an inch each at a distance
of 40 inches) when the magnetic directive force is such as to give
a period of vibration equal to about 1*5 second. When the jar is
discharged and the four quadrants are connected with one another
and with the inner coating of the jar, lower degrees of sensibility
may be attained better by increasing the magnetic directing-force
than by diminishing the charge of the jar. Thus, for instance,
when it is to be used for measuring and photographically recording
the potential of atmospheric electricity at the point where the
stream of the water-dropping collector^ breaks into drops, the
magnetic directing-force may be made from 10 to 100 times more
than that just described. When this is lo be done it may be
convenient to attach a somewhat more powerful magnetic needle
than that which has been made in the most recent instruments
where a high degree of sensibility is desired. But it is to be
remarked ^that in general the directing-force of the external steel
magnets cannot be too strong, as the stronger it is the less is
the disturbance produced by changing magnetic bodies in the
neighbourhood of the instrument. In laboratory work, where
numerous magnetic experiments are being performed in the
immediate neighbourhood, and in telegraph factories, where there
is constant disturbance by large moving masses of iron, the
artificial magnetic field of the electrometer ought to be made very
strong. To allow this, and yet leave, sufficient sensibility to the
instrument, the suspended magnetic needle has been made smaller
and smaller, until it is now reduced to two small pieces of steel
side by side, 6 millimetres (^ of an inch) long. For a meteorological
observatory all that is necessary is, that the directing magnetic
force should be so great that the greatest disturbance experienced
in magnetic storms shall not sensibly deflect the luminous image J.
§ 15. The sensibility of the gauge should be so adjusted that

* That is to say, the namber of soale-divisionB over which the laminoas image
moTes when the chief electrodes are disconnected fh>m one aoother and pat in
metallic connexion with the two plates of a Daniell's battery.

t Bee Boyal Institution Lecture, May 18, 1860 {Proceedings of the R. I.), or
Nichol's Cyclopadia, article " Electricity, Atmospheric '* (edition 1860).

X All embarrassment from this source wiU be done away with if the bifilar plan
be adopted {vide footnote to 8 8).

2^2 PRACTICAL STANDARDS

a variation in the charge of the jar, producing an easily perceived
change in the position of the hair, shall produce no sensible change
in the deflection of the luminous image produced by the greatest
difference of potentials between the quadrants, which is to be
measured in the use of the instrument. I believe the instruments
already made, when adjusted to fulfil these conditions, may be
tested to measure the difference of potentials produced by a
single cell of Daniell's to an accuracy of a quarter per cent. It
must be remembered that the constancy of value of the unit of
each instrument depends not only on the constancy of the potential
indicated by the gauge, but also on the constancy of the force in
the field traversed by the suspended needle. As both these may
be expected to decrease gradually from year to year (although
very slowly after the first few hours or weeks), rigorous methods
must be adopted to take such variations into account, if the
instrument is to be trusted to as giving accurately comparable
indications at all times. The only method hitherto provided for
this most important object consists in the observation of the
deflection produced by a measured motion of one of the quadrants
by the micrometer-screw (i) when the four quadrants are put in
metallic communication with one another through the principal
electrodes — the force producing this deflection when the potential
of the jar is constant ; and therefore, the jar being brought to one
constant potential by aid of the gauge, the amount of the deflection
will show whether or not the force of the magnetic field has
changed, and will render it easy at any time to adjust the strength
of the magnets, if necessary, to secure this constancy. But to
attain this object by these means, the three quadrants not moved
by the micrometer-screw must be clamped by their fixing-screws
so that they may be always in the same position.

§ 16. The absolute constancy of the gauge cannot be altogether
relied upon. It certainly changes to a sensible degree with
temperature, and to very different degrees, and even in different
directions, as will be seen (§ 32) in connexion with the. description
of the portable electrometer to be given later. But this temperature
variation does not amount in ordinary cases probably to as much
as one per cent.; and it is probable that after a year or two any
further secular variation of the platinum torsion spring will be
quite insensible. It is to be remarked, however, that secular
experiments on the elasticity of metab are wanting, and ought

FOR EXECTBICAL l^ASUREMENTS 233

at least to be commenced in our generation. In the meantime
it will be desirable, both on account of the temperature variation
adid of the possible secular variation in the couple of torsion^ to
check the gauge by accurate measurements of. the time of oscillation
<^ the needle with its appurtenances. The moment of inertia of
this rigid body, except in so far as it may be influenced by oxida-
tion of the metal, of which I have as yet discovered no signs, may
be regarded as constant; and therefore the amount of the directing
couple due to the magnets may be determined with great accuracy
by finding the period of an oscillation when the four quadrants
are put in connexion through the charging rod with the metal
mounting bearing the guard-plates, etc. I have not as yet put
into practice anj*" of the obvious methods, founded on the general
principle of coincidences used in pendulum observations, for
determining the period of the oscillation ; but although not more
than twenty or thirty oscillations can be counted, it seems certain
that with a little trouble the period of one of them may be de-
termined without much trouble to an accuracy of about -j^ per cent.

Absolute Electrometer.

§ 17. The absolute electrometer (fig. 11» Plate 6) and the
other instruments of Class III. are founded on a method of
experimenting introduced by Sir Wm. Snow Harris, and described
in his first paper "On the Elementary Laws of Electricity*"
thirty-four years ago. In these experiments a conductor, hung
from one arm of a balance and kept in metallic communication
with the earth, is attracted by a fixed insulated conductor, which
is electrified, and, for the sake of keeping its electric potential
constant, is connected with the inner coating of a Leyden battery.
The first result which he announced is, that, when other circum-
stances remain the same, the attraction varies with the square
of the quantity of electricity with which the insulated body is
charged; but "it is readily seen that, in the case of Mr Harris's
experiments, it will be so slight on the unopposed portions that
it could not be perceived without experiments of a very refined
nature, such as might be made by the proof plane of Coulomb,
''which is, in fact, with a slight modification, the instrument
** employed by Mr Faraday in the investigation. Now to the

* Philoiophieal Tramactiont, 1884.

4(
4t

234 PRACTICAL STANDARDS

" degree of approximation to which the intensity on the unopposed
"parts may be neglected, the laws observed by Mr Harris when
" the opposed surfaces are plane may be readily deduced from the
" mathematical theory. Thus let v be the potential in the interior
'* of J., the charged body, a quantity which will depend solely on
" the state of the interior coating of the battery with which, in
"Mr Harris's experiments, A is connected, and will therefore be
"sensibly constant for difiTerent positions of A relative to the
"uninsulated opposed body B. Let a be the distance between
" the pl€ine opposed faces of A and J3, and let S be the area of the
"opposed parts of these faces, which will in general be the area
" of the smaller, if they be unequal. When the distance a is so
"small that we may entirely neglect the intensity on all the
"unopposed parts of the bodies, it is readily shown, from the
" mathematical theoiy, that (since the difference of the potentials
" at the surfaces of A and B is v) the intensity of the electricity
" produced by induction at any point of the portion of the surface

" of B which is opposed to A is 7— , the intensity at any point

" which is not so situated being insensible. Hence the attraction
" on any small element co, of the portion S of the surface of B,
"will be in a direction perpendicular to the plane and equal to

" 2'jr i 1 — I *. Hence the whole attraction on B is ^ — r .
V47ra/ oTra"

"This formula expresses all the laws stated by Mr Harris as
" results of his experiments in the case when the opposed surfaces
"are planet."

§ 18. After many trials to make an absolute electrometer
founded on the repulsion between two electrified spherical con-
ductors for which I had given a convenient mathematical formula
in § 4 of the paper just quoted, it occurred to me to take advantage
of the &ct noticed by Harris, but easily seen as an immediate
consequence of Green's mathematical theory, that the mutual
attraction between two conductors used as in his experiments
is but little influenced by the form of the unopposed parts ; and
in 1853, in a paper " On transient Electric Currents J," I described

* See Mathematical Journal^ vol. iii. p. 275.

t *' On the Elementary Laws of Statical Electricity," Cambridge and Dublin
Mathematical Journal^ 1846; and Phil, Mag, Jaly 1854.
t Phil, Mag. June 1858.

FOR ELECTRICAL MEASUREMENTS 235

a method for measuring differences of electric potential in absolute
electrostatic measure founded on that idea. The "absolute
electrometer" which I exhibited to the British Association at
its Olasgow Meeting in 1855 was constructed for the purpose
of putting these methods in practice. This instrument consists
of a plane metal disk insulated in a fixed horizontal position, with
a somewhat smaller fixed metal disk hung centrally over it
from one end of the beam of a balance. In two papers entitled
"Measurement of Electrostatic Force produced by a Battery"
and " Measurement of the Electromotive Force required to produce
a spark in Air between parallel metal plates at different distances/'
published in the Proceedings of the Royal Society* for February
1860, I described applications of this electrometer, in which, for
the first time, I believe, absolute electrostatic measurements were
made. The calculations of differences of potentials in absolute
measure were made according to the formula quoted above (§ 17)
from my old paper on "The Elementary Laws of Statical
Electricity."

§ 19. This formula is rigorous only if the distance between
the disks is infinitely small in comparison with their diameters;
and therefore, in my earliest attempt to make absolute electrostatic
measurements, I used very small distances. I found great difficulty
in securing that the distance should be nearly enough equal
between different parts of the plates, and in measuring its absolute
amount with sufficient accuracy; and found besides serious in-
conveniences in respect of sensibility and electric range: later
I made a great improvement in the instrument by making only
a small central area of one of the disks movable. Thus the electric
part of the instrument becomes two large parallel plates with
a circular aperture in one of them, nearly filled up by a light
circular disk supported properly to admit of its electrical attraction
towards the other being accurately measured in absolute units
of force. The disk and the perforated plate surrounding it will
be called, for brevity, the disk and the guard-plate. The faces
of these two next the other plate must be as nearly as possible in
one plane when the disk is precisely in the position for measuring
its electric force, which, for brevity, will be called its sighted
position. The space between the disk and the inner edge of its

* Phil. Mag, September and October 1860.

236 PRACTICAL STANDARDS

guard-ring must be a very small part of the diameter of the.
aperture, and must be very small in comparison with the distance
between the plates; but the diameter of the disk may be greater
than, equal to, or less than the distance between the platea

§ 20. Mathematical theory shows that the electric attraction
experienced by the disk is the same as that experienced by a
certain part of one of two infinite planes at the same distance,
with the same difference of electric potentials, this area being
very approximately the mean between the area of the aperture
and the area of the disk, and that the approximation is very good,
even although the distance between the plates be as much as
a fourth or fifth, and the diameter of the disk as much as three-
fourths of the diameter of the smaller of the two plates. This
conclusion will be readily assented to when we consider that* the
resultant electric force at any point in the air between the two
plates is equal numerically to the rate of conduction of heat per
unit area across the corresponding space in the following thermal
analogue. Let a solid of uniform thermal conductivity replace all
the air between and round the plates; and in place of the plates
let there be hollow spaces in this solid. Let these hollow spaces
be kept at two uniform temperatures, differing by a number of
degrees equal numerically to the difference of potentials in the
electric system, the space corresponding to the disk and guard-
ring being at one temperature, and that corresponding to the
opposite plate at the other temperature; and let the thermal
conductivity of the solid be unity. If we attempt to draw the
isothermal surfaces between the hollow corresponding to the
continuous plate on the one side, and that corresponding to the
disk and guard-ring on the other side, we see immediately that
they must be very nearly plane from very near the disk all the
way across to the corresponding central portion of the opposite
plate, but that there will be a convexity towards the annular
space between the disk and guard-ring.

§ 21. Thus we see that the resultant electric force will, to

V
a veiy close approximation, be equal to jz for all points of the air

between the plates at distances fix)m the outer bounding edges

* "On the Uniform Conduction of Heat through Solid Bodies, and its connexion
with the Mathematical Theory of Electricity,'* Cambridge Mathematical Journal^
Feb. 1S42, and Phil Mag. July 1654.

FOR ELECTRICAL MEASUREMENTS 237

exceeding two or three times the distance between the plates,
and at distances from the interstice between the guard-ring and
disk any greater than the breadth of this interstice. Hence if p
denote the electric density of any point of the plate or disk fer
enough from the edges, we have

P =

47ri)'

But the outward force experienced by the surfistce of the
electrified conductor per unit of area at any point is iirp^] and
therefore if F denote the force experienced by any area A of the
fixed plate, no part of which comes near its edge, we have

SwD^'

which will clearly be equal to the attraction experienced by the
movable disk, if il be the mean area defined above. This gives

F=2)^--j— , the formula by which difference of potentials in

absolute electrostatic measure is calculated fi'om the result of a
measurement of the force F, which, it must be remembered, is to
be expressed in kinetic units. Thus if W be the mass in grammes
to which the weight is equal, we have

where g is the force of gravity in centimetres per second.

The difficulty which, in first applying this method about twelve
years ago, I found in measuring accurately the distance D between
the plates and in avoiding error from their not being rigorously
parallel, I now elude by measuring only differences of distance,
and deducing the desired results from the difference of the
corresponding differences of potentials. Thus let F' be the
difference of potentials between the plates required to give the
same force F; when the difference of potentials is V instead
of F, we have

§ 22. The plan of proceeding which I now use is as follows: —
Each plate (fig. 11, Plate 6) is insulated; one of them, the
continuous one, for instance, is kept at a potential differing from

238 PRACTICAL STANDARDS

the earth by a fixed amount tested by aid of a separate idiostatic*
electrometer; the other plate (the guard-ring and movable disk
in metallic communication with one another) is alternately
connected with the earth and with the body whose potential is
to be measured. The lower plate is moved up or down by a
micrometer-screw until the movable disk balances in a definite
position, indicated by the hair (with background of white with
black dots) seen through a lens, as shown in fig. 11. Before and
after commencing each series of electrical experiments, the amount
of weight to be placed on the upperside of the disk to bring the
hair to its sighted position when there is no electric force is deter-
mined. This last condition is secured by putting the two plates
in metallic communication with one another. For the electric
experiments the weight is removed, so that when the hair is in
the sighted position the electric attraction on the movable disk
is equal to the force of gravity on the weight. The electric
connexions suitable in using this instrument for determining in
absolute electrostatic measure the diiference of potentials main-
tained by a galvanic battery between its two electrodes are indicated
in fig. 11. No details as to the case for preventing disturbance
by currents of air, and for maintaining a diy atmosphere, by aid
of pumice impregnated with strong sulphuric acid, are shown,
because they are by no means convenient in the instrument at
present in use, which has undergone so many transformations that
scarcely any part of the original structure remains. I hope soon
to construct a compact instrument convenient for general use.
The amount of force which is constant in each series of experiments
may be varied firom one series to another by changing the position
of a small wire rider on the lever from which the movable disk
is hung.

The electric system here described is heterostatic (§ 40 below),
there being an independent electrification besides that whose
difference of potential is to be measured.

* See §40 below.

for electrical measurements 239

Portable Electrometer.

§ 23. In the ordinary use of the portable electrometer (figs. 8,
9, and 10, Plate 6) the electric system is heterostatic and quite
similar to that of the absolute electrometer, when used in the
manner described above in § 22. But the balance is not adapted
for absolute measure of the amount of force of attraction experienced
by the movable disk ; on the contrary, it is precisely the same as
that described for the gauge of the quadrant electrometer in § 13
above, only turned upside down. Thus, in the portable instrument,
the square disk (/) forming part of the lever of thin sheet aluminium
is attracted upwards by a solid circular disk of sheet brass (g),
thick enough for stifihess. Every part of the aluminium lever
except this square portion is protected from electric attraction by
a fixed brass plate (A, h) with a square hole in it, as nearly as may
be stopped by the square part of the sheet aluminium destined to
experience the electric attraction, all other parts of the aluminium
balance-lever being below this guard-plate. The aluminium lever
(t, k), as shown in figs. 8 and 10, is shaped so that when the hair
{I) at the long end of its lever is in its sighted position, the upper
surfaces of the fixed guard-plate (h) and movable aluminium square
(/) are as nearly as may be in one plane. The mode of suspension
is precisely the same as that described (§ 13) for the gauge of the
quadrant electrometer. In the portable instrument, careful atten-
tion is given by the maker to balance the aluminium lever by
adding to it small masses of shellac or other convenient substance,
so that its centre of gravity may be in the line of its platinum*
wire axis, or, more properly speaking, in such a position that the
instrument shall give, when electrified, the same "earth-readings"
when held in any positions, either upright, or inclined, or inverted
(§ 30 below). Thus the condition of equilibrium of the balance,
when the hair is in its sighted position, is that the moment of
electric attraction round the axis of suspension shall be equal
to the moment of the couple of torsion, the latter being as constant
as the properties of the matter concerned (platinum wire, brass
stretching-springs, etc.) will allow.

§ 24f. The guard-plate carrying, by the platinum-wire suspen-
sion, the aluminium balance, is attached to the bottom of a small
glass Leyden jar (m, m), and is in permanent metallic communica-
tion with its inside coating of tinfoil. The outside tinfoil coating

240 PRACTICAL STANDARDS

of this jar is in permanent metallic communication with the outside
brass-protecting case. The upper open mouth of this case is closed
by a lid or roof, which bears on its underside a firm frame projecting
downwards. This firame has two V notches, in which a stout brass
tube (o) slides, kept in the Vs by a properly placed spring (p)
giving it freedom to slide up and down in one definite line*.
Firmly fixed in the upper end of this tube is a nut (a, fig. 8),
which is caused to move up and down by a micrometer-screw.
The lower end of the shaft of this screw has attached to it a convex
piece of polished steel {b, fig. 8), which is pressed upon a horizontal
agate plate rigidly attached to the framework above mentioned by
a stiff brass piece projecting into the interior of the brass tube
through a slot long enough to allow the requisite range of motion.
This arrangement will be readily understood from the accompanying
drawings. It has been designed upon obvious geometrical prin-
ciples, which have been hitherto neglected, so far as I know, in all
micrometer-screw mechanisms, whether for astronomical instruments
or other purposes. The screw-shaft is turned by a milled head,
fixed to it at its top outside the roof of the instrument, and the
angles through which it is turned are read on a circle divided into
100 equal parts of the circumference (or 3°'6 each) from a fixed
mark on the roof of the instrument. The hole in the roof through
which the screw-shaft passes is wide enough to allow the shaft to
turn without touching it, and the lower edge of the graduated
circle turning with the screw is everywhere very near the upperside
of the roof, but must not touch it at any point. A second nut
(c, fig. 8) above the eSective nut fits easily, but somewhat accurately,
in the hollow brass tube, but is prevented from turning round in
the tube by a proper projection and slot. Thus the screw is
rendered sufficiently steady, with reference to the sliding-tube ;
that is to say, it is prevented from any but excessively small rota-
tions round an axis perpendicular to the length of the screw-shaft;
and when the nut is kept from being turned round its proper axis,
it forms along with the sliding-tube virtually a rigid body. A

* In conseqaence of baggestions by Mr Jenkln, it is probable that the spring
maj be done away with, and the Vb replaced by rings approximately fitting round
the tube, but leaving it quite free to fall down by its own weight. In consequenoe
of the symmetrical position of the convex end of the screw over the centre of the
attracted disk, slight lateral motions of the tube produce no sensible effect on the
electric attraction.

.> XiA'^V

^-^aftiM^

VtAJ^

•i

•'^ V— ^;>,:S

1.4

..V-*

'-^ ^.'■'

T - t---r:

O.'- .'.^ > C •

•,tr

)i!'

> ,1
3 -

- i

/. 'i. f'. x

' '^NVji ^\<i'^

sar^-^i^^A. e«t.

^-i-^

[•\.

C iii?

t
J

' -»]

J *

•• K :

:i.

♦. ,i-'

,■ — — ^

T— ■

.•\ I ,Un...'.\

r- .rr-^S-^^i-^.T^..?^

!!•

: i I

"?:i

i^r

"^'

FOR ELECTRICAL MEASUREMENTS 241

carefully arranged spiral spring presses the two nuts asunder, and
so causes the upperside of the thread of the screw-shaft always to
press against the underside of the thread of the effective nut, thus
doing away with what is technically called in mechanics 'Most
time." In turning the micrometer-screw, the operator presses its
head gently downwards with his finger, to secure that its lower
end bears firmly upon the agate plate. It would be the reverse
of an improvement to introduce a spring attached to the roof of
the instrument outside to press the screw-head downwards, inasmuch
as however smooth the top of the screw-shafb might be made, and
however smooth the spring pressing it down, there would still be
a very injurious friction impeding the proper settlement of the
sliding-tube into its Vs. A stiff fork (q) stretching over the
graduated circle is firmly attached to the roof outside, to prevent
the screw from being lifted up by more than a very small space,
perhaps not more than ^ of an inch at most. In using the
instrument, the observer should occasionally pull up the screw-
head and press it down again, and give it small horizontal motions,
to make sure that when he is using it it is pressed in properly to
its Vs and down upon the agate plate. A long arm (dy figs. 8
and 9) (or two arms one above the other), firmly attached to the
sliding-tube, carries a pointer which moves up and down with it.
Two fixed guiding-cheeks on each side of this pointer prevent the
tube from being carried round too far in either direction when the
screw is turned: one of these cheeks is graduated so that each
division is equal in length to the step of the micrometer-screw;
this enables the operator to ascertain the number of times he has
turned the screw. These two cheeks must never simultaneously
press upon the sliding-pointer; on the contrary, they must leave
it a slight amount of lateral fireedom to move. If this does not
amount to 0'36 of a degree, the amount of " lost time *' produced by
it will not exceed ^'(7 of a division of the micrometer-circle, and
will not produce any sensible error in the use of the instrument.
A glass rod cemented to the lower end of the tube prolongs its
axis downwards, and bears the continuous attracting-plate of the
electrometer at its lower end.

The object aimed at in the mechanism just described is to
prevent the nut and other parts rigidly connected with it from
any other motion than parallel to one definite line, and to leave
it freedom to move in this line, unimpeded by any other fi-iction

B. A. 16

242 PRACTICAL STANDARDS

than that which is indispensable in the arrangement for keeping
the sliding-tube in its Vs.

§ 25. If the inner tinfoil covering of the Leyden jar were
completed up to the guard-plate bearing the aluminium balance,
the long arm of this lever being in the interior of a hollow conductor
would experience no electric influence and no force from the
electrification of the Leyden jar, or from separate electrification
of the upper attracting-plate, or, more strictly speaking, the electric
density and consequent electric force on the long arm of the lever
would be absolutely insensible to the most refined test we could
apply, because of the smallness of the gap between the movable
aluminium square and the boundary of the square aperture in
the guard-plate. But to see the hair on the long end of the lever,
und the white background with black dots behind it, a good portion
of the glass under the guard-plate must be cleared of tinfoil outside
and inside. Thjis the electric potential of the inner coating of the
Leyden jar will not be continued quite uniformly over the inner
surface of the bared portion of the glass, and a disturbance affecting
chiefly the most sensitive part of the lever will be introduced.
To diminish this as much as possible without inconveniently
impeding vision, a double screen of thin wire fences, in metallic
communication with the inner tinfoil coating and the guard-plate,
is introduced between the end of the lever and the glass through
which it is observed.

§ 26. A very light spiral spring (r) connects the upper
attracting-plate with a brass piece supported upon a fixed vertical
glass column projecting downwards from the roof of the instrument
This brass piece bears a stout wire («), called the main electrode,
projecting vertically upwards along the axis of a brass tube open
at each end, fixed in an aperture in the roof so as to project
upwards and downwards, as shown in fig. 9.

§ 27. The top of the main electrode bears a brass sliding-pieoe
(t), which, when raised a little, serves for umbrella and wind^guard
without disturbing the insulation ; and when pressed down closes
the aperture and puts the electrode in metallic connexion with
the roof of the instrument. When the instrument is to be used
for atmospheric electricity (unless at a fixed station) a steel wire,
about 20 centimetres long, is placed in the hole on the top of the
sliding brass piece just mentioned, and is thus held in the vertical
position. A burning match is attached to its upper end, which

FOR ELECTRICAL MEASUREMENTS 243

bos the effect of bringing the potential of the chief electrode and
upper attracting-plate etc. all to the potential of the air at the
point where the match bums*. The instrument is either held
in the observer's hand, or it is placed upon a fixed support, and
care taken that its outer brass case is in connexion with the earth.
When the difference of potentials between two conductors is to
be tested, one of these is connected with the brass case of the
instrument, and the other with the chief electrode, the umbrella
being kept up. If both of these conductors must be kept insulated
from the earth, the brass case of the electrometer must be put
on an insulating stand, and the micrometer-screw turned by an
insulating handle.

§ 28. A lead cup (e, fig. 8), supported by metal pillars from
the roof and carrying pieces of pumice-stone, held in their place
by india-rubber bands, completes the iustrument The inner
surface of the glass must be clean, and particles of dust, minute
shreds or fibres, etc. removed as carefully as possible, especially
from the lower surface of the upper attracting-plate, and the upper
sur&ce of the guard-plate and aluminium square facing it from
below. The pumice is prepared by moistening it with a few drops
of strong pure sulphuric acid. Ordinary sulphuric acid of commerce
should be boiled with sulphate of ammonia to free it from volatile
acid vapours, and to strengthen it sufficiently by removing water
if the acid be not of the strongest. There should not be so much
acid applied to the pumice as to make it have the appearance of
being moist, but there must be enough to maintain a suflSciently
dry atmosphere within the instrument for very perfect insulation
of the Leyden jar, which I find does not in general lose more of
its charge than 5 per cent, per week when the pumice is properly
acidulated. Thus there is no tendency of the liquid to drop out
of the pumice; and the pumice being properly secured by the
india-rubber bands, the instrument may be thrown about with
any force, short of that which might break the glass jar or either
of the glass stems, without doing any damage ; but to ensure this
hardiness the sheet aluminium of which the balance is made must
be very thin. After several weeks' use the pumice may commence
to look moist, and even slight traces of moisture may be seen
on the outside- of the lead cup, in consequence of watery vapour

* See Kiehors Cyclopadia^ artiole "Eleetrioity, Atmospheric," 2nd edition, i860 ;
4ft Royal tfuHttUion Lecture on Aimotpheric Eleetrieity, May 1860.

16—2

244 PRACTICAL STANDARDS

attracted by the flulphuric acid from the atmosphere; but the
pumice should then be taken and dried. At all events this must
be done in good time, before enough of the liquid has collected
to give any tendency to drop. In all climates in which I have
hitherto tested the instrument, I have found the pumice effective
for insulation and safe in keeping all the liquid to itself for two
months. But it having been reported to me by Mr Becker that
many instruments have been returned to him in a ruinous condition
from drops of sulphuric acid having become scattered through their
metal work, I now cause to be engraved conspicuously on the outer
case of the instrument " pumice dangerous, if not dried once
A month"; also a frame carrying a card, on which the dates of
drying are inscribed, to be placed in a convenient position on the
roof of the instrument.

§ 29. To prepare the instrument for use, the inner coating
of the Leyden jar must be charged through a charging-rod,
insulated in a vulcanite or glass tube and let down for the occasion
through a hole in the roof of the instrument, by aid of a small
electrophorus, which generally accompanies the instrument, or by
an electrical machine. I generally prefer to give a negative charge
to the inner coating, as I have not found any physical reason, such
as that mentioned in § 9 above, to prefer a positive charge to a
negative charge ; and the negative charge gives increased readings
of thv micrometer, in the ordinary use of the instrument, to
correspond to positive charges of the principal electrode, as will
be presently explained. Before commencing to charge the jar,
the upper attracting-plate should be moved to nearly the highest
position of its range by the micrometer-screw, otherwise too strong
a force of electric attraction may be put upon the aluminium
square; and, besides, the jar will discharge itself between the
upper plate and the extreme edge of the aluminium square^
pulled as it is very much above the level of the guard-plate by
the electric attraction. I have not found any injury or change
of electric value of the scale-divisions to arise from any such rough
usage ; but still, to guard against such a possibility, I propose to
add to the guard-plate checks to prevent the corners of the
aluminium from rising much, if at all, above its level, and to
conduct the discharge and protect the aluminium and platinum
from the shock, in case of the upper plate being brought too near
the lower. When the instrument is being charged, or when it is

FOR ELECTRICAL MEASUREMENTS 245

out of use at any time, the umbrella should always be kept down ;
but it must be raised to insulate the principal electrode, of course,
before proceeding to apply this to a body whose difference of
potential from a body connected with the case of the instrument
is to be measured.

§ 30. In using the instrument the umbrella must very
frequently be lowered, or metallic communication established in
any other convenient way between the chief electrode and the
outer brass case, the micrometer-screw turned until the hair takes
its sighted position, and the reading taken, the hundreds being
read on the interior vertical scale, and the units (or single divisions
of the circle) on the graduated circle above. The number thus
found is called the earth-reading; it measures the distance from
an arbitrary zero position to the position in which the upper
attracting-plate must be placed to give the amount of electric
force on the aluminium square which balances the lever in its
sighted position. A constant added to the earth-reading, or
subtracted from it, gives (§ 1) a number simply proportional to
the difference of potentials between the upper and lower plate ;
that is to say, between the two coatings of the Leyden jar. The
vertical scale and micrometer-circle are numbered, so that increased
distances between the plates give increased readings; and the
zero reading should correspond as nearly as may be to zero distance
between them, although in the instruments hitherto made no pains
have been taken to secure this condition, even somewhat approxi-
mately. If it is desired to know the constant, an electrical
experiment must be made to determine it, which is done with
ease ; but this is not necessary for the ordinary use • of the
instrument, which is as follows.

§ 31. First an earth-reading is taken, then the upper electrode
is insulated by raising the umbrella, or otherwise breaking connexion
between the principal electrode and the outer metal case of the
instrument. The principal electrode and the outer case are then
connected with the two bodies whose difference of potential is to
be determined, and the micrometer-screw is turned until the hair
is brought to its sighted position. The reading of hundreds on
the vertical scale and units on the circle is then taken. Lastly,
the principal electrode is again connected with the case of the
instrument and another earth-reading is taken. If the second
earth-reading differs from the first, the observer must estimate

246 PRACTICAL STANDARDS

the most probable earth-reading for the moment when the hair
was in its sighted position, with the upper plate and the metal
case in connexion with the two bodies whose difference of potential
is to be measured. The estimated earth-reading is to be subtracted
&om the reading taken in connexion with the bodies to be tested.
This difference measures (§ 21) the required difference of potentials
between them in units of the instrument The value of the unit
of the instrument ought to be known in absolute electrostatic
measure ; and the difference of reading found in any experiment
is to be multiplied by this, which is called (§1) the absolute
coefficient of the instrument, to give the required difference of
potentials in absolute measure. It so happens that, in the
portable electrometers of the kind now described which have
been hitherto constructed, the absolute coefficient is somewhere
about 001, so that one turn of the screw, or 100 divisions of the
circle, corresponds to somewhere about one electrostatic unit, with
a gramme for the unit of mass, a centimetre for the unit of distance^
and a second for the unit of time ; but the different instruments
differ from one another by as much as ten or twenty per cent,
in their absolute coefficients. In all of these I have found between
three and four Daniell's cells to correspond to the unit division ;
that is to say, between three hundred and four hundred cells to
a full turn of the screw. With great care, the observer may
measure small differences of potentials by this instrument to the
tenth part of a division (or to about half a DanielFs cell). With
a very moderate amount of practice and care, an error of as much
as a half division may be avoided in each reading.

§ 32. But there are imperfections in the instrument itself
which make it difficult or impossible to secure very minute accuracy,
especially in measurements through wide ranges.

(1) In the first place, I am not sure that the end of the
needle carrying the hair is protected sufficiently by the wire
fences (§ 25) fix)m electric disturbance to provide against any
error from this source, which possibly introduces serious irregu-
larities.

(2) In the second place, the capacity of the jar in the small
portable instrument is not sufficient to secure that the potential
of its inner coating shall not differ sensibly with the different
distances to which the upper plate is brought to balance the
aluminium lever with the hair in its sighted position. But on

FOR ELECTRICAL MEASUREMENTS 247

this point it is to be remarked that the electric density on the
upper surfieu^ of the guard-plate is in its central parts always
the same when the hair is in its sighted position; and it is
therefore only the comparatively small difference of the quantity
of electricity on this sur&ce, towards the rim, corresponding to
different distances of the attracted plate, that causes difference
of potential in the inner coating of the jar. But if the upper
attracting-plate be kept for several minutes at any distance,
differing by a few turns of the screw, from that which brings the
hair to its sighted position, the electricity creeps along the inner
unconnected sur&ce of the glass, so as to increase the charge of
the inner metallic coating or diminish it, according as the distance
is too great or too small. If then quickly the screw be turned
and the earth-reading taken, it is found greater or smaller, as the
case may be, than previously; but after a few minutes more it
returns to its previous value very approximately. Elrror from this
source may be practically avoided by taking care never to allow
the hair to remain for more than a few minutes far from its sighted
position — never so far, for instance, as above the centre of the
upper, or below the centre of the lower dots.

(3) A third source of error arises from change of temperature
influencing the indications. In most of the instruments hitherto
made I have found that the warmth of the hand produces in
a few minutes a very notable augmentation of the earth-reading
(as it were an increased charge in the jar); but in the last
instrument which I have tested (White No. 18) I find the reverse
effect, the earth-reading becoming smaller as the instrument is
warmed, or larger when it is cooled. I have ascertained that these
changes are not due to changes in the electric capacities of the
Leyden jars ; and I have found that the change, if any, of specific
inductive capacity of glass by change of temperature is excessively
small, in comparison to what would be required to account for the
temperature errors of these instruments, which probably must be
due to thermo-elastic properties of the platinum wire, or of the
stretching-springs, or of the aluminium balance-lever, or to a
combination of the effects depending on such properties; but
I have endeavoured in vain, for several years, and made many
experiments, to discover the precise cause. It surely will be
found, and means invented for remedying the error, now that
I have an instrument in which the error is in the opposite

248 PRACTICAL STANDARDS

direction to that of most of the other instruments. It is of coarse
much greater in some instruments than in others: in some it is
so great that the earth-reading is varied by as much as twenty
divisions by the warmth of the hand in the course of five or ten
minutes after commencing to use the instniment, if it has been
previously for some time in a cold place. Its influence may be
eliminated, not quite rigorously, but nearly enough so for most
practical purposes, by frequently taking earth-readings (§ 30) and
proceeding according to the directions of § 31.

(4) A fourth fault in the portable electrometer is, that the
diameter of the guard-plate and upper attracting-disk, which
ought to be infinite, are not sufficiently great, in proportion to
the greatest distance between them, to render the scale quite
uniform in its electric value throughout. A careful observer,
however, will remedy the greater part of the error due to this
defect, by measuring experimentally the relative (or absolute)
values of the scale-division in different parts of the range. There
will, however, remain uncorrected some irregularity, due to
influence of the distribution of electricity over the uncoated inner
surface, in the instruments as hitherto made, in all of which the
inner surface of the jar is coated with tinfoil only below the
guard-plate, so that the upper surface of the guard-plate may be
seen clearly, in order that the observer may always see that all
is in order about the aluminium square and aperture round it;
and particularly that there are no injurious shreds or minute
fibres. But the irregular influence of the electrification of the
uncoated glass, if found sensible, will be rendered insensible
by continuing the tinfoil coating an inch above the upper surface
of the guard-plate.

§ 33. All faults, except the temperature error, depend on the
smallness of the instrument ; and if the observer chooses to regard
as portable an instrument of thirty centimetres (or a foot) diameter,
with all other dimensions and all details of construction the same
as those of the instniment described above, he may have a portable
electrometer practically free fix)m three of the four faults described.
But it is scarcely to be expected that a small instrument (12^
centimetres high, and 8^ centimetres in diameter) which may be
carried about in the pocket can be ft^e firom such errors. They
are, however, so far remedied as to be probably not perceptible in
the large stationary instrument which I now proceed to describe.

for electrical measurements 249

Standard Electrometer.

§ 34. This instrument (figs. 12, 13, and 14, Plate 6) differs
from the portable electrometer only in dimensions, and in certain
mechanical details, which are arranged to give greater accuracy
by taking advantage of freedom from the exigencies of a small
portable instrument. It is at present called the standard elec-
trometer, in anticipation of either remedying or of learning to
perfectly allow for the temperature error, and of finding by secular
experiments on the elasticity of metals that their properties used
in the instrument are satis&ctory as regards the permanence from
year to year, and from century to century, of the electric value of
its reading. It is an instrument capable of being applied with
great ease to very accurate measurements of differences of potential,
in terms of its own unit. The value of the unit for each such
standard instrument ought, of course, to be determined with the
greatest possible accuracy in absolute measure ; and until con-
fidence can be felt as to its secular constancy, determinations should
frequently be made by aid of the absolute electrometer.

§ 35. The Leyden jar of the standard electrometer consists
of a large thin white-glass shade coated inside and outside to
within G centimetres of its lip, and placed over the instrument as
an ordinary glass shade, to protect against dust, currents of air,
and change of atmosphere. It may be removed at pleasure from
the cast-iron sole of the instrument, and then the interior works are
seen, consisting of: —

(1) A continuous disk of brass supported on a glass stem, in
prolongation of a stout brass rod or tube sliding vertically in Vs,
in which it is kept by a spring, and resting with its lower flat end
on the upper end of a micrometer-screw shaft, shown in fig. 13,
where the screw, graduated circle, and stout brass rod are as seen
in the instrument ; the perforated brass disk (which is intended
to keep the round upper end of the screw-shaft in position) is
shown in section in fig. 14.

(2) Resting on three glass columns, a guard-plate with a
square aperture in its centre, and carrying on its upperside
stretching-springs and thin platinum-wire suspension of an
aluminium balance-lever, shaped like those of the gauge (§ 13)
and the portable electrometer (§ 23) already described, but some-
what larger. The tops of the three glass columns are rounded ; a

260 PftACTICAL STANDARDS

round hole and a short slot in line with this hole are cut in the
guard-plate and receive the rounded ends of two of the columns,
which are somewhat longer than the third. The flat smooth lower
sur&ce of the guard-plate rests simply on the top of the third
glass column. The diametet of the round hole and the breadth

of the slot in the guard-plate may be about -^ of the diameter

of curvature of the upper hemispherical rounded ends of the
glass column, so that the bearing portions of the rounded ends in
the round hole and in the slot respectively may be inclined some-
where about 45'' to the plane of the plate. This well-known but
too often neglected geometrical arrangement gives perfect steadi-
ness to the supported plate, without putting any transverse strain
upon the supporting glass columns, such as was almost inevitable,
and caused the breakage of many glass stems, before mental
inertia opposing deviations from the ordinary instrument-maker's
plan (of screwing the guard-plate to brass mountings cemented to
the tops of the glass columns) was overcome. It has also the
advantage of allowing the guard-plate to be lifted off and replaced
in a moment.

(3) Principal electrode projecting downwards through a hole
in the sole of the instrument, and rigidly supported from above
by a brass mounting cemented to the top of a thick vertical glass
column, connected by a light spiral spring with the lower attract-
ing-plate moved up and down by the micrometer-screw. The
aperture round the principal electrode may be ordinarily stopped
by a perforated column of well-paraflSned vulcanite projecting
some distance above and below the aperture, which I find to
insulate extremely well, even in the smoky, dusty, and acidulated
atmosphere of Glasgow. When an extremely perfect insulation of
the principal electrode and connected attracting-plate is required,
the vulcanite stopper surrounding it may be removed, so that the
only communication between the electrode and the case of the
instrument may be along the two glass columns in the artificially
dried interior atmosphere of the case ; but from day to day, when
the instrument is out of use, the aperture round the principal
electrode should be kept carefully stopped, if not by a vulcanite
insulator, by a perforated cork (although I find but little loss of
insulation, either by the inner glass surface of the Leyden jar or by
the three glass columns, when this precaution is neglected).

FOR £LECTB[CAL MEASUREMENTS 251

(4) Temporary charging-rod supported by a vertical perforated
column of paraffined vulcanite, or a glass tube well varnished
outside and thickly paraflBned inside. The insulating column
bearing this charging-rod is turned round till a horizontal spring
projecting from its upper end touches the iimer coating of the jar,
when this is to be charged frt)m an independent source, or when, for
any other experimental reason, it is to be put in connexion with a
conductor outside the case of the instrument

(5) A small replenisher of the kind described for the quadrant
electrometer (§ 12), but with much wider air-spaces to prevent
discharge by sparks.

(6) A large glass or lead dish to hold as large masses of
pumice as may be, which are to be kept sufficiently impregnated
with strong sulphuric acid.

§ 36. A considerable portion of the jar above the guard-plate
is left uncoated to allow the observer to see easily the hair and
white background with black dots ; also several other smaller parts
of the glass above the guard-plate are left uncoated to admit light
to allow a small circular level on the upper side of the guard-plate
to be seen. The long arm of the aluminium balance-lever is very
thoroughly guarded by double cages and fences of wire (§ 25), so
that it can experience no sensible influence from electric disturbing
forces when the covering jar is put in position and electric con-
nexion is established between its inner coating and the guard-
plate by projecting flexible wires or slips of metal.

§ 37. The aluminium square plate is somewhat larger and
the platinum bearing wire somewhat longer in this instrument
than in the portable electrometer, to render it sensible to smaller
differences of potential. The step of the screw is the same as in
the portable (-^ of an inch), and one division (j^ of the circum-
ference of the screw-head) corresponds to a difference of potentials
which, roughly speaking, is equal to about that of a single cell of
Danieirs. The effective range of the instrument is about sixty
turns of the screw, and therefore about 6000 cells of DanielFs;
that of the portable electrometer is about 15 turns of the screw
(equivalent to about 1500 cells). Neither of these instruments
has sufficient range to measure the potential to which Leyden jars
are charged in ordinary electric experiments, or those reached by
the prime conductor of a powerful electric machine. The sta-
tionary instrument with its long screw and its large plates now

252 PRACTICAL STANDARDS

described would go far towards meeting this want if its aluminium
lever and platinum suspension were made on the same scale as
those of the portable electrometer ; but for an instrument never
wanted to directly measure differences of potentials of less than
two or three thousand cells, the heterostatic (§ 40) principle is in
general not useful ; and therefore I have constructed the following
very simple idiostatic (§ 40) instrument, which is adapted to
measure with considerable accuracy differences of potential from
4000 cells upwards to about 80,000 cells.

Long-range Electrometer.

§ 38. In this (fig. 15, Plate 6) the continuous attracting-
plate is above, and the guard-plate with aluminium balance below,
as in the portable electrometer; but, as in the standard stationary
electrometer, the upper plate is fixed and the lower plate is moved
up and down by a micrometer-screw. The mechanism of the
screw and slide has all the simplicity and consequent accuracy
of that of the standard electrometer. In the only long-range
instrument yet constructed the step of the screw is the same as
that of the others (^ of an inch). In future instruments it
would be well either to have a longer step or to have a simple
mechanism (which can be easily added) to give a quick motion,
as in the use of the present instrument the turning of the screw
required for great changes of the potential measured is very
tedious. The guard-plate projects by more than an inch all
round beyond the rim of the upper attracting-plate — partly to
obviate the necessity of giving it a thick rim, which would be
required to prevent brushes and sparks originating in it if it
had only the same diameter as the continuous plate above, and
partly to guard the observer from receiving a spark or shock in
measuring the potential of an electric machine or of a Leyden
battery, and to prevent the hair from being attracted to the
upper plate. Thus the guard-plate is allowed to be no thicker
than suffices for stiffness ; and this allows the observer to see the
hair at the end of the aluminium balance-lever without the lever
being made of a djrnamically disadvantageous shape, as would be
necessary if the guard-plate were thick or had a thick rim added
to it. No glass case is required for this instrument. The small-
ness of the needle and the greatness of the electric force acting
on it are such that I find in practice no disturbance to any

FOR ELECTRICAL MEASUREMENTS 253

inconveniejit degree by ordinary currents of air ; although it and
all these attracted disk instruments show the influence of sudden
change of barometric pressure, such as that produced by opening
or shutting a door. If not kept under a glass shade when out of
use, the lower surface of the upper attracting-plate and the
lower surface of the guard-plate and attracted aluminium square
should be carefully dusted by a dry cool hand. Generally speaking,
none of the vital electric organs of an electrometer should be
touched by a cloth, as this is almost sure to leave shreds fatal to
their healthy action.

§ 39. The effective range of this instrument is about 200
turns of the screw ; rather greater force of torsion is given than
in the portable electrometer, and a rather smaller attracted disk
may be used, so that upwards of four cells may be the electric value
of one division. The instrument in its present state measures
nearly, but not quite, the highest potential I can ordinarily produce
in the conductor of a good Winter's electric machine, which some-
times gives sparks and brushes a foot long.

§ 40. The classification of electrometers given above is
founded on the shape and kinematic relations of their chief organic
parts ; but it will be remarked that another principle of classifica-
tion is presented by the different electric systems used in them,
which may be divided into two classes : —

I. Idiostatic, that in which the whole electric force depends
on the electrification which is itself the subject of the test.

II. Heterostatic, in which, besides the electrification to be
tested, ano'ther electrification maintained independently of it is
taken advantage of.

Thus, for example, the long-range electrometer (§§ 38, 39) is
simply idiostatic and is not adapted for heterostatic use; but
each of them may be used idiostatically. The absolute electro-
meter was at first simply idiostatic (^ 17-21); more recently it
has been used heterostatically, and is about to acquire (§ 22)
special organs adapted for heterostatic use ; as yet, however, no
species of the absolute electrometer promising permanence has
come into existence.

§ 41. It is instructive to trace the origin of various hetero-
static species of electrometers by natural selection. A body
hanging, or otherwise symmetrically balanced, in the middle of a
symmetrical field of force, but firee to move in one direction or the
other in a line tangential to a line of force, moves in* one directioA

254

PRACTICAL STANDARDS

or the opposite when electrified positively or negatively. Bohnen*
berger's arrangement of this kind has a convenient and approxi-
mately constant field of force; and his instrument was chosen
in preference to others which may have been equally sensitive,
but were less convenient and constant, and it became a permanent
species.

§ 42. Bennet's gold-leaf electroscope, constructed with care
to secure good insulation, electrified sufficiently to produce a
moderate divergence, has been often used to test, by aid of this
electrification, the quality of the electrification of an electrified body
brought into the neighbourhood of its upper projecting electrode,
causing, if its elasticity is of the same sign as that of the gold
leaves, increase of divergence ; if of the opposite sign, diminution.
By connecting the upper electrode with the inner coating of a
Leyden jar with internal artificially dried atmosphere, the charge
of the gold leaves may be made to last with little loss from day
to day ; and by insulating Faraday's metal cage (§ 2) round the
gold leaves and alternately connecting it with the earth and with
a conductor whose difference of potentials firom the earth is to be
tested, an increase or a diminution of divergence is observed
according as this difference is negative or positive, the gold leaves
being positive. Hence (through Peltier's and Delmann's forms)
the heterostatic stationary and portable repulsion electrometers,
described in the Royal Institution Lecture on "Atmospheric
Electricity" and in Nicholas CyclopcBdia, article "Electricity,
Atmospheric," already referred to, of which one species still sur-
vives in King's College, Nova Scotia, and in the Natural Philosophy
Class-room of Edinburgh Uni-
versity. The same form of the
heterostatic principle applied
to Snow Harris's attracted-disk
electrometer gave the portable
and standard electrometers de-
scribed above.

§ 43. A modification of
Bohnenberger's electroscope, in
which the two knobs on the two
sides of the hanging gold leaf
became transformed into halves
of a circular cylinder, with its axis horizontal and the gold leaf
hung on a wire insulated in a position coinciding with its axis,

FOR ELECTKICAL MEASUREMENTS 255

producing a species designed for telegraphic purposes, but which
did not acquire permanence by natural selection, and is only
known to exist in one fossil specimen. In this instrument the
wire bearing the gold leaf was connected with a charged Leyden jar,
and the semicylinders with the bodies whose difference of potential
was to be tested. But various modifications of the divided-
cylinder or divided-ring class with the axis vertical and plane of
motion horizontal have done some practical work, and one
species, the new quadrant electrometer (§ 6), promises to become
permanent.

§ 44. The heterostatic principle in one form or other is
essential to distinguish between positive and negative. As re*
marked above (§ 42), the original type of this use of it is to be
found in the old system of testing the quality of the charge taken
by the diverging straws or gold leaves of the electroscopes used
for the observation of atmospheric electricity, which was done by
bringing a piece of rubbed sealing-wax into the neighbourhood,
and observing whether this caused increase or diminution of the
divergence. A doubt which still exists as to the sign* of the
atmospheric electricity observed by Professor Piazzi Smyth on
the Peak of Teneriffe, is owing to the imperfection of this way of
applying the principle. It is, indeed, to be doubted in any one
instance whether it is not vitreous electricity that the rubbed
sealing-wax acquires ; and, again (§ 2), it is not certain that the
glass case enclosing the gold leaves, especially if very clean and
surrounded by a very dry natural atmosphere, screens them suffi-
ciently from direct influence of the piece of sealing-wax to make
sure that the divergence due to vitreous electricity could not be
increased by the presence of the resinously electrified sealing-wax
if held nearer the gold leaves than the upper projecting stem.

§ 45. The heterostatic principle has a very great advantage
as regards sensibility over any simple idiostatic arrangement, inas-
much as, for infinitely small differences of potential to be measured,
the force is as the squares of the differences in any idiostatic
arrangement, but is simply proportional to the difference in every
heterostatic arrangement.

* NiohoPs Cyclopadia, article ** Eleotrioity, Atmospheric,'* edition 1860.

256 PRACTICAL STANDARDS

VL Deterrmnation of the Dynamical Equivalent of Heat from the
thermal effects of Electric Currents. By J. P. Joule, D.C.L,
F.R.S., etc.

Sir W. Thomson, as long ago as 1851, showed that it was
desirable to make experiments such as are the subject of the
present paper. They have necessarily been delayed until a suffi-
ciently accurate method of measuring resistance was discovered.
Such a method having been described by Sir William, and carried
out into practice by Professor C. Maxwell and his able coadjutors^
the task ^issigned to me by the Committee of Electric Standards
was comparatively simple.

My experiments were commenced nearly two years ago, and
the apparent ease with which they could be executed gave
promise of their early completion. It was, however, found essen-
tial that careful observations of the earth's horizontal magnetic
intensity should be frequently made, and these required the con-
struction of apparatus whereby this element could be determined
with accuracy and rapidity.

The apparatus finally adopted for this purpose consists of a
suspended horizontal flat coil of wire between two fixed similar
coils. A current of electricity can be made to traverse all three,
communication with the suspended coil being made by the
suspending wires themselves according to Sir W. Thomson's plan.
The strength of a current is found by observing the sum of the
forces of attraction and repulsion by which the suspended coil is
urged. The strength of a current can in this manner be deter-
mined in absolute measure ; for the area of each of the three
equal coils being called a, the weight required to counterpoise the
force with which the suspended one is urged w, the force of
gravity g^ and the length of wire in each of the coils Z, the

current c = ^ a/ -^— (1 -h correction), the correction being prin-
cipally due to the distance between the fixed coils. In my instru-
ment, in which this distance is 1 inch, the diameter of the coils
being 12 inches and their interior core 4 inches, this correction
was proved by experiment to be '1185.

There was, however, considerable difficulty in obtaining an exact
measure of the distance between the fixed coils; and I therefore

FOR ELECTRICAL MEASUREMENTS 257

judged that the measure of the currents used in the experi-
ments would be most accurately obtained by means of a tangent
galvanometer, the above-described current-meter being employed
to determine the horizontal intensity.

This determination was effected as follows: — Many careful
observations of the horizontal intensity by an improved method
on Gauss and Weber's system were made alternately with obser-
vations of the deflections of a tangent galvanometer and the
weighings of the current-meter when the same currents traversed
both instruments in succession. Then calling the horizontal in-
tensity H, the angle of deflection 6, and the weighing w, there

was obtained a constant c = — — — = 0*17676. Hence with these

. rr 0-17676 Vw
mstruments -a = — - — ^ — •

The experiments for the determinations of horizontal intensity
by the use of this formula could be effected in a few minutes, and
did not require an alteration in the disposition of any part of the
apparatus. It was satisfactory to find that, although the presence
of masses of iron at only a few yards distance made the field in
which I worked considerably more intense than that due to the
latitude, and although I worked at different times of the day,
the highest intensity, out of upwards of seventy observations
distributed over a year, was 3*6853, and the lowest 3*6607, in-
dicating a much greater degree of constancy than might have
been expected.

The galvanometer above mentioned was that employed in the
thermal experiments. It had a single circle of ^^-inch copper
wire, the diameter of which, being measured in many places by a
standard rule, gave a radius of 0*62723 of a foot. The needle was
half an inch long, and furnished with a glass pointer traversing a
divided circle of 6 inches diameter. In the experiments the
deflections were not far fix)m 26° 34', the angle at which the
influence of the length of the needle within certain limits is
inappreciable. It was easy by a magnifier, arranged so as to avoid
parallax, to read to one minute. The torsion of the fibre gave
only 3''5 for an entire twist. The trifling correction thus required
is applied to the recorded observations of deflection.

The calorimeter first used was a copper vessel upwards of a
gallon in capacity, filled with distilled water. It had a conical

B. A. 17

258 PRACTICAL STANDARDS

lid, atteiched by screws, in which were two tubulures, one for the
introduction of a copper stirrer, the other for the thermometer,
around the immersed stem of which a wire of platinum silver,
having a resistance nearly equal to that of the Association unit,
was coiled.

The resistance of the wire was found by comparing it with the
Association unit, sent me by the Committee, using Ohm's formula^

R = -^{ r^ — p" )' w^®^® ^8' ^s» *°^ ^1 ^^ ^^^ tangents of deflection

with the battery and connexions only with these and the unit and
with the coil respectively. This, though by no means so delicate
a method as that of the Wheatstone balance improved by Thomson,
was able to give a final result certainly accurate to the two-
thousandth part. The results for the resistance of the coil in the
first series of experiments are as follows. They were obtained
before and after those experiments. A large galvanic cell, con-
sisting of cast iron and amalgamated zinc plunged in dilute
sulphuric acid, was the source of electricity, which was measured
by a galvanometer with a coil of nine turns, 17 inches in
diameter.

c, c. c. '^^"sr^r ""T^"" ^^^'

of unit

tan 65 6*75 tan 28 18 tan 28 13 637 62*65 1-01901

tan 59 32-5 tan 32 396 tan 32 22 5924 58*39 1*01825

The average resistance 1*01863 being reduced fix)m the tem-
perature 14° '5 Cent., at which the unit was adjusted, to 69* 9
Fahr., the average temperature of the calorimeter in the first
series of experiments, becomes 1*0191, which, multiplied by
32808990, gives 33435640 as the resistance in British absolute
measure*.

A delicate thermometer was placed at a few inches distance
from the calorimeter, for the purpose of registering the temperature
of the air. In the Tables its indications are reduced to the scale
of the instrument plunged in the calorimeter. A string attached

* Note added in 1911 : The resistance is expressed in terms of the foot and the
■econd. Written folly the conversion is

(a0'48om. = l foot.)

10191 ohm = .^^,Z, = 82808990.

FOR ELECTRICAL MEASUREMENTS

259

the handle of the stirrer to a stick, so that the water could be
effectually stirred without communicating the heat of the hand.
A wooden screen separated the observer from the apparatus.

In the experiments of the first series a battery of five large
DanieU's cells, arranged in series, transmitted the current through
the coil for 40 minutes exactly, determined by chronometer.
During this time twenty-eight observations of deflection were
obtained, seven at each end of the pointer directed N.E. and S.W.,
and seven when it was directed N.W. and S.E. by reversing the
current in the galvanometer for the latter half of the time. The
water was stirred twenty-eight times. Its temperature was taken
at the beginning, middle, and end of an experiment. There were
also fourteen observations of the temperature of the air.

Immediately after each experiment the horizontal intensity of
magnetic force was obtained by observing the deflection of the
galvanometer and the weighing of the current-meter produced by
the same current.

Before and after each experiment, two others were made in
precisely the same manner, but excepting the current, in order to
discover the influence of radiation and the conducting power of
the atmosphere.

First Series of Thermal Experiments.

Date

Deflection

tan^
Deflection

Temperatare
of air

Temperatare
of water

Rise of
temperature

Horizontal
intensity

1866
Aug. 22...

» 23...
Sept 8...

„ lO...

M 11...

n 12' ••
1, 13...

n 15...

„ 15...

yy 18...

32 46-86
34 0-29
32 24-83
31 50-22
31 31-02
31 14*42
30 67-51
30 24-86
30 20-61
30 34*34

•414719
•465133
•403156
-386642
•376024
•367944
•369850
-344607
•342610
•348982

492-36*

494-77

400-4

44111

367-0

344*33

361 -54

346-7

381-41

342-64

497-42

493-27

401-8

433-85

392-89

344-45

368-47

330-01

367-56

324-32

23-56

26-65

22-8

22-214

18-61

21-9

20-95

21-98

21-07

22-29

3-6763
3-6815

3-6737
3-6758
3-6656
3-6671
3-6638
3-6711
3-6607

Average...

•379857

397-226

394-406

22-0914

3-67073

* 12-951 dinsions of the thermometer are equivalent to 1** F.

17—2

S6d

PRACTICAL STANDARDS

First Series of Radiation Experiments.

Date

1866
Aug. 22.

Aug. 23.

Sept. 8.

Sept. 10.

Sept. 11.

Sept. 12.

Sept. 13.

Sept. 15.

n •
99 •

Sept. is!
j>

Average.

Temperatare
of air

Temperatare
of water

495-93

469-14

502-22

477-83

476-37

458-96

490-81

499-22

393-5

382-75

395-82

414-15

444-31

419-4

437-15

396-96

373-07

384-72

36714

391-76

334-0

332-42

365-34

360-2

352-82

343-11

366-65

36916

330-78

315-41

381-47

347-14

378-93

350-67

381-05

379-51

326-99

309-28

339-9

338-35

Rise of

temperatare

of water

373-058

2-88
315
3-08
0-55
2-0
1-7
2-9
4-83
0-63
1-75
0-44
1-6
1-83
•008
2-78
3-72
3-34
0-22
2-55
004

364-686

1-3806

— la Applying the preceding Table for the purpose of correcting
the results of the thermal experiments, it must be first observed
that the external influences on the calorimeter are not zero when
the temperature of the air-thermometer coincides with the in-
dication of that immersed in the calorimeter. This might arise
partly from the locality of the two instruments not being the
same, but was, I found, principally owing to the different radiating
and absorbing powers of the air- thermometer bulb and of the
surface of the calorimeter. Taking, then, the number of instances
in which the temperature of the air appeared to exceed that of the
water, there are fifteen, with a total excess of 259*63 and a re-
sulting gain of temperature of 35*36 ; also those in which the air
appeared to be colder than the water were five, giving a total
deficiency of 65*5 with a loss of temperature 47 1. Hence

65*5 -5a? 259-63 + 15a?

471

35*36

FOB ELECTRICAL MEASUREMENTS 261

whence a; =» 4*418, which must be added to the indications of the
thermometer registering the temperature of the air. After, this
correction has been made, it will be found that the effect of a
difference of temperature between the air and water of 9*216 is
unity.

4*418 added to 397*226 gives 401*644 for the corrected tem-
perature of the air in the thermal experiments ; and this being
7*238 in excess of the tempeirature of the calorimeter, the corrected
thermal effect will be

22*0914-^ = 21*306,

which, after applying the needful correction for the immersed
portion of the thermometer-stem, becomes ultimately 21*32&,

The thermal capacity of the calorimeter was made up of
95525 grains of distilled water, 26220 grains of copper, equivalent
to 2501 grains of water^^ and the thermometer and coil equivalent
to 80 grains, giving a total capacity equal to 98106 grains of
water. 12*951 divisions of the thermometer are equivalent to one
degree Fahr.

The dynaitiicdl equivdleht is the quotient of the work done by
the thermal effect, or

tan* em-

j J2723 ^ 3.57073) ' ^ 379867 x 33435640 x 2400

\^'^^^ ) . s 25335 *.

It appeared to be desirable to diminish the atmospheric
influence; I therefore commenced a second series, in which the
calorimeter was covered with two folds of cotton wadding. The
bulb of the air-registering thermometer was also placed in a small
bag made of the same material. In this fresh series each ex-
periment occupied one hour, as I had learned by experience that
with my battery arrangement the current would be suflSciently
uniform. In fact the highest reading in an experiment was not

* Note (1911) : The units are the foot, the grain, and the second, with the degree
Fahrenheit. In o.o.b. measure and the degree Centigrade, the value becomes: —

.25835 X 9/5 x (d0-48)>'=> 42-866 x lO^.

262

PRACTICAL STANDARDS

more than ^ higher than the lowest. There were, evenly dis*
tributed through the hour, forty observations of deflection, twenty
of the air, and three of the water-thermometer ; and the water
was stirred forty times. Two minutes were allowed for the
complete equalization of temperature previous to the final thermo-
meter reading. The experiments on radiation were also similarly
extended.

The coil was the same as that used in the first series ; it had a
coat of shellac varnish. Five determinations of its resistance were
made, using a single Daniell's cell with various resistances in-
cluded in the circuit. The galvanometer had a coil 17 inches in
diameter consisting of nine turns. The results are as follow : —

Temperatare
of unit

Temperatnre
of ooil

Resistance of

coil in tenns

of unit

tan 79 39*5
tan 71 39-5
ten 70 16
tan 71 54*33
ten 62 6

ten 52 33-3
ten 47 1706
ten 46 18*11
tan 47 7-66
ten 41 30*43

ten 52 9*3
ten 46 55*6
ten 45 57*4
ten 46 45*93
ten 41 13*46

59-25

48*6

54*68

58*6
48*5
57*4

1*0192
1*0198
1*0194
1*0198
1*0187

Average ...

1*01938

The average temperature of the calorimeter in the experiments
being 13° '55 Cent., and that at which the unit was adjusted 14'' *5,
the resistance during the experiments must have been 1*01906,
which is equal to 33434330 in British measure.

The correction to be applied to the thermometer immersed in
air as deduced from the above Table is given by

123-66 -lar 356-65 + I&1:

12-74

30-99

whence a;B~l'1835. It appears also that a difference between
the temperatures of the calorimeter and air-registering thermo-
meter so corrected, equal to 10*822, gives the unit effect on the
former.

Hence the corrected indication of the air-thermometer in the
second series of thermal experiments will be

349-63 - 1-1835 == 348-4465.

FOR ELECTRICAL MEASUREMENTS

263

This being 12*5345 in excess of the temperature of the calori-
meter, the corrected thermal effect will be

12-5345

25-65 -

= 24*4917,

10-822

which, after a small further correction for the immersed stem,
becomes 24*512.

The thermal capacity in this second series was made up of
95561 grains distilled water, copper as water 2501, thermometer
and coil as water 80, and cotton-wool as water 200 grs., giving a
total of 98342 grains.

The equivalent, as deduced from the second series, is therefore

USS ^ ^'^^^4' ^ '2^^^*^ ^ 33434330 x 3600

24-512
12-951

= 25366.

X 98342

The equivalents obtained in the two foregoing series of ex-
periments are as much as one-fiftieth in excess of the equivalent
I obtained in 1849 by agitating water. I therefore instituted a
strict inquiry with a view to discover any causes of eiror, so that

Second Series of Thermal Experiments,

Date

Defleotion

tan<
Deflection

Temperature
of air

Temperatare
of water

Else of
temperature

Horizontal
intensify

1866

Sept 21...
n 22...
,f «0 ...

„ 26...

»> 27...

Oct. 5...

>» "•••
n S'"

„ 20...
„ 22...
„ 23...
„ 26...
„ 26. ••
„ 27...

29 51-68

28 58-4

29 14-63
29 51-46

28 54-78

29 6-05
28 22-54
28 8-74
28 42-81
27 40-13

26 40-5

27 28-1
27 9-63

27 42-56

28 7-84

-329623
•306585
-313472
-329526
•305064
-309393
-291761
•286198
•300074
•274910
-252409
•270252
•263230
•275855
•286838

397^4

362-51

346-19

370-84

365-91

380-66

426-55

338*49

398-56

395-18

371-72

320-07

275-65

249-75

245-96

363-42
348 06
386-94
360-64
361-71
387-57
392-77
335-54
332-35
3&1-90
388-63
318-09
286-25
257-54
247-27

30-38
26-95
29-75
29-92
25-88
24-90
27-40
24-04
31-08
26-08
19-12
22-55
20-98
2215
23-67

3-6668
3-6707
3-6724
3-6644
3-6665
3-6612
3-6688
3-6595
3-6659
3-6654
3-6702
3-6638
3-6620
3-6623
3-6641

Average...

•292946

349-63

335-912

25-65

3-6656

264

PRACTICAL STANDARDS

Second Series of Radiation Experiments.

Date

Temperature
of air

Temperature
of water

Rise of

temperature

of water

1866
Sept 21

Sept 22!!!!!!

Sept 25

Sept 26!!!!!!

Sept. 27!!!!!!

Oct 5

Oct" 6!!!!!!
»i

Oct 8

Oct 19!!!!!!
Oct? 20!!!!!!
Oct? 22...!..
Oct? 23!!!!!!
Oct'- ^r.;!T!

Oct? 26!!!!!!

Oct? 27!!!!!!

378-84

390-13

326-32

360-71

330-67

347'56

352-15

377-56

365-81

388-0

376-9

385-8

402-94

433-28

319-5

356-02

365-08

398-49

357-9

395-66

371-24

362-7

297-96

33407

261-67

277-59

233-31

264-37

237-05

251-15

344-95

381-34

334-37

361-13

287-94

326-13

33312

36812

347-9

375-69

375-04

396-95

376-47

411-33

323-51

347-79

303-94

356-29

344-01

377-40

380-45

392-44

305-0

329-05

277-01

294-31

247-61

265-97

234-85

257-24

3-0

0-32
-0-43
-0-41

4-06

1-59

212

0-70

0-74

1-31

0-
-1-15

2-13

1-52
-0-29

0-33

5-96

3-57

1-61

1-43
-0-95
-3-18
-0-50

0-5
-1-26
-1-86
-1-40
-0-66

0-1 -
-0-65

Average

343-011

335-245

0-6083

they might be avoided in a fresh series. The most probable
source of error seems to be insufficient stirring of the watef'bf the
calorimeter. Although agitated so frequently as forty times in
the hour, there could be no doubt that, during any intervals of
comparative rest, a current of heated water would ascend from the
coil, and that if a thin stratum of it remained any time at the top,
some loss of heat would result. I resolved therefore to'!use a
fresh calorimeter, and to introduce into it a stirrer which could be
kept in constant motion by clockwork.

Another source of error which, though it would be finally |

FOR ELECTRICAL MEASUREMENTS 265

eliminated by frequent repetition of the experiments, it seemed to
be desirable to avoid, was the hygrometric quality of the cotton-
wool which enveloped the calorimeter in the second series of
experiments. I therefore sought for a material which did not
present that inconvenience. The plan finally adopted was to
cover the calorimeter first with tinfoil, to place over that two
layers of silk net (tulle), and to finish with a second envelope of
tinfoil.

A third source of possible error was the circumstance that the
silver-platinum alloy, when made positively electrical in distilled
water, is slowly acted upon, an oxide of silver as a bluish-white
cloud arising from the metal, while hydrogen escapes from the
negative electrode. On this account the coil in the experiments
of the last series, as well as the subsequent, was well varnished.
But it was found at the conclusion of the experiments that the
varnish had in a great measure lost its protecting power. This
circumstance gave me considerable anxiety: I was, however,
-ultimately able, by the following facts arrived at after the thermal
experiments were completed, to satisfy myself that no perceptible
influence had been produced by it on the results : —

1st. The resistance of the coils, afber long-continued use had
deteriorated the varnish, was not sensibly less than it was after
they had been fireshly varnished.

2nd. The coil of the 3rd series (p. 267) was, in the unprotected
state, immersed in distilled water, and compared with many hundred
yards of thick copper wire, unimmersed, having nearly equal
resistance. The result showed that the resistance to the current
was sensibly the same whether a single cell or five cells of Daniell
in a series were used. Now, had any considerable leakage by
electrolytic action taken place, it would have been very much less
in proportion in the former than in the latter instance.

3rd. When the coils of the second and third series, in the un-
protected state, were placed in distilled water, and made the
electrodes of a battery of five cells, the deflection was 40' of a
degree on a galvanometer with a coil of 17 inches diameter com-
posed of 18 turns of wire. This deflection indicates a current pf
about ^ of the average current in the thermal experiments. In .
this.. case the chemical action was distinctly visible, but quite
ceased to be so when the electrodes were connected by a wire of

266

PRACTICAL STANDARDS

unit resistance, so as to reduce the potential to that in the
thermal experiment&

4 th. The ooil of No. 2 series being used as a standard, that
of No. 3 series, in the unprotected condition, was immersed, first
in water, then in oil The resistance to the current of five
DanieU's cells was found to be sensibly equal in the two cases.

Hence there could be no doubt that the loss of heat during
the experiments by electrolytic action could not possibly in any
instance have been so great as one-thousandth of the entire effect,
and was probably not one-quarter of that small quantity ; whilst
in the larger number of experiments, when the varnish was firesh,
it must have been nil.

The coil used in the third series of experiments was made by
bending four yards of platinum-silver wire double, and then coiling
it into a spiral, which was supported and kept in shape by being
tied with silk thread to a thin glass tube. The terminals were
thick copper wires, and the whole was coated with shellac and
mastic varnish. The following results were obtained for its re-
sistance. In the first three trials the current was measured by a
galvanometer with a circle of nine turns 17 inches diameter, and
in the last six with an instrument with eighteen turns of wire.
In the first six there was an extra unit of resistance included in
the circuit : —

Battery

Unit

Temp.

of
unit

Temp.

ooil

Berisfc.
anoe in
termeof
my unit

OneceU,Daniell
Ditto

Mine ...
Jenkin's

„

Mine ...
„ ...

« •••

Jenkin's

Mine ...

tan 52 53
tan 52 2412
tan 52 3*62

tan 50 25-8

tan 49 4812
tan 48 17*62
tan 75 28
tan 75 17*25
tan 75 59*6

iATi 37 3-15
tan 36 29*02
tan 36 6*45

tan 35 21*88

tan 34 57*36
tan 34 5*48
tan 49 58*6
tan 49 44*93
tan 49 18*97

tan 37 10*6
tan 36 37*27
tan 36 14*79

tan 35 29*27

tan 35 5*62
tan 34 12*24
tan 50 11*98
tan 49 57*51
tan 49 33*08

63-27
59-03
60*88

59-78

60-03
60-50
61-27
61-96
69-36

62-78
60-07
00-57

60-46

00-30
60-88
61-08
61-27
70-28

-98963
-98823
•98752

-98818

•98754

-98816
-98863
-98871
-98820

Ditto

Daniell'8 cell.)
PoBitive V
metal iron j

Ditto

A «r4k**A AA

1
.... '

-98831

^.vexagu

FOR ELECTRICAL MEASUREMENTS

267

Third Series of Thermal Experiments,

Date

Defleotion

tan*
Defleotion

Tempera-
tare of

air

Tempera-
ture of
water

Rise of
tempera-
ture

Fall of
weight

1867

June 28, 12.54 p.m.
„ 28, 5.36
„ 29, 1.30

July 1, 10.30 a.m.
„ 1, 4.24 p.m.
„ 2, 12.45
„ 2, 6.0
„ 4, L20
„ 20,11.11a.m.
„ 20, 3.45 p.m.
„ 22, 12.36
„ 22, 5.21
„ 23, 1.7
„ 24,11.0 A.M.
„ 24, 4.5 P.M.
„ 25, 12.15
„ 25, 4.55
„ 26, 12.58
„ 27, 11.13a.m.
„ 27, 4.14 p.m.

Aug. 2,12.31
„ 2, 5.18
„ 3, 12.56
„ 6, 11.18a.m.
„ 6, 3.55 p.m.
„ 8, 12.17
„ 8, 5.45
„ 9, 1.27

„ 10,11.9 A.M.

„ 10, 3.56 p.m.

& 18-25
30 56-37

28 55-45

29 411

30 19-4
30 1012
30 30*98
3123-4

30 21-72

31 37-55

32 0-6
32 23-47
31 18-43

31 4-75

30 49-15

32 39-5

33 10

32 33-95

33 1-6
32 58*22

31 52-98
31 53-77
31 3718
26 34-35

28 42*8

29 29*25
29 39-25
29 33-2
29 12-65
28 14-47

-290024
•359310
*305345
-324949
•342107
•337891
•347424
•372299
•343170
•379241
•390765
•402470
•369881
•363299
-355900
-410832
•427129
•407920
•422590
•420777
•386923
•387325

•379056
•250162
•300070
•319773
•324137
•321491
•312626
•288500

488-660

534155

509-172

428-81

508-78

4a5-343

401-822

516-992

385-622

454-19

482-44

493-087

465-238

430-688

439-007

465-354

521-569

445-009

391-0

418-11

385-876

407-781

453-66

439-906

457-145

465-586

499-874

478-a'>8

468-344

519-082

494-17

524-214

490-13

425-67

467-214

450-73

458-104

462-97

394-0

430-97

460-621

498-573

473-167

448*043

470-954

432-46

486-049

464-267

419-21

446-623

390-911

422-843

421-948

435-699

462-056

443-204

480-564

469-296

455-304

493-136

251

32-08

27-82

28-52

3305

25-13

24-99

57-98

28-98

34*92

35-48

34*47

31*27

30*24

28-14

38-48

39-72

33-61

34-46

34-09

331

32-25

35-37

22-32

25*67

29-6

29*67

28*8

28*21

27*28

in,

30*6

28*4

28*7

27*9

28*2

29*4

28*4

29-4

29-75

29*7

296

29^7

264

27*4

28^4

A VOFAfle ••«.«•■*•...

-3547795

458-699

455-436

31-02666

28*362

The above average resistance, reduced to 18° '63 C, the mean
temperature in the third series, is 0*98953 of the Association unit,
or in British measure 32465480.

In the third series, the experiments for the heat of the current,
of radiation, and for horizontal magnetic intensity were alternated
in such a manner that each class occupied the same portions of
the day that the others did. I sought in this way to avoid the
effects of any horary change in the humidity etc, of the atmo-
sphere or in the magnetic force. Of the thirty experiments

268

PRACTICAL STANDARDS

comprising each class, six were performed at about each of the
several hours — 11 A.M., 12 J P.M., 1^ P.M., 4 p.m., and 5^ P.M.

The calorimeter, protected as already described, was supported
on the edges of a light wooden frame. It was carefully guarded
against draughts by screens coated with tinfoil placed at a foot
distance. The stirrer consisted of a vertical copper rod, to which
vanes, on the plan of a screw-propeller, were soldered at four equi-
distant places. The rod extended 2 inches above the calorimeter,
and was there affixed to a light wooden shaft 2 feet long, attached
at the upper end to the last spindle of a train of clock-wheels.

Third Series of Radiation Experiments.

Date

Temperature

Rise of

Fall of

of air

of water

temperature

weight •

1867

in.

June 28, 10.38 a.m.

460-527

481-990

-1-48

))

28, 3.53 p.m.

513-687

506-770

0-75

28-2

))

29, 11.65 a.m.

493-088

473-930

1-82

29, 4.40 p.m.

526-185

508-480

1-88

28-5

July

1, 1.23

469-368

442-114

2-46

27-5

2, 10.58 A.M.

404-842

439-790

-2-82

2, 4.5 P.M.

401-779

450-930

-4-1

28-5

4, 11.46 a.m.

492-210

427-517

6-97

>»

4, 4.42 p.m.

541007

484-927

5-1

26-5

20, 1.0

416-237

409-044

1-03

28-75

22, 11.5 A.M.

474-393

439-140

3-32

22, 3.50 p.m.

486-267

480-106

O'S

28-76

23, 11.41 A.M.

451029

456-947

-0-1

28-4

23, 4.49 p.m.

475-319

486-113

-0-65

28-5

24, 12.54

441-677

460-780

-1-48

26-6

25, 10.40 A.M.

435-863

410-237

2-43

25, 3.27 p.m.

515-653

460-939

5-03

28-8

26, 11.29 a.m.

441-256

447-526

-0-2

28-6

26, 4.49 p.m.

435-776

472-503

— 30

27, 1.7

404-58

433-444

-2-28

29-8

Aug.

2, 10.55 A.M.

369-966

374-18

-0-15

29-75

2, 3.50 P.M.

407-34

406-42

0-17

27-8

3, 11.30 a.m.

435-813

401-187

3-24

28-6

3, 4.33 p.m.

476-691

446-393

2-9

6, 1.15

457-87

447-843

1-05

28-9 .

8, 10.46 A.M.

442-403

426-304

1-68

8, 4.17 P.M.

489-901

463-143

2-42

29-7

9, 11.51a.m.

466-428

453-149

1-27

26-6

9, 5.37 P.M.

490-308

484-753

0-66

27-9

10, 1.20

502-96

472-469

2-82

28-6

Average :

460*6808
1_

461-6356

1-018

28-488

FOR ELECTRICAL MEASUREMENTS

269

The weight was 35 lbs., which, falling about 2 feet per hour,
produced a continuous revolution of the stirrer at a rate of about
200 in the minute. The action of the stirrer left nothing to be
desired. It was started five minutes before an experiment
commenced, and kept going until the last observation of the ther-
mometer had been made.

Each experiment, as in the second series, lasted one hour,
during which were made eight observations of the thermometer-
immersed in the calorimeter, twenty of the temperature of the air,
and forty of the deflection of the galvanometer.

Determinations of Horizontal Magnetic Intensity.

Date

June

))

July

n
V

»»
»

n
»

Aug.

»>

>»
>»

1867

28, 1.30 p.m.

29, 10.50 A.M.
29, 3.50 P.M.

1, 12.25

1, 5.20

2, 1.40

4, 10.45 A.M.

4, 3.45 P.M.
20, 12 Noon.
20, 4.40 p.m.

22, 1.30

23, 10.45 A.M.

23, 3.45 P.M.

24, 11.51 A.M.

24, 5.0 P.M.

25, 1.10

26, 10.30 A.M.

26, 3.33 P.M.

27, 12 Noon.
27, 5.12 P.M.

2, L30

3, 10.25 A.M.
3, 3.33 p.m.
6, 12.12

6, 4.50

8, 1.11

9, 10.53 A.M.
9, 4.42 P.M.

10, 12.12
10, 450

Galvanometer
deflection,

37 21-42
26 43-06

25 12-56

38 23-56
38 59-25

38 49-94

26 24-55
26 10-55

39 18-9
41 11-35
41 21-4

32 51
31 56-15

39 52-95

40 24-9

41 27-95
34 40-45

33 25-5
43 19-55

42 48-53
41 15-35

34 13-9

33 40-3

35 9-8
37 8-1
37 44-55
31 23-65
30 43-4

36 25-4

34 49-5

Average

Weighing by .i7676Vir

ourrent-meter, i H= — - — ^ —

grs.

253-04
109-28
9604
272-35
284-95
280-9
106-25
104-99
289-875
332-825
335-13
169-616
168-608
301-591
315-092
338 391
206-658
188-675
386-0
372-658
332-733
198-99
191-983
214-117
248-258
259-867
160-708
152-75
235-433
209-C08

tan 9

3-68334

3-67114
3-67964
3-68144
3-68634
3-68034
3-66894
3-68474
3-67484
3-68504
3-67594
3-67194
3*68224
3-67364
3-68474
3-67964
3-67324
3-67864
3-68194
3-68414
3-67584
3-66464
3-67628
3-67156
3-67784
3-68110
3-67186
3-67590
3-67557
3-67864

3-67771

270 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

The correction to be applied to the air-registering ther-
mometer, as deduced from the radiation experiments of this
third series, is found from

217-452 - 10a? ^ 488-807 + 20a?
16-26 - 46-8

whence x, the quantity to be added to the observed temperature
of the air in the thermal experiments, = 2*81. The temperature
of the air was therefore virtually 6-073 higher than that of the
water. The results also show that the unit of eflfect on the
calorimeter was produced by a diflference of temperature of
11-645.

Hence 310266 - ^^^^ = 305051 ;

ll-64o

and adding 0-077 for the unimmersed part of the thermometer-
stem, the corrected thermal effect in the third series is found to
be 30-5821.

The average capacity of the calorimeter was equal to that of
93859 grs. of water, being made up of 91531 grs. distilled water,
22364 grs. of copper, 486 grs. of tin (the weight of the coating
next the calorimeter), 52 grs. silk net (half that employed), the
thermometer, coil, and corks.

The equivalent deduced from the third series is therefore

* X 3-6777 [ X -35478 x 32465480 x 3600

(6-2832

^l X ^^^

= 25217.

The equivalents above arrived at are : —

From Series 1. Average of 10, 25335 ♦.
From Series 2. Average of 15, 25366 *
From Series 3. Average of 30, 25217 ♦.

The extra precautions taken in the last series entitle the last
figure to be taken as the result of the inquiry. Reduced to
weighings in vacuo it becomes 25187.

* See footnote p. 261 for conyersion to c.o.b. measure.

SIXTH KEPORT— EXETER, 1869.

The Electrical Standard Committee have this year had com-
paratively few meetings, and the results of the experiments made
by the individual Members do not call for a Report of any length.
It is, however, thought desirable to print at once, as Appendices,
the important results obtained by Professor Clerk Maxwell, in
determining the ratio of the electromagnetic and electrostatic
series of units, and also a description of Sir Wm. Thomson's
experiments on the same subject.

Deacriptiim of Sir Wm. Thomsons Experiments made for the
Determination of v, the Number of Electrostatic Units in the
Electromagnetic Unit By W. F. KiNO»

The two principal pieces of apparatus used in these experi-
ments were the absolute electrometer and the electrodynamometer.
The former of these instruments was described at the last Meeting
of the Association, and a description of it is printed in the Report.
Plate 7 illustrates the arrangements described in what follows.

The electrodjmamometer consists of two large coils of fine
copper wire, and a smaller coil of still finer wire. The two large
coils are about 30 centims. diameter, and are placed vertical, in
planes parallel to one another; the distance between the large
coils is 15 centims. (equal to their radius). The smaller coil is
suspended between the large coils by a copper wire of such a
thickness as to give the coil a time of vibration such that it com-
pletes a period in about thirteen seconds. The upper end of the
suspending wire is attached to a milled head, and this head can
be turned round by the fingers. The lower end of the wire is
firmly fixed to the coil, and is in metallic connexion with one end
of it. To the other end of the coil is soldered a spiral of very fine
platinum wire, which hangs directly below the coil, and its lower
end is cemented to the dry woodwork of the instrument. To the
fixed end of the spiral coil a copper wire is attached, whose other
end is soldered to a binding-screw in an accessible position.

272 PRACTICAL STANDARDS

On one side of the small or movable coil is fixed a plane
mirror, and in front of the mirror, at a distance of about 450
centims., the scale is fixed on which the observations are read.
A paraffin-lamp wire, to give dark line in image of fiame, and lens
are used in the ordinary way for finding accurately the angle
through which the coil turns. It is never greater than 0*05. Its
true amount can be determined to within -^ per cent.

The connexions are not very intricate, and are traced thus : —
Starting from one pole of the battery (the battery used was sixty
sawdust Daniell's in series), the current goes in at one end of
large coil No. 1, and from the other end of No. 1 the current goes
to either end of the movable coil ; and the end of the movable
coil at which we suppose the current to be coming out is connected
with the end of No. 2 large coil, similar to the end of No. 1, to
which the battery was first attached, that is to say, the end which
will make the current go round in the same direction in both the
large coils. When the current leaves thle extreme end of No. 2,
it passes through a 10,000 B.A. resistance-box; the current is
completed by connecting the other end of the resistance-box with
the pole of the battery not already engaged.

The absolute electrometer is used in the ordinary way for
measuring differences of potential, and its electrodes are con-
nected, one to the end of the dynamometer coil No. 1, which is
joined to the battery, and the other electrode is fixed to the end
of the resistance-box, which is connected to the other pole of the
battery. Thus the greatest difference of potential in the arrange-
ment is measured by the absolute electrometer. An electrometer-
key is used to reverse these connexions in the course of the
experiments.

There is only one other part of the arrangement to be ex-
plained, and that is the method of observing the resistance of
the dynamometer coils while the experiments are going on. This
was done by means of the resistance-box in the circuit and an
electrometer. At one time the standard electrometer was used
for this purpose, but more lately the quadrant, rendered un-
sensitive, was employed. Both these instruments are described
in the last Report.

To take the resistance of the coils, the electrodes of the
electrometer were first placed on the extreme ends of the three
coils, and the difference of potential was ascertained. The

EUdnmitter Sr^*

LEUctrvTTUter Key

( V,

1 J

r :

FOR ELECTRICAL MEASUREMENTS 273

electrodes were then shifted to the ends of the resistance-box, and
the difference of potentials of its two ends was found. This gives
ac once the resistance of the coils.

There are two things which have to be done before the ex-
periments are commenced. One is the determination of the
moment of inertia of the movable coil. This is done at the
beginning and end of a long series of experiments, by comparing
it with a ring whose moment of inertia is known. The other is
done every day, and it is finding the time of vibration of the
small coil after all the connexions have been made and the coil
put into its place. This was done both with the current from the
battery flowing through the coils and with no current flowing;
but this variation was of very little consequence, as no difference
could be detected in the time. When the dynamometer is set up,
c€Lre is taken to neutralize the effects of the earth's magnetism
by a large number of magnets fixed at a great distance from the
coils. If the adjustment of the magnets is perfect, there is no
alteration of the position of the spot of light when the current is
reversed through the coils by the battery-key. Up to the present
time (May 1868) various causes have prevented the obtainment of
as satisfactory results as the method described above allows us to
expect. Eleven sets of experiments, made at various dates, from
March 10 to May 8 of the present year, have indicated values
for V, of which the greatest was 292 x lO^, the smallest 275*4 x 10«,
and the mean 282*5 x 10^ centimetres per second. Sir W. Thomson
intends to continue the investigation, hoping to attidn much
greater accuracy.

[P.S. Nov. 1869. A new form of absolute electrometer has
now been completed and brought into use, with good promise as
to accuracy and convenience. A glass jar constituting the
" Leyden battery '* contains within it the " absolute electrometer "
proper, the "idiostatic gauge," and the " replenisher." One
observer can use it effectively ; although it is more easily worked
by two, one maintaining constant potential in the Leyden jar by
aid of the idiostatic gauge and the replemsher, and the other
attending to the absolute electrometer (main balance and micro-
meter-screw). The main balance, giving electric weighing iii
known weights, is as steady and as easily used as any of the
"attracted disk" electrometers, whether portable or stationary,
described in previous Reports.]

B. A. 18

274 PRACTICAL STANDARDS

Experiments on the Value of v, the Ratio of the Electromagnetic to
the Electrostatic Unit of Electricity, By J. Clerk Maxwell.

The experiments consisted in observing the equilibrium be-
tween the electrostatic attraction of two disks, at a certain
difference of potential, and the electromagnetic repulsion of two
coils traversed by a certain current. For this purpose one of the
disks, with one of the coils at its back, was attached to one arm
of a torsion-balance, while the other, with the other coil at its
back, was capable of being moved to various distances from the
suspended disk by a micrometer-screw. Another coil, traversed
by the same current in the opposite direction, was attached to the
other arm of the torsion-balance, so as to do away with the effect
of terrestrial magnetism.

The fixed disk was larger than the suspended disk, and the
latter, when in its zero position, had its sur&ce in the same plane
as that of a " guard-ring," as in Sir W. Thomson's electrometers.
Its position and motion were observed by means of a microscope,
directed to a graduated glass scale, connected with the disk.
When the microscope was adjusted so that the image of the zero
line on the glass scale coincided with the cross wires of the micro-
scope, the very smallest motion of the scale could be easily
detected, so that the observations were very rapid. The disk was
brought to zero by the tangent-screw at the top of the suspension-
wire, and its equilibrium was also observed at zero. The equi-
librium, when the electrical forces were applied, was always
unstable. This electrical balance was made by Mr Becker. The
experiments were made in the laboratoryof Mr Oassiot, who kindly
gave the use of his great battery for the purpose. Mr Willoughby
Smith lent his resistance-coils, of 1,102,000 ohms; Messrs Forde
and Fleeming Jenkin lent a galvanometer, a resistance-box, a
bridge and a key ; and Mr C. Hockin undertook the observation
of the galvanometer, and the testing of the galvanometer, the
resistances, and the micrometer-screw.

The difference of potentials of the disks was compared with
the current in the coils as follows : — One electrode of the great
battery was connected with the fixed disk, and the other with the
case of the instrument and the guard-ring and the suspended
disk. They were also connected through the great resistance JB,

FOR ELECTRICAL MEASUREMENTS

276

and the primary coil of the galvanometer Oi, shunted with a
resistance S.

A small Grove's battery was employed to send a current
through the three coils and the secondary coil of the galvano-
meter &,.

£ar/A

A, Saspended disk and ooil.

i4'. Ootmterpoise disk and ooil,

C Fixed disk and coil.

By. Great bottezy. J?,. Small battezy.

-Gi. Primaxy coil of galvanometer.

G%, Secondary coil.

R. Great resistance. S, Shunt.

JT. Double key.

g. Graduated glass scale.

X, Current through R.

af. Current through G, .

s-zf. Current through 8.

y. Current through the three ooila

and Of, *
Jkf. Mercury cup.
T. Torsion head and tangent screw.

One-quarter of the micrometer-box, disks, and coils is cut away to show the
interior. The case of the instrument is not shown. The galvanometer and shunts
were 10 feet from the electric balance.

Equilibrium of the electric balance was obtained by working
the micrometer, and so adjusting the distance of the disks. At

18—2

276 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

the same time equilibrium of the galvanometer was obtained by
altering the resistance of the shunt S,

The simultaneous values of the micrometer and the shunt
formed the result of each experiment. It was necessary also to
ascertain the ratio of the magnetic effects of the two coils of the
galvanometer immediately after each set of experiments.

The experiments are described at greater length in the Philo-
sophical Transactions for 1868, p. 643.

The method of experimenting appeared capable of considerable
accuracy; but some difficulties arose from want of constancy in
the batteries, from leakage of electricity, etc., so that many of the
experiments were known to be faulty. Twelve experiments, how-
ever, against which nothing could be proved at the time of making
them, in which the distance of the disks ranged from ^ to ^ an
inch, and the power of the battery from 1000 to 2600 cells, gave
values of v of which the least was 28*4 and the greatest 29*4 ohms;
and in nine of these the values lay between 28*68 and 28*91. The
mean of the 12 was —

V = 28*798 ohms
= 288,000,000 metres per second
= 179,000 statute miles per second.

This result is much lower than that of MM. Weber and
Kohlrausch, which was v = 310,740,000 metres per second, but
agrees, I believe, more nearly with values recently obtained by
Sir W. Thomson, whose method, as well as mine, depends on the
B.A. unit Weber's method depends on the measure of capacity^
It is to be hoped that this important physical quantity may soon
be determined by methods founded on capacity, and disem-
barrassed from the phenomena of "electric absorption," which
occurs in all solid condensers, and which would tend to give too
high values of v.

REPORT* ON THE NEW UNIT OF ELECTRICAL RESIST-
ANCE PROPOSED AND ISSUED BY THE COMMITTEE
ON ELECTRICAL STANDARDS APPOINTED IN 1861
BY THE BRITISH ASSOCIATION.

[From Proc. Roy, Soc. xiv. 154 to 164 (April 6, 1865).]

By Fleeming Jenkin, Esq.

Sir Humphry Davy, in 1821 f, published his researches
proving a difference in the conducting power of metals and
the decrease of that power as their temperature rose. This
quality of metals was examined by Snow Harris, Gumming, and
E. Becquerel, whose table of conducting powers, compiled by the
aid of his differential galvanometer, and published in 1826 :[> ^
still frequently quoted, and*^ is indeed remarkable as the result of
experiments made before the publication by Ohm, in 1827 §, of
the true mathematical theory of the galvanic circuit.

The idea of resistance as the property of a conductor was
introduced by Ohm, who conceived the force of the battery over-
coming the resistance of the conductors and producing the current
as a result. Sir Humphry Davy, on the contrary, and other
writers of his time, conceived the voltaic battery rather as con-
tinually reproducing a charge, somewhat analogous to that of a
Leyden jar, which was discharged so soon as a conductor allowed
the fluid to pass. The idea of resistance is the necessary corollary
of the conception of a force doing some kind of work||, whereas
the idea of conducting power is the result of an obvious analogy
when electricity is conceived as a fluid, or two fluids, allowed to
pass in different quantities through different wires from pole to

* This did not form part of a Report to the British Association.

t PhiL Trans, 1821, vol. czi. p. 425.

t Ann. de Chim. et de Phy$, 2nd series, vol. xxxil p. 420.

§ Die galvanitehe Kette, mathematUch bearbeitett 1827 ; also Taylor's SeiefUiJU
Memoirs, vol. ii. p. 401.

II The writer does not mean by this that eleotrioal and mechanical resistance are
truly analogous, or that a current truly represents work.

278 PRACTICAL STANDARDS

pole. When submitted to measurement, the qualities of con-
ducting power and resistance are naturally expressed by reciprocal
numbers; and the terms are used in this sense in the early
writings of Lenz (1833)*, who, with Fechnerf and PouilletJ,
established the truth of Ohm's theory shortly after the year 1830.

The conception of a unit of resistance is implicitly contained
in the very expression of Ohm's law ; but the earlier writers seem
to have contented themselves with reducing by calculation the
resistance of all parts of a heterogeneous circuit into a given
length of some given part of that circuit, so as to form an ima-
ginary homogeneous conductor, the idea of which lies at the basis
of Ohm's reasoning. These writers, therefore, generally speak of
the resistance as the " reduced length " of the conductor, a term
still much used in France (vide Daguin, Jamin, Becquerel, De la
Rive, and others). The next step would naturally be, when com-
paring different circuits, to reduce all resistances into a length of
some one standard wire, though this wire might not form part of
all or of any of the circuits, and then to treat the unit length of
that standard wire as a unit of resistance. Accordingly we find
Lenz (in 1838§) stating that 1 foot of No. 11 copper wire is his
unit of resistance, and that it is 19*9 times as great as the unit
he used in 1833 1|, which was a certain constant part of the old
circuit. In the earlier paper the resistances are treated as lengths,
in the later as so many ''units."

Lenz appears to have chosen his unit at random, and ap-
parently without the wish to impose that unit upon others. A
further advance is seen when Professor Wheatstone, in his well-
known paper of 184311, proposes 1 foot of copper wire, weighing
100 grains, not only as a unit, but as a standard of resistance,
chosen with reference to the standard weight and length used in
this country. To Professor Wheatstone also appears due the
credit of constructing (in 1840) the first instruments by which
definite multiples of the resistance-unit chosen might be added
or subtracted at will from the circuit!. He was closely followed

* Pogg. Ann, vol. xxxiv. p. 418.
t Maatbeitimmungen, etc. 1 vol. 4to. Leipzic, 1881.

X EUmenB de Phyiique, p. 210, 6th edition; ftnd Comptes Rendu$t vol. it. p. 267.
§ Pogg, Ann. vol. xlv. p. 105.
)l Pogg* Ann. vol. xxxiv. p. 418.
IT Phil. Tram. 1843, vol. cxxxm. p. 808.

FOR BLECTRICAL MEASUREMENTS 279

by Poggendorff * and Jacobif , the description of whose apparatus,
indeed, precedes that of the Rheostat and Resistance-coils,
although the writer understands that they acknowledge having
cognizance of those inventions. Resistance-coils, as the means of
adding, not given lengths, but given graduated resistances to any
circuit, are now as necessary to the electrician as the balance to
the chemist.

In 1846 Hankelt used as unit of resistance a certain iron
wire ; in 1847 I. B. Cooke§ speaks of a length of wire of such
section and conducting-power as is best fitted for a standard of
resistance. Buff|| and Horsfordf in the same year reduce the
resistance of their experiments to lengths of a given German-
silver wire, and as a further definition they give its value as
compared with pure silver. To avoid the growing inconvenience
of this multiplicity of standards, Jaoobi** (in 1848) sent to
Poggendorflf and others a certain copper wire, since well known
as Jacobins standard, desiring that they would take copies of it,
so that all their results might be expressed in one measure. He
pointed out, with great justice, that mere definition of the
standard used, as a given length and weight of wire, was insuffi-
cient, and that good copies of a standard, even if chosen at
random, would be preferable to the reproduction in one laboratory
of a standard prepared and kept in another. The present Com-
mittee fiilly indorse this view, although the definition of standards
based on weights and dimensions of given materials has since then
gained greatly in precision.

Until about the year 1850 measurements of resistance were
confined, with few exceptions, to the laboratory ; but about that
time underground telegraphic wires were introduced, and were
shortly followed by submarine cables, in the examination and
manufacture of which the practical engineer soon found the
benefit of a knowledge of electrical laws. Thus in 1847 the
officers of the Electric and International Telegraph Company used
resistance-coils made by Mr W. F. Cooke, apparently multiples of
Wheatstone's original standard, which was nearly equal to the

* Pogg, Ann, vol. Ln. p. 511.
t Pogg. Ann. toI. lii. p. 526, vol. liv. p. 847.

t Pogg. Ann. vol. lxix. p. 255. § Phil. Mag. New Seriefl, vol. xxz. p. SS5.

I! Pogg. Ann. vol. lxxiii. p. 497.

H Pogg. Ann. vol. lxx. p. 238, and 8iUiman*9 Journ. vol. v. p. Sii.
** Contptet RendtUt 1851, voL xxxni. p. 277.

280 PRACnOAL STANDARDS

No, 16 wire of commerce; and Mr C. F. Varley* states that,
even at that date, he used a rough mode of '' distance testing/'
In 1850, Lieut. Werner Siemensf published two methods for
determining, by experiments made at distant stations, the position
of '' a fault " — that is to say, a connexion between the earth and
the conducting-wire of the line at some point between the
stations. In one of these plans a resistance equal to that of the
battery is used, and the addition of resistances is also suggested ;
and Sir Charles Bright, in a Patent dated 1852}, gives an account
of a plan for determining the position of a fault by the direct
use of resistance-coils. Since that time new methods of testing
for faults and of examining the quality of materials employed,
and the condition of the line, have been continually invented,
almost all turning, more or less, on the measurement of re-
sistance ; greater accuracy has been continually demanded in the
adjustment of coils and other testing-apparatus, until we have now
reached a point where we look back with surprise at the rough
and ready means by which the great discoveries were made on
which all our work is founded.

The first efiect of the commercial use of resistance was to turn
the "feet" of the laboratory into "miles" of telegraph wire.
Thus we find employed as units, in England the mile of No. 16
copper wire§, in Germany the German mile of No. 8 iron wire,
and in France the kilometre of iron wire of 4 millimetres
diameter. Several other units were from time to time proposed
by Langsdorf |{, Jacobif , Marie-Davy**, Weberf+, W. ThomsonJ},
and others, with a gradually increasing perception of the points
of chief importance in a standard ; but none of these were
generally accepted as the one recognized measure in any country.
To remedy the continually increasing evils arising from the
discrepancies invariably found between different sets of coils,
Dr Werner Siemens (in 1860§§) constructed standards, taking as

•

Letter to writer, 1865. f Pogg. Ann, vol. lxxix. p. 481.

X Patent No. 14,831, dated Oct. 21, 1852.

§ A size mnoh used in underground conductors, and equal in resistance to
about double the length of the common No. 8 iron wire employed in aerial lines.
II Liehig^i Ann. vol. lxzxv. p. 155. IF Pogg, Ann, vol lxxvui. p. 178.

** Ann. Chim, et Phys. 3rd series, vol. ix. p. 410.
ft Pogg. Ann. vol. lxxxii. p. 887.
Xt Phil. Mag. Deo. 1851, 4ih series, vol. n. p. 551.
S§ Pogg. Ann. vol. ex. p. 1.

FOR ELECTRICAL MEASUREMENTS 281

unit the resistance of a column of chemically pure mercury 1 metre
long, having a section equal to 1 millimetre square, and main-
tained at the temperature of 0° Centigrade*. Dr Siemens
supposed that this standard could be reproduced without much
difficulty where copies could not be directly obtained. Mercury
had been proposed before as a fitting material for a standard by
Marie-Davy and De la Rive; but Dr Siemens merits especial
recognition, as the coils and apparatus he issued have been made
with great care, and have materially helped in introducing strict
accuracy f.

The question had reached this point when (in 1861) the
British Association, at the suggestion of Professor W. Thomson,
appointed a Committee to determine the best standard of elec-
trical resistance. This Committee, aided by a grant from the
Royal Society, have now issued a new standard, the subject of the
present paper.

The writer has hitherto described those units only which are
founded on a more or less arbitrary size and weight of some more
or less suitable material; but measurements of resistance can
be conceived and carried out entirely without reference to the
special qualities of any material whatever. In 1849 Kirchhoff t
had already effected a measurement of this kind; but it is to
W. Weber§ that we owe the first distinct proposal (in 1851) of a
definite system of electrical measurements, according to which
resistance would be measured in terms of an absolute velocity.
This system of measures he called absolute electromagnetic
measure, in analogy with Gauss's nomenclature of absolute mag-
netic measure. The Committee have decided that Weber's
proposal is far preferable to the use of any unit of the kind pre-
viously described. Setting aside the difficulties in the way of
their reproduction, which are by no means contemptible, arbitrary

* Br Siemens, while retaining his definition, has altered the valae of his
standard about 2 per cent, since the first issue ; and it is doubtful whether even
the present standard represents the definition truly: his experiments were made
by weight ; and in reducing the results to simple measurements of length he has
used a specific gravity for mercury of 13*557 instead of 18*596 as given by Begnault,
18*595 by H. Kopp, and 13*594 by Balfour Stewart. (1873. The error due to this
cause has since been corrected.)

t Many of the different units described above were represented by resistance-
coils in the International Exhibition of 1862 : vide Jury Report, Glass XIII. p. 83,
where their relative values are given : vide also Appendix A to present paper.

X Pogg, Ann» vol. lxxvi. p. 412. § Pogg, Ann. vol. Lxxxn. p. 337.

282 PRACTICAL STANDARDS

material standards, whether of mercury, gold, silver, platinum, or
any other material, would be heterogeneous isolated units without
any natural connexion with any other physical units. The unit
proposed by Weber, on the other hand, forms part of a symme-
trical natural system, including both the fundamental units of
length, time, and mass, and the derived electrical units of current,
quantity and electromotive force. Moreover it has been shown by
Professor W. Thomson*, who accepted and extended Weber's
proposal immediately on its appearance, that the unit of absolute
work, the connecting link between all physical forces, forms part of
the same system, and may be used as the basis of the definition of
the absolute electromagnetic units.

The fiill grounds of the choice of the Committee could only
be explained by a needless repetition of the arguments given in
the reports already made to the British Association. It will be
sufficient here to state that, in the absolute electromagnetic
system, the following equations exist between the mechanical and

electrical units: —

W^C'Rt (1)

where W is the work done in the time t by the current G conveyed
through a conductor of the resistance R. This equation expresses
Joule and Thomson's law.

C = ^ (2)

where E is the electromotive force. This equation expresses
Ohm's law.

Q=ct. (3)

expressing a relation first proved by Faraday, where Q is the
quantity of electricity conveyed or neutralized by the current in
the time t Finally, the whole system is rendered determinate by
the condition that the unit length of the unit current must
produce the unit force on the unit pole (Gauss) at the unit
distance. If it is preferred to omit the conception of magnetism^
this last statement is exactly equivalent to sa}ring that the unit
current conducted round two circles of unit area in vertical planes
at right angles to each other, one circuit being at a great distance
D above the other, will cause a couple to act between the circuits
of a magnitude equal to the reciprocal of the cube of the distiince

* PhiL Mag, Deo. 1851, 4th series, vol. n. p. 551.

FOR ELECTRICAL MEASUREMENTS 283

D, This last relation expresses the proposal made by Weber for
connecting the electric and magnetic measure. These four re-
lations serve to define the four magnitudes R, C, Q, and E, without
reference to any but the fundamental units of time, space, and
mass; and when reduced to these fundamental units, it will be
found that the measurement of R involves simply a velocity, %,e.
the quotient of a length by a time. It is for this reason that the

absolute measure of resistance is styled j or j , pre-

•^ second second '^

cisely as the common non-absolute unit of work involving the

product of a weight into a length is styled kilogrammetre or

foot-pound. The Committee have chosen as fundamental units

the second of time, the metre, and the mass of the Paris gramme.

The metrical rather than the British system of units was selected,

in the hope that the new unit might so find better acceptance

abroad, and with the feeling that while there is a possibility that

we may accept foreign measures, there is no chance that the

Continent will adopt ours. The unit of force is taken as the

force capable of producing in one second a velocity of one metre

per second in the mass of a Paris gramme, and the unit of work

as that which would be done by the above force acting through

one metre of space. These points are very fully explained in the

British Association Report for 1863, and in the Appendix C to

that Report by Professor J. Clerk Maxwell and the writer.

The mamitude of the r is far too small to be practically

® second ^ ''

convenient, and the Committee have therefore, while adopting the

system, chosen as their standard a decimal multiple 10^^ times as

great as Weber's unit [ the -z — j , or 10' times as great as

the , . This macmitude is not very different from Siemens's

second ° "^

mercury unit, which has been found convenient in practice. It is
about the twenty-fifth part of the mile of No. 16 impure copper
wire used as a standard by the Electric and International Com-
pany, and about once and a half Jacobi's unit*.

It was found necessary to undertake entirely fresh experiments

* This last number may be 30 per cent, wrong, as the writer has never been in
possession of an authenticated Jacobi standard, and has only arrived at a rough
idea of its value by a series of published values which afford an indirect comparison.

284 PRACTICAL STANDARDS

in order to determine the actual value of the abstract standard,
and to express the same in a material standard which might form
the basis of sets of resistance-coils to be used in the usual
manner. These experiments, made during two years with two
distinct sets of apparatus by Professor J. C. Maxwell and the
writer, according to a plan devised by Professor W. Thomson, jvre
fully described in the Reports to the British Association for 1863
and 1864.

The results of the two series of experiments made in the two
years agree within 0*2 per cent., and they show that the new
standard does not probably differ from true absolute measure by
0*1 per cent.* It is not far from the mean of a somewhat widely
differing series of determinations by Weber.

In order to avoid the inconvenience of a fluctuating standard,
it is proposed that the new standard shall not be called " absolute

measure," or described as so many j- , but that it shall receive

•^ seconds

a distinctive name, such as the B.A. unit, or, as Mr Latimer Clark
suggests, the "Ohmad^f; so that, if hereafter improved methods
of determination in absolute measure are discovered or better
experiments made, the standard need not be changed, but a small
coefficient of correction applied in those cases in which it is
necessary to convert the B.A. measure into absolute measure.
Every unit in popular use has a distinctive name ; we say feet or
grains, not units of length or units of weight ; and it is in this
way only that ambiguity can be avoided. There are many absolute
measures, according as the foot and grain, the millimetre and
milligramme, the metre and gramme, etc. are used as the basis of
the system. Another chance of error arises from the possibility
of a mistake in the decimal multiple used as standard. For all
these reasons, as well as for convenience of expression, the writer
would be glad if Mr Clark's proposal were adopted and the unit
called an Ohmad.

Experiments have been made for the Committee by Dr
Matthiessen, to determine how far the permanency of material
standards may be relied on, and under what conditions wires un-
altered in dimension, in chemical composition, or in temperature
change their resistance. Dr Matthiessen has established that in

* Vide Appendix B.

t NoUf 1873. The name Ohm has been adopted.

FOB ELECTRICAL MEASUREMENTS 285

some metals a partial annealing, diminishing their resistance,
does take place, apparently due to age only. Other metals
exhibit no alteration of this kind ; and no permanent change due
to the passage of voltaic currents has been detected in any wires
of any metal — a conclusion contrary to a belief which has very
generally prevailed.

The standard obtained has been expressed in platinum, in a
gold-silver alloy, in a platinum-silver alloy, in a platinum-iridtum
alloy, and in mercury. Two equal standards have been prepared
in each metal ; so that should time or accident cause a change in
one or more, this change will be detected by reference to the
others. The experiments and considerations which have led to
the choice of the above materials are fully given in the Report to
the British Association for 1864. The standards of solid metals,
are wires of from 0*5 millim. to 0*8 millim. diameter, and varying
from one to two metres in length, insulated with white silk
wound round a long hollow bobbin, and then saturated with solid
paraffin. The long hollow form chosen allows the coils rapidly
to assume the tempeniture of any surrounding medium, and they
can be plunged, without injury, into a bath of water at the
temperature at which they correctly express the standard. The
mercury standards consist of two glass tubes about three-quarters
of a metre in length. All these standards are equal to one another
at some temperature stated on each coil, and lying between 14''*5
and 16'*'5C. None of them, when correct, diflfer more than 0"03
per cent, from their value at 16° '5 C.

Serious errors have occasionally been introduced into observa-
tions by resistance at connexions between different parts of a
voltaic circuit, as perfect metallic contact at these 'points is often
prevented by oxide or dirt of some kind Professor Thomson's
method of inserting resistances in the Wheatstone balance (dif-
ferential measurer) has been adopted for the standards; but in the
use of the copies which have been issued it has been thought that
sufficient accuracy would be attained by the use of amalgamated
mercury connexions.

In the standards themselves permanence is the one paramount
quality to be aimed at ; but in copies for practical use a material
which changes little in resistance with change of temperature is
very desirable, as otherwise much time is lost in waiting till coils
have cooled after the passage of a current ; moreover large correc*
tions have otherwise to be employed when the coils are used at

286 PRACTICAL STANDARDS

various temperatures ; and these temperatures are frequently not
known with perfect accuracy. German-silver, a suitable material
in this respect, and much used hitherto, has been found to alter
in resistance, in some cases, without any known cause but the
lapse of time, since the change has been observed where the wires
were carefully protected against mechanical or chemical injury.
A platinum-silver alloy has been preferred by the Committee to
German-silver for the copies which have been made of the
standard. These have been adjusted by Dr Matthiessen so as to
be correct at some temperature not differing more than 1° from
15°'6C. The resistance of platinum silver changes about 0*031
per cent, for each degree Centigrade within the limits of 5** above
and below this temperature ; this change is even less than that of
German-silver. The new material seems also likely to be very
permanent, as it is little affected by annealing. The form of the
copies is the same as that of the standard, with the exception of
the terminals, which are simple copper rods ending in an amal-
gamated surface. Twenty copies have been distributed gratis,
and notices issued that others can be procured from the Committee
for £2. lOs. The Committee also propose to verify, at a small
charge, any coils made by opticians, as is done for thermometers
and barometers at Eew.

Dr Matthiessen reports, with reference to the question of
reproduction, that given weights and dimensions of several pure
metals might be employed for this purpose if absolute care were
taken. The reproduction, in this manner, of the mercury unit, as
defined by Dr Siemens, differs frt)m the standards issued by him
in 1864 about 8*2 per thousand if the same specific gravity of
mercury be used for both observations*. Each observer uses for
his final value the mean of several extremely accordant results.
It is therefore to be hoped that the standard will never have to
be reproduced by this or any similar method. On the other hand,
four distinct observers, with four different apparatus, using four
different pairs of standards issued respectively by Dr Siemens
and the Committee, give the B.A. unit as respectively equal to
10456, 10455, 10466, and 10467 of Siemens's 1864 unitf. It
is» certain that two resistances can be compared with an accu-
racy of one part in one hundred thousand — an accuracy wholly

* If Dr Matthiesseu uees the sp. gr. of 13*596, as given by Regnault, the
difference from Dr Siemens's standard is 5 per thousand,
t 1873. The value now adopted is 1*0436.

FOR KLECTRICAL MEASUREMENTS 287

unattainable in any reproduction by weights and measures of a
^ven body, or by fresh reference to experiments on the absolute
resistance. The above four comparisons, two of which were made
by practical engineers, show how far the present practice and
requirements differ from those of twenty and even ten years ago,
when, although the change of resistance due to change of tem-
perature was known, it was not thought necessary to specify the
temperature at which the copper or silver standard used was correct.
The difficulty of reproducing a standard by simple reference to a
pure metal, further shows the unsatisfactory nature of that system
in which the conducting power of substances is measured by com-
parison with that of some other body, such as silver or mercury.
Dr Matthiessen has frequently pointed out the discrepancies thus
produced, although he has himself followed the same system
pending the final selection of a unit of resistance. It is hoped
that for the future this quality of materials will always be ex-
pressed as a specific resistance or specific conducting power referred
to the unit of mass or the unit of volume, and measured in terms
of the standard unit resistance, that the words conducting power
will invariably be used to signify the reciprocal of resistance, and
that the vague terms good and bad conductor or insulator will be
replaced, in all writings aiming at scientific accuracy, by those
exact measurements which can now be made with far greater ease
than equally accurate measurements of length.

There is every reason to believe that the new standard will
be gladly accepted throughout Great Britain and the colonies.
Indeed the only obstacle to its introduction arises from the diffi-
culty of explaining to inquirers what the unit is. The writer has
been so much perplexed by this simple question, finding himself
unable to answer it without entering at large on the subject of
electrical measurement, that he has been led to devise the follow-
ing definitions, in which none but already established measures
are referred to.

The resistance of the absolute j is such that the current

second

generated in a circuit of that resistance by the electromotive force

due to a straight bar 1 metre long moving across a magnetic field

of unit intensity*, perpendicularly to the lines of force and to its

own direction, with a velocity of 1 metre per second, would, if

* OauBB's definition.

288 PRACTICAL STANDARDS

doing no other work or equivalent of work, develop in that circuit
in one second of time a total amount of heat equivalent to one
absolute unit of work — or sufficient heat, according to Dr Joule's
experiments, to heat 00002405 gramme of water at its maximum
density 1** Centigrade.

The new standard issued is as close an approximation as could
be obtained by the Committee to a resistance ten million times

as great as the absolute -^. The straight bar moving as

described above in a magnetic field of unit intensity would require
to move with a velocity of ten millions of metres per second to
produce an electromotive force which would generate in a circuit
of the resistance of the new standard the same current as would

be produced in the circuit of one j resistance by the electro-

^ second "^

motive force due to the motion of the bar at a velocity of one

metre per second. The velocity required to produce this particular

current* being in each case proportional to the resistance of the

circuit, may be used to measure that resistance ; and the resistance

of the B. A. unit may therefore be said to be ten millions of metres

per second, or 10' r .

'^ second

It is feared that these statements are still too complex to fulfil
the purpose of popular definitions ; but they may serve at least to
show how a real velocity may be used to measure a resistance by
using the velocity with which, under certain circumstances, part
of a circuit must be made to move in order to induce a given
current in a circuit of the resistance to be measured. That
current in the absolute system is the unit current, and the work
done by that unit current in the unit of time is equal to the
resistance of the circuit, as results from the first equation stated
above.

Those who from this slight sketch may desire to know more
of the subject will find fiill information in the Reports of the
Committee to the British Association in 1862, 1863, and 1864.
The Committee continue to act with the view of establishing and

* This carrent is the unit current, and, if doing no other work or equivalent of
work, would develop, in a circait of the resistance of the B.A. anit, heat equivalent
to ten millions of units of work, or enough to raise the temperature of 2405 grammes
of water at its maximum density V Centigrade.

[To face page 288.]

Thomson's
old anit

0-»520

1000

1-988

2-871

2-979

3-123

28-94

30-50

32-56

42-43

79-96

179-4

■

OermaD
Miles

0005307

0-005574

0-01108

0-01655

0-01661

001741

01613

0-l7(X)

0-1815

0-2365

0-4457

1*000

Observations

Calculated from the B.A. unit

From an old determination by Weber.

No measurement made; ratio be-
tween Siemens (Berlin) aod Jacobi,
taken from Weber's Oalvano-
metrie.

Measurement taken from a deter-
mination in 1862 of a standai-d
sent by Prof. Thomson; does not
agree with Weber's own measure-
ment of Siemens's units; by We-
ber 1 Siemens's unit « 1*025x107
metres-second.

Measurement taken from three coils
issued by Messrs Siemens.

Equal to 10,000,000 5??!£^, accord-
ing to experiments of Standards
Committee.

I From coils exhibited in 186Ji (pretty
) well adjusted).

I From coils exhibited in 1862 (in-
\ differently adjusted).

I From coils exhibited in 1862 (badly
( adjusted).

) From a coil lent by Dr Matthiessen
I (of (German -silver wire).

{ From coils lent by Mr Varley (well
I adjusted).

I From coils exhibited in 1862 by
/ Messrs Siemens, Halske & Co.t

Hiemens's unit Iftfie gravity alinded to in the text,
^is unit, whi«b ^

FOR ELECTRICAL MEASUREMENTS

289

isBumg the correlative units of current, electromotive force,
quantity, and capacity, the standard apparatus for which will, it
is proposed, be deposited at Eew along with the ten standards
of resistance already constructed with the funds voted by the
Royal Society.

Appendix B.

The following Table shows the degree of concordance obtained
in the separate experiments used to determine the unit. The

1. : 3.

Value of B. A. unit in
Time of 100 , . . ,^ metres

revolutiomi of coil, | *«™* ®^ ^^ ^^35Hd»
in seconds as calculated from

each experiment

17-64

17-58

77-62

7617

63-97

64*53

41-76

41-79

64-07

63-78

17697

17-783

17-81

17-78

17-01

16-89

21-36

21-38

21-362

21-643

11-247

16-737

1-01211
0-9836 1
1-04681
0-9613
0-9985
0*9998
0-9916
0-9936 f

0-9961 1
0-98861
0-98781
1-01361
0*9952 {
1-01741
1-0191 1
0-9895 f
1-00341
1-00111
0-99681
1-0096 1
1-04241
0-97071

Value from mean

of each pair of

experiments

Percentage error of

pair of observations

from mean value

0-9978

-0-22

1*0040

+0-28

0-9992

-0-08

0-9926

-0-75

0-9924

-0-76

1-0007

+0-07

1-0063

+0*63

1-0043

+0-43

1-0022

+0-22

1-0040

+0-40

0-9981

-019

Probable error of -B (1864) =0'1 percent.

Probable error of iJ (1863) =024

Difference in two values 1864 and 1863 = 0*16
Probable error of two experiments = 0*08 „

determinations were made by observing the deflections of a certain
magnet when a coil revolved at a given speed, first in one direc-
tion, and then in the opposite direction. The first column shows

a A. 19

290 PRACnCAL STANDARDS FOR ELECTRICAL MEASUREMENTS

the speed in each experiment ; the second shows the value of the

B.A. unit in terms of 10' r , as calculated from the sinele

second ^

experiments. A difference constantly in one direction may be

observed in the values obtained when the coil revolved different

ways. This difference depended on a slight bias of the suspending

thread in one direction. The third column shows the value of the

B.A. unit calculated from the pair of experiments; the fourth

shows the error of the pair from the mean value finally adopted.

In the final mean adopted, the 1864 determination was allowed

five times the weight allowed to that of 1863.

SEVENTH REPORT— LIVERPOOL, 1870.

The Committee are unable to report any material progress
during the last year in the work which remains to be done, and
beg leave to suggest that this work may probably be more
effectually expedited by the appointment of several small Com-
mittees than by retaining the large but somewhat cumbrous
organization by which their work was commenced. When the
Committee were first appointed, no coherent system of units for
the measurement of electrical resistance, currents, quantity,
capacity, or electromotive force had met with general acceptance.
The so-called absolute system existed indeed on paper, but in far
too intangible a form to be either understood or used by practical
men. At the same time, proposals for the adoption of isolated
units, variously determined, had been carried out, with more or
less success, so as to meet in some degree the immediate require-
ments of telegraphy. Many competing units of this nature were
in the field. The Committee chose a system based on the abso-
lute measure, and so, at least as far as electrical resistance was
concerned, made this measurement a tangible and practical opera-
tion ; and their choice has been ratified by men of science over
a great portion of the globe. Copies of the unit of resistance
adopted by the Committee in 1864 were deposited at the Eew
Observatory; and others exist in the hands of electricians in
various parts of the world. Comparisons of several of the copies,
which were published in the Report of the Committee for 1867,
showed that, with one or two exceptions, the ratio of their resist-
ances remained unchanged. It is, however, desirable that addi*
tional comparisons should be made fix)m time to time. Incidentally
many researches of considerable value were carried out by the
members of the Committee ; and the yearly reports have been so
generally in request that it may be advisable to reprint the entire
series.

19—2

292 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

No second unit, however, has been issued by the Committee,
although apparatus for the determination of the units of capacity,
quantity, potential, and intensity of current have been con-
structed, both with the funds of the Association and from the
private means of its members. The great numbers of the Com-
mittee render meetings of rare occurrence; and the Sub-com-
mittees appointed to undertake the work have been lately remiss
in its execution ; the Committee, believing that direct responsi-
bility to the Association and greater freedom of action will act
as a stimulus to individual members, beg to suggest that the
Electrical Standards Committee be not reappointed, but that
three new Committees of smaller numbers be chosen, to determine
and issue: 1st, a condenser representing the unit of capacity;
2nd, a gauge for showing the unit difference of potential ; 3rd, an
^lectrodynamometer adapted to measure the intensity of currents
in a decimal multiple of the absolute measure.

They would also suggest that it be an instruction to each
Committee that it shall carry out the system adopted by the
Electrical Standards Committee, and that these new Committees
shaiU have the use of all instruments hitherto constructed with
the funds of the Association, a list of which is appended (in
account book).

Considering that the principal instruments have already been
constructed, the Committee believe that a small grant of, say, £20
to each Committee, will be sufficient to meet the expenses of the
next year.

In conclusion, should this suggestion be adopted, they beg to
recommend that a volume, containing the complete series of
reports, be issued by the Association, and sold to the public^
feeling assured, from the demand for isolated copies, that such an
issue would involve no expense to the Association.

NoTB. The Buggestion that the Electrioal Standards Committee be not re-
appointed wae approved in 1870 by the General Committee of the British Association.
In 1880 seYeral members of the Association asked for the reappointment of thft
standards Committee and this was done.

EIGHTH REPORT— YORK, 1881.

It appeared to the Committee that in order to perform the
task entrusted to them, they had two principal questions to
consider: First, to select or prepare a well-defined standard of
accurately known absolute value for each kind of magnitude ; and,
secondly, to take measures for making certified copies of each of
the adopted standards accessible to the public.

The standard magnitudes which the Committee have had
under consideration as yet are —

1. The Standard of Resistance.

2. The Standard of Capacity.

3. The Standard of Electromotive Force.

As to the first of these, the standard of Resistance, the Com-
mittee were of opinion that, in view of the discrepant results
obtained by experimenters who have re-examined the absolute
resistance of the B.A. unit, it might be well to reconsider the
question whether the "ohm" should be defined by reference to
a particular coil of wire preserved as a concrete standard, or
whether the term *' ohm " should be understood to mean a resist-
ance of 10® C. G. s. units. They were also of opinion that it was
desirable to continue the experimental investigation of the abso-
lute resistance of the existing standards.

The repetition of the determination with the original apparatus,
by Lord Rayleigh and Professor Schuster in the Cavendish Labora-
tory, has gone far to supply this requirement. Experiments by
another method have also been carried on by Professor G. C.
Foster in the Physical Laboratory of University College, London.
Some account of these experiments is given in Appendix I. to
this Report, but the results hitherto obtained can only be regarded
as preliminary. With regard to the issue of authorised copies of
the ohm for general use, the Committee did not see their way to
making arrangements for actual construction of standard coils.

294 PRACTICAL STANDARDS

They were of opinion that it would be best to limit their action
to drawing up a detailed specification for the construction of
standard resistance coils, and to arranging for the systematic
testing of coils which are certified to them as being made in
accordance with this specification, issuing certificates showing
their actual resistance. Such a system would be analogous to the
system adopted by the Kew Committee for the testing of metero-
logical instruments at the Kew Observatoiy. It has not yet been
settled by whom this duty should be undertaken.

An important point of detail connected with the practical
construction and use of standard coils has been investigated by
Mr Herbert Taylor. The material adopted by the former Com-
mittee for the wire of the standards issued by them, was an alloy
of platinum and silver, containing one part platinum to two of
silver ; and the same material is very often used for the coils in
tlie " resistance-boxes " issued by instrument-makers. One special
reason for the selection of this alloy for the purpose named is its
small temperature-rate of variation of resistance, — 0031 per cent,
per degree, according to the late Dr Matthiessen. Mr Taylor has
now found that the rate of variation of resistance of wire made of
this alloy depends upon the diameter of the wire, the percentage
amount for one degree varying fi"om 00299 for a wire nearly
7 millim. in diameter to 0"0231 for a wire of diameter 0*168 millim.
A detailed account of Mr Taylor s experiments forms Appendix II.
to this Report.

With regard to standards of Capacity y the Committee are
able, thanks to the zealous co-operation of one of their number,
to report somewhat more complete arrangements, Dr Muirhead
having undertaken for the present to make and issue Standard
Condensers adjusted in accordance with one whose absolute
capacity has been determined by himself and Mr Hockin. {Brit
Assoc. Rep. 1879, pp. 283 and 285.)

With a view to testing the permanency of condensers made
with mica, paraffined paper, or other solid insulators, Dr Muirhead
is also having constructed a large air-condenser.

In reference to the standard of Electromotive Force, the Com*
mittee have had to consider whether this ought to be based upon
a particular combination of chemicals, forming a galvanic cell of
definite electromotive force, such, for instance, as a DanielFs cell,
constructed in a specified manner from materials of guaranteed

FOR ELECTRICAL MEASUREMENTS 295

purity, or the cell introduced by Mr Latimer Clark (Proc. Roy.
Soc. XX. 444), or whether they should not rather aim at the con-
struction of some convenient form of electrometer cap&ble of
indicating with sufficient accuracy an electromotive force of about
a volt. The first plan would be comparable with a supply of ice
and boiling water as affording a standard interval of temperature ;
the second would be comparable with a thermometer showing the
two limits of the standard interval.

The Committee are not yet prepared to make a final recom-
mendation as to either method of embodying the standard electro-
motive force, though they are strongly inclined to believe that
an electrometer or gauge, capable of showing when a definite
electromotive force has been developed, by whatever means, will
ultimately be found more satisfactory than any system in which
the constancy of an electromotive force is inferred irom the
supposed constancy of the conditions under which it has been
developed.

Another question of a more general kind, which, though it
may not be of much immediate practical importance, will eventually
have to be carefully considered, has occupied the attention of the
Committee to some extent. It is the question as to what concrete
standards should finally be recognised as fundamental standards.
Supposing that we already had independent standards of Resist-
ance, Capacity, Electromotive Force, Quantity, and Current Strength,
each of them defined with all the accuracy that our present
experimental methods admit of, they would infallibly be found to
exhibit small discrepancies when compared together. For in-
stance, the current of standard strength would not be exactly the
same as that produced by the standard electromotive force acting
in the circuit of the standard resistance, and we should then have
to consider which of the three standards was to be corrected so as
to bring it into harmony with the other two.

Similarly, in the case of the other electrical magnitudes. The
known relations between these are sufficient to enable us to define
the unit of any one of the five magnitudes just mentioned in
terms of the units of any two of the rest. Hence it appears that
the electrical standards of ultimate authority cannot be more than
two in number, and it will have to be decided which pair of con-
crete standards are to be recognised as ultimate or fundamental,
and what are to be left to be defined by reference to them.

296 PRACTICAL STANDARDS

A further question arising out of the mutual relations of the
fundamental units was that of the magnitude of the practical units
to which distinctive names should be attached. The present
usage with respect to this matter is that a resistance of 10^ C. o. s.
units is called an Ohm ; an electromotive force of 10^ c. g. s. units
is called a Volt; and the current produced by a Volt acting
through an Ohm, that is to say, a current of 10® 4- 10* or 0*1 c. G. s.
unit is called a Weber, In the opinion of the Committee it was
considered highly desirable, irom a scientilBc point of view, that
the relations among these standards should be simplified by de-
fining them as follows: —

Ohm = 10® c. G. s. units of resistance.

Volt = 10* c. G. s. units of electromotive force.

Weber = 1 c. G. s. unit of current.

It was felt, however, that any recommendation involving a
change in the value attached to terms which are rapidly coming
into extensive use among practical electricians, might give rise to
serious inconvenience. Therefore, although with regard to the
scientific aspect of this proposal the Committee were decidedly in
fovour of the change, they felt that a public recommendation could
not well be made until the practical inconveniences likely to foUoV
had been very carefully investigated.

Appendix I.

Account pf Preliminary Experiments on the Determination of
Electrical Resistances in Absolute Measure, By Professor
G. Carey Foster, F.R.S.

The experiments to be described in what follows were made
in the Physical Laboratory of University College, London. The
principle of the method employed is essentially the same as that
of the method long ago pointed out by Sir William Thomson, and
adopted by the first Committee of this Association upon Electrical
Standards in their experiments of 1863 and 1864, as well as by
Lord Rayleigh in the repetition of these experiments recently
carried out by him in conjunction with Professor Schuster.

FOB ELECTRICAL MEASUREMENTS

297

Every absolute measurement of resistance is, by the nature of
the case, fundamentally the determination of the ratio of an
electromotive force to the strength of the current which it pro-
duces in the conductor whose resistance is to be measured. In
Sir William Thomson's methods, as is well known, the electro-
motive force is due to the action of the earth's magnetism upon a
coil of wire spinning about a vertical diameter, and its numerical
value is known from the rate at which the coil spins, the total
area enclosed by its windings, and the intensity of the horizontal
component of the earth's magnetism. The electromotive force
generates a current in the coil, the strength of which is known
from the deflection of a small magnet, hung at the centre of the
coil, and from the intensity of the earth's horizontal magnetic
force. This last factor, entering similarly into the expressions for
the electromotive force and for the current, disappears from their
ratio, which gives in absolute measure the resistance of the wire
forming the revolving coil.

Fig. 1.

In the method now to be described there is again an electro-
motive force generated in a revolving coil, just as in Sir William
Thomson's method, but the current is measured by an independent
galvanometer, and the conductor, whose resistance is given by the
experiments, is entirely distinct from the revolving coil. So far
as this method possesses any particular advantages they arise from
the circumstance last mentioned. The conductor of which the
resistance is measured being at rest, it may be a coil of wire of

298 PRACTICAL STANDARDS

any material, wound in whatever way may be most convenient ;
and it may be immersed in a bath of liquid so as to keep it at an
accurately known temperature. Moreover, several independent
coils of different resistances can be experimented upon one after
another, and the resistance of each determined.

The nature of the method and the arrangement of the essential
parts of the apparatus may be explained by help of the adjoining
figure. In this, R stands for the wirCy of which the resistance is to
be measured ; and P for a thermopile whereby a current is produced
through 12 and through a tangent-galvanometer G, The ends of the
wire of the revolving coil (7 are connected, through a commutator K
upon the axle, with the ends of the wire R, a delicate reflecting
galvanometer g, called in what follows the zero-galvanometer^ being
interposed on one side of the commutator. When the speed of the
coil is so adjusted that the zero-galvanometer is not deflected, the
electromotive force developed in the coil by the magnetism of the
earth is equal to the electromotive force exerted by the thermopile
between the ends of the conductor R, Consequently, the resist-
ance of this conductor is obtained in absolute measure by dividing
the electromotive force of the coil by the strength of the current
indicated by the tangent-galvanometer.

This result may be expressed in terms of the experimental
data, as follows. Let A be the total area included by all the con-
volutions of the revolving coil — that is, the sum of the areas
included by all the turns taken severally, Uq the horizontal
magnetic intensity at the place occupied by the coil, to the angular
velocity of the coil, and 2a the arc of contact made by the commu-
tator, then E, the electromotive force of the coil, is

„ ry . sin a
^ a

Again, if F is the strength of the magnetic field produced at the
centre of the tangent-galvanometer, where the needle is hung, by
a current of unit strength flowing through the galvanometer,
Hq the horizontal intensity of the earth's magnetic field at the
same point, and 6 the deflection of the galvanometer, the strength
G of the current in the galvanometer, and therefore also in the
wire Ry is

(7= Stan 5.

FOR ELECTRICAL MEASUREMENTS 299

Hence, putting p for the resistance to be measured, that of the

wire iJ, we have

u- . psina

^"C HatsLaO '

_ He a

^^ ^"if^' ytan^ '

if T is the period of one revolution of the coil.

If the experiment could be made in a region of uniform
magnetic force, we should have Hq^ Hq, and therefore the ratio

-^ = 1, as in Sir William Thomson's method. Owing to the

neighbourhood of rather large masses of iron, this condition was
not fulfilled in the actual experiments. The ratio in question was
accordingly determined by noting the time of vibration of the
same magnet when it was suspended alternately in the position of
the revolving coil and in that of the galvanometer respectively.

It was thus found at the beginning of the experiments that j~

was equal to 0'9889. A repetition of the measurement that was
afterwards attempted was made useless by some large masses
of iron being brought just outside the Laboratory while it was
going on.

The ring upon which the revolving coil was wound, as well as
the frame in which it was mounted, were in the main copied from
those used in the experiments of the former Committee, but both
ring and frame were made considerably stouter in the metal, and
the ring had only one groove instead of two. The upper and lower
halves were insulated from each other, to prevent the formation of
induced currents.

To determine the area Ay we have A = nm^j^, where n is the
number of turns of wire on the coil and m the circumference of
the mean layer. The value of m was ascertained by measuring
with a steel tape the circumference of the groove in which the
coil was wound, as well as the circumference after each successive
layer of wire had been put on. The mean of all these measures,
corrected for the thickness of the tape, about 0*01 cm., was taken
as the value of m. In order to guard against accidental error, the

300 PRACTICAL STANDARDS

separate measurements of the circumference were combined in
pairs, thus Wo+win, m, + mn_i, ... the suffixes denoting the
numbers of layers of wire which had been wrapped on when the
several measurements were made ; these sums, which ought to be
constant, varied between 19325 cm. and 193*50 cm., the number
of layers being 32. In this way the circumference of the mean
layer was found to be 96*669 cm., which gives for the area enclosed
by it 743*66 sq. cm. Each layer of wire contained 30 turns, and
therefore n = 30 x 32 = 960, and the total effective area of the coil
was 4 = 960 X 743*66 = 713914 sq. cm.

The tangent-galvanometer had two equal parallel coils, of
approximately square section, placed at a distance apart nearly
equal to their mean radius, which was about 18*25 cm. Each coil
consisted of 22 layers of 20 turns each ; the galvanometer had thus
altogether 880 turns of wire. The needle consisted of three short
bits of hardened and magnetised watch-spring, &8tened one above
another at the back of a light plane-glass mirror. The deflections
were read upon a straight glass scale, divided at the back into
millimetres. The distance from mirror to scale was 137*25 cm., of
which about 0*45 cm. was occupied by glass ; the optical distance
was therefore taken as 136*95 cm. The galvanometer-constant F
was calculated by the formula

where n is the number of turns of wire (880) in the two rings
taken together, 8 and 8 the areas of the cross-sections of the two
coils, ai the external radius of each coil, a, the internal radius,
hi the half-distance measured parallel to the axis between the
outer surfaces of the coils, and 6, the half-distance between their
iuner surfaces. The numerical values were Oj = 18*945 for one
coil, =18-953 for the other; 02 = 17*518 and 17*524; fci = 9*851,
and 6s = 8*429, all in centimetrea The values for ai and also for
02 being so nearly alike for the two coils, the means Oi = 18*949
and o,= 17*521 were used in the calculation of F. The numerical
value of F was thus found to be 1/0*004618, so that the absolute
strength of a current measured upon this galvanometer is

0*0046185" tan ^.
The commutator of the revolving coil consisted of a cylindrical

FOR ELECTRICAL MEASUREMENTS 301

piece of ivory about 7*6 cm. in diameter, with two pieces of
platinum let in upon opposite sides. One end of the wire was
fastened to one of these platinum pieces and the other end to the
other piece ; and contact with the external circuit was made
through two platinum-faced gim-metal wheels, each about 16 cm.
diameter, which revolved in contact with the ivory cylinder. The
wheels revolved in insulated bearings about vertical axes, nearly in
the same plane as the axis of rotation of the coil. The upper end
of the axle of each wheel carried a small copper mercuiy cup into
which a well-amalgamated copper wire dipped for connecting the
coil with the end of the wire R (see figure), of which the resistance
was to be measured. This arrangement was adopted in order to
avoid the heating, and consequent thermo-electric action, which
would probably have resulted firom the use of rubbing contacts.
It was found very efficient for this purpose.

In order to avoid as far as possible the effects of self-induction
in the revolving coil, the platinum contact-pieces had an angular
breadth of only about 20 degrees, so that the coil was in metallic
connexion with the rest of the circuit during only about ^th of
each revolution. By adjusting the contact- wheels so that the
vertical plane containing their axes coincided with the magnetic
meridian, the middle of the period of contact was made to coincide
with the instant of maximum intensity and minimum rate of
variation of the electromotive force in the coil. The arc of con-
tact actually employed was 20° 3', which gives for the ratio of
the maximum and minimum electromotive force due to the earth's
magnetism the value 1:0*9817, or an extreme variation of less
than 2 per cent.

Putting together the values of the constant factors in the
expression for the resistance to be determined, we get

95561 X ly
^^ TtsLnd '

leaving T, the. period of rotation of the coil, and d, the deilectio9
of the tangent-galvanometer, to be observed in each experiment.

To determine the speed of the coil, the following method was
adopted. Three glass pens, each controlled by a small electro-
magnet, were caused to mark side by side upon a strip of pap^r
drawn forward by clock-work, as in an ordinary Morse receiver.

302 PRACTICAL STANDARDS

The pens, when left to themselves, ruled parallel straight lines on
the paper, but when any of the electro-magnets was excited, the
corresponding pen was pulled to one side and a notch was made
in the line the pen was drawing. By means of a wheel of 100
teeth, carrying a pin which made contact with a light spring once
in every revolution, and gearing into a screw cut upon the upper
part of the axle of the coil, the circuit of one of the electro-
magnets was completed for an instant at every hundredth revolu-
tion of the coil, and an indentation was made in the corresponding
line. The circuit of the second electro-magnet was broken for an
instant by a clock at intervals of one second, thus making notches
on the second line. By afterwards measuring the distances be-
tween the notches on the two lines, the speed with which the coil
was spinning at any instant could be ascertained. This measure-
ment was made by laying over the paper a strip of glass divided
on its lower surface into centimetres and millimetres. The degree
of accuracy attainable in this way, independently of error of the
clock, was about one part in one thousand. The speeds used in
the experiments varied from about nine to about twelve revolu-
tions per second. The electro- magnet acting on the third pen was
under the control of an observer who watched the zero-galvano-
meter {g in the iBgure), and held down a contact-key, which com-
pleted the corresponding circuit whenever and as long as this
galvanometer showed no deflection. In this way the third line on
the recording strip was displaced to one side whenever the speed
was such as to cause the electromotive force of the coil and that
due to the thermo-electric pile accurately to balance each other,
and thus the parts which were to be measured of the other two
lines were indicated. A second observer noted the reading of the
tangent-galvanometer when the zero-galvanometer was undeflected,
and thus determined the angle 0.

It is evident, from the formula given above, that the product
T tan 6 should be constant in all experiments in which the wire
whose resistance was to be determined was the same. The amount
of agreement in the value of this product in different experiments
therefore affords a criterion of the consistency of the results with
each other. The results obtained in two series of experiments
were as follows: —

FOR ELECTRICAL MEASUREMENTS 303

TiBuB T tan ^

(Series I.) (Series IL)

0-01291 0-01192

-01296 -01196

•01309 -01194

-01312 -01196

-01309 -01189

•01283 -01192

•01298 -01193

-01306 -01194

-01296 —

-01306 —

-01302 —

-01294 —

•01310 —

It will be seen that the second set of values agree better
together than the iSrst set. This is probably chiefly due to greater
practice in observing, and to the adoption of an artifice whereby
the speed of the gas engine employed to drive the coil was kept
more constant. In calculating the final result from each set,
weight was given to each separate observation in proportion to
the square of the number of revolutions of the coil over which it
extended ; for it was assumed that the accuracy with which the
speed of the coil was determined was proportional to the number
of revolutions included in the record ; moreover, the number of
galvanometer-readings obtained in each experiment was propor-
tional also to the number of revolutions, and hence it was assumed
that the accuracy with which the product was determined was
proportional to the square of the number of revolutions. The
weighted means thus calculated are, for the

First series . . 0013017,
Second seiieB . . 0011932.

Calling pi and pt the resistances measured in the two series of
experiments respectively, these results give

*"*• P* " 011932 '^ ^^ " ®^'^^ "" ^^-

The wires measured belonged to a set of Qerman-silver resist-
ance-coils, which were very carefully adjusted by my assistant,
Mr W. Grant, by comparison with a " B.A.-unit " coil issued by
the former Committee. The nominal values were 73 and 80 ohms
in the two sets of experiments respectively. Applying a not very

304 PBACTICAL STANDARDS

certain correction for the diflference between the temperature of
the coils during these experiments and that at which they were
adjusted, we get for the nominal values

Pi « 73*16 ohms and p, = 8018 ohms.

Hence, according to the first series of experiments,

1 ohm= Jj^ X 10» = 1-003 x 10»;

according to second series

QA.AQQ

1 ohm = ^°- xlO» = 0-999 x 10».
80*18

I do not attach any particular importance to these values, or
to the close agreement of their mean with the intended value of
the ohm, as the experiments, so £Etr, have only been undertaken
with the view of ascertaining how far the method that has been
described is capable, when employed under favourable circum-
stances, of giving good results. In this repect I think the experi-
ments may be considered fairly satirfactory, but the numbers
obtained for the value of the ohm are subject to several correc-
tions, the most important of which are probably that for errors of
the clock, which I had no means of rating more accurately than
by comparison with a good watch ; that due to slight uncertainty
as to the value in ohms of the resistances measured ; that due to
self-induction in the revolving coil, which, however, I believe must
be very small ; and perhaps errors due to unobserved disturbances
of the magnetic iBeld during the experiments.

I wish, in conclusion, to acknowledge with warm thanks the
obligations I am under to Mr Charles Hockin for most valuable
aid of various kinds, — important practical suggestions as to the
construction of the apparatus, information as to the conditions
required in order to ensure sensitiveness, and the loan, for a long
time, of a very delicate zero-galvanometer and a set of resistance
coils. I am also greatly indebted to Mr Qrant and Mr G. W. von
Tunzelmann, B.Sc, by whom conjointly the actual observations
were almost entirely made.

FOR ELECTRICAL MEASUREMENTS 305

Appendix IL

On the Causes of the Variation in the Temperature-Coefficient of
the Alloys of Platinum and Silver. By Herbert Taylor, Esq.

In his report to the Cominittee of the British Association in
1862 Dr Matthiessen proposed for the constraction of standard
resistance coils the now widely-used alloy of silver and platinum,
consisting of two parts of silver and one part of platinum by
weight. In the same paper he gave the specific conducting power
of the material and also its percentage-variation in resistance due
to a change of temperature of 1° Centigrade.

The latter value, called the temperature-coefficient in what
follows, he stated to be 0*031 per cent, per degree.

It was, however, found, after the alloy came into general use,
that the temperature-coefficient varied within moderately wide
limits.

And it was noticed, by the writer amongst others, that having
determined by experiment the coefficient of a particular wire, it
was necessary to make a fresh determination for the same wire
when drawn down to a finer gauge.

An investigation into the causes of these variations was there-
fore desirable ; and at the request of your Committee it has been
undertaken by myself. As yet, no very definite result has been
reached, but the observations already made, involving much care
and a very large expenditure of time, and the method of expeii*
menting employed, may perhaps be worth recording.

I should here say that throughout the investigation I have
had the benefit of the co-operation of Mr Charles Hockin, by whom
many of the observations were made.

To better observe the variation in the temperature-coefficient
with change of diameter, rods of considerable sectional area, to be
afterwards drawn into wire, were obtained fix>m Messrs Johnson
and Matthey.

The first rod (called hereafter Bar A) was of the commercial
alloy manu&ctured specially for electrical purposes; the metab
used are commercially pure, and are melted together in large
quantities. The alloy is then cast as a flat ingot, not more than
an inch or an inch and a half in depth, and this ingot is next

B. A. 20

306 PRACTICAL STANDARDS

rolled into a large sheet about 0*3 inch thick, which is cut by
shears into narrow strips.

These strips are finally passed between grooved rollers, to give
them an approximately circular section, and the rods thus formed
are ready for the draw-plates to reduce them to wire of the
required diameter.

The bar experimented on was, in the "rod" stage, about
8 inches long, and 0*27 inch in diameter.

The second rod (called hereafter Bar B) was specially made
for these experiments of pure materials, by Messrs Johnson and
Matthey.

It was cast in the form of a bar, about 8 inches long. On
leaving the mould it was about 0*3 inch square, slightly tapering,
itfid more or less irregular in section, but was reduced in the lathe
and by filing to a section almost absolutely square and uniform.

The third rod (called Bar G) was of an alloy made by the same
firm for the use of dentists — the method of casting and rolling
being the same as that described for the electrical alloy, but rather
less attention is paid to the purity of the components.

It was procured in the form of a narrow strip, about a quarter
of an inch thick, and reduced to a uniform square section by the
lathe and file.

As it was necessary to observe accurately the small percentage-
variation due to change of temperature, in the resistance of these
ban, which was itself exceedingly minute, recourse was had to the
method of observation originally proposed by Mr C. Hockin, and
daecribed and figured in Clerk Maxwell's Electricity and Magnttism^
pp. 406 and 407, vol. i., by means of which the unavoidable resist-
aace of the connexions can be altogether eliminated.

Instead, however, of using, as shown in the figure referred to, a
comparatively short wire, with resistance-bobbins at its ends, a
wire 40 feet in length, wound on a cylinder, was employed, so that
the bobbins could be dispensed with, without loss of accuracy, and
with a great gain in simplicity of calculation and in range.

To avoid the very great expense of a necessarily thick wire of
iridio-platinum, which was, however, recognised to be the best
material, a Qerman-silver wire was in the first instance fitted to
the cylinder and calibrated, but after a short time it was found to
get loose in its groove, so that the readings obtained on it could
not be depended on.

FOR ELECTRICAL MEASUREMENTS 307

A platinum-silver wire was next tried, and though most care-
fully drawn, the calibration showed such a want of uniformity in
the conducting power at different parts of its length that it was at
once discarded.

Finally, an iridio-platinum wire was obtained from Messrs
Johnson and Matthey, fitted to the cylinder and calibrated, but
the result not being considered quite satis£Eustory, the wire was
removed, carefully annealed, drawn through one hole in a draw-
plate and remounted, but with very little better results. This
operation was repeated without advantage, and it became evident
that the want of uniformity in the conducting power was not due
to irregularities in the section of the wire, but was to be attributed
in all probability, to want of uniformity in the composition of the
alloy.

To avoid further loss of time it was therefore decided to make
use of the wire as it then was, and to correct all readings by the
result of an accurate and close calibration.

The wire was therefore calibrated in 100 equal parts by a
method devised after ttying one or two others, and found to be
very accurate and convenient.

It is fully described at the end of this paper.

The wire is wound in twenty convolutions in a spiral groove,
accurately formed in the cylindrical surface of an insulating drum.
The ends of the wire are soldered to massive bars, each brazed to
one axle of the drum, which terminates in an amalgamated copper
disk, half immersed in a cup of mercury.

The mercury cups are themselves connected with the rest of
the apparatus by means of very stout copper rods.

The contact-piece, by means of which the galvanometer is put
in circuit, is mounted on a brass block, moving between two brass
rods, and traversed by a screw after the fashion of the slide rest of
A lathe. One end of the screw carries a toothed wheel, gearing
with another wheel attached to the drum and concentric with it.
The pitch of the screw and the gearing are so calculated that,
when the drum is made to revolve, the contact-piece, whilst moving
in a horizontal line parallel with the axis of the drum, is always in
•contact with some point of the wire, upon which it presses lightly
by means of a spring.

The brass toothed wheel has a slightly greater diameter than
the drum to which it is attached. On the flat exterior side, near

20—2

808 PBACTICAL STANDARDS

its circumference, it is divided into 1,000 equal parts, and by
means of a microscope with cross-wire eyepiece, the divisions can
be read by estimation to tenths and easily to fifths.

As there are twenty turns, the whole wire can therefore be
accurately divided into 100,000 parts.

Whole turns of the wire are shown by the divisions of a hori-
zontal scale close to the contact-piece.
Fi«. a.

To maintain the wire throughout its length at a uuifoTm
temperature, the drum is enclosed in a wooden case in which
openings are left for the contact-piece and microscope. A. sketch
of the apparatus is given in fig. 2.

To determine the temperature-coefficients of the various bare
and wires, their resistances at two different temperatures were
compared with that of a standard, maintained as nearly as possible
at a constant temperature.

The higher temperature of comparison was generally nearly
that of boiling water, and was maintained by means of steam.

FOR ELECTRICAL MEASUREMENTS 309

The bare were imcaeraed in a bath of melted paraffin wax, the
interior surface of the bath being lined throughout with convolu-
tions of ^-inch "compo" gas-tubing through which steam was
caused to flow.

The lower temperature was about that of the air, the bath just
described was again used ; but paraffin oil was substituted for the
wax, and cold water from the main was made to circulate in the
pipe instead of steam. In both cases the paraffin oil or wax was
kept continually stirred. The standard was also immersed in
paraffin oil, kept at a uniform temperature by the circulation of
water from the main, through tubing in the containing vessel.

The paraffin oil used possesses remarkably high insulating
qualities, bobbins of ailk-covered wire, even of many thousand
ohms resistance, could be immersed
in it, without the least sensible
leakage, from spire to spire. It is -
sold under the name of'Strange's
Crystal Oil."

The apparatus for maintaining
the desired temperature when the
alloy is in the form of wire is fer ^
more convenient than thatdescribed '
above for the bars. It resembles
the instrument used for testing the
boiling point of thermometers.

It is shown in section in the
accompanying sketch, fig. 3.

a is an open tube about |-inch
in diameter, surrounded by a second -
tube, 6, closed at the bottom and
opening at the top into the outer
tube, c, which is closed at both
ends. The supply pipe leads into
the annular space between a and b,
near the bottom, and the exhaust,
OF escape, pipe from the space be-
tween b and c.

When steam is allowed to enter by the supply pipe, it completely
envelopes, in its passage, the inner tube a, and external influences,
such as sudden draughts of air, are effectually cut off by the steam

SIO PRACTICAL STANDARDS

jacket between 6 and c. For observations at ordinary tempera-
tures water is allowed to flow into the supply pipe instead of steam,
and a convenient system of pipes and cocks allows the change to be
made from steam to water, or vice versa, with facility. When steam
is used, a mercurial gauge is provided, to indicate the difference
of pressure between the inside of the tube and the atmosphere.

The temperature calculated from the reading of this gauge,
and that of the barometer, was found to agree within -^ degree
with the indication of the thermometer inserted in the inner
tube a.

The bottom of the inner tube a is closed by an ebonite stopper,
through which pass two stout copper rods of semicircular section,
80 as to have the greatest available cross section; each is held
rigidly in position by a screw passing radially through the ebonite
stopper, and they are insulated one from the other by an air space,
through which pass the wires for the galvanometer connexions,
the interstices being afterwards filled with shellac.

The ends of the wire to be tested are soldered to the upper
extremities of the copper rods. When the wire is short and thick,
it is covered with silk, bent double, and passed into the tube;
when thin, it is wound on a glass rod, and the whole coated with
silk ribbon. The space between the wire and tube is filled with
lead shot of the smallest size, and a thermometer is inserted in the
top of the tube, which is then filled quite up with shot.

It was found that the shot had an excellent effect in prevent-
ing minute oscillations of temperature, due to draughts or similar
causes.

The batteiy used was one large groove cell ; the galvanometer,
a " dead beat *' by White, having a resistance of 1*4 ohms.

In most cases the absolute resistance of the specimens of alloy
was determined as well as the temperature-coefficient.

For the bars this was done by comparison with a rod of pure
lead of known dimensions.

The wires were compared directly with bobbins of known
resistance.

The sectional area of the bars and wires was always calculated
from their weight, specific gravity, and length.

All bars and wires were annealed before making the electrical
observations.

The following table shows the resistance and temperature-^

FOB ELECTRICAL MEASUREMENTS

311

coefficient of the first sample of the alloy when in the original bar
shape, and when drawn down to wire of various gauges.

Bar A.

Besistanoe of a

Percentage-

Length

Diameter

wire 1 metre long

Specific

variation in

Remarks

mm.

and 1 millimetre

gravity

resistance for

diameter at O^Cent.

1° Cent.

ohms.

175-67

6-826

0-3658

12-740

•0299

Rod

996-40

1-578

0-3824

12-733

-0276

3132-70

0-525

—

12-613

•0273

2655-25

0-525

0-3937

12-811

•0263

—

0-203

•0232

0168

— —

—

•0231

A complete analysis of this alloy from the end of the rod was
made for me by Messrs Johnson and Matthey. The composition
was as follows: —

Platinam 3295

Silver 6665

Iridium 0-03

Rhodium 0-01

Iron 0-01

Lead, with minute trace of copper . . . 0-01

Loss . . . ..... . 0-34

100-00

The platinum and silver are thus very nearly in the prescribed
proportion, and the amount of impurity is inconsiderable. An
assay of a piece cut from one end of the wire 1*578 mm. in diameter,
gave the proportion of platinum as 31*34 per cent.

A calibration made from inch to inch in its length showed the
conducting power to be sensibly uniform.

Bar B.

The temperature-coefficient of this bar was 0*0308 per cent*
The specific gravity and absolute resistance were abnonnal, and a
calibration showed it was quite irregular in conducting power^
the casting being porous and full of holes.

Messrs Johnson and Matthey failed in their attempt to roll or
draw it down into wire, but a small portion was remelted and
drawn into wire 0*67 mm. in diameter, of which the specific gravity

312

PRACTICAL STANDARDS

Was 13*854, the temperature-coefficient 0'0282 per cent., and the
calculated resistance of 1 metre of the wire, 1 millimetre in
diameter, 0'238 ohms. It was obvious, therefore, that even if the
alloy was of the desired composition, on the average, it had such a
want of uniformity that an analysis was not worth making.

Bar C {dental alloy).

The following table gives the observations made with this
sample of alloy from the bar stage to that of very fine wire.

Resistance of a

Percentage-

Length

Diameter

wire 1 metre

Specific

variation in

Remarks

mm.

long and 1 mm.

gravity

resistance

in diameter

for r Cent.

ohms.

168-00

4-314

0-3770

12-444

•0269

Square

996-95

1-625

0-3611

12-449

•0261

2416-15

0-535

0-3690

12-630

-0266

0-203

^—

— -

•0248

Messrs Johnson and Matthey's analysis of this alloy was as
follows : —

Platinum 28-95

Silver 70*50

Rhodium 0-01

Iridium 0-02

Iron 0-02

Lead, with minute traces of copper and loss . . 0*50

100-00

An assay of one end of the wire of the diameter 1*625 mm.
gave 28*44 per cent, as the proportion of platinum ; but it would
appear, from the third observation of specific gravity in the above
table, that the composition could not be uniform — indeed the bar
was calibrated in six parts, both at the high and low temperature,
and the temperature-coeflScients for the different parts were found
to be as follows : 0*028, 0-032, 0029, 0*036, 0*031, and 0-029 per
cent, per degree.

To examine the effect on the temperature-coefficient of a
variation in the percentage of platinum and silver, the following
alloys were cast by Mr Hockin and drawn to wire. They were
made with pure platinum black and precipitated silver.

FOR ELECTRICAL MEASDREHENTS

313

Composition of alloy

Diameter
approximate

Peroentage-

yariation in

resistance for

V Cent.

Remarks

33|
30
25
25

60
66|

70
75
75

mm.

0-234

0096

0-234

0-063

•0259
•0265
•0301
•0313
■0407
•0377

[ different castingH

From the foregoing tables it would appear that a moderate
variation in the percentage-composition on either side of the
normal proportions of two parts silver, by weight, to one part
platinum, produces less effect on the temperature-coefficient than
does a change in the diameter of the wire.

It also appears, from these and other experiments, that it is
practically impossible to ensure a uniform mixture of the metals,
even when the alloy is of the normal composition in the aggregate.

Thus the wire from an unit coil of the B. A. pattern, made by
Dr Matthiessen, or under his superintendence, was drawn down to
a diameter of 0*168 mm.

The whole wire had a temperature-coefficient of 0*0250 per cent.,
but on examining the wire in two approximately equal portions,
the temperature-coefficients of the two halves were found to be
0*0237 and 0*0266 per cent.

It is therefore evident that when the highest attainable
accuracy is acquired, as in the construction of standard coils, it is
not sufficient to depend upon the general temperature-coefficient
of the alloy, but that a determination of the coefficient of the
particular coil or instrument is required.

For less accurate work it would seem that Dr Matthiessen's
value for the coefficient, viz., 0*031 per cent, per 1** C. should be
reduced by from 5 to 10 per cent, for wires of large diameter, by
about 15 per cent, for wires of 0*25 mm. diameter, such as, in
general, resistance-coils of from 100 to 1,000 ohms are made of,
and by from 20 to 25 per cent, for wires of the smallest gauge
usually drawn.

314 PRACTICAL STANDARDS

Calibration of Wire.

The calibration of the wire was effected thus — Six coils of the
B. A. pattern, viz., la, 16, 2, 3, 5, and 8 ohms, whose resistance
amounted, in the aggregate, to 20 ohms, were arranged in a con-
tinued series by means of mercury cups, and the ends of this series
were connected by copper bars with the two ends of the wire to
be calibrated, properly mounted on its drum. To these copper
bars the poles of the battery could be connected by means of a
contact-key.

The series of coils was arranged, in the order above enumerated,
in a trough of water, to maintain uniform temperature. One
galvanometer wire was permanently attached to the sliding contact
of the drum, whilst the loose end of the other galvanometer wire
could be dipped in either of the mercury cups joining the coils, the
arrangement thus forming a Wheatstone's bridge.

The loose galvanometer wire, being first inserted in the mercury
cup joining the outer or left-hand terminal of the coil la. with the
connecting bar, a balance was obtained, the reading r^ on the drum
being of course quite close to the 0 end of the wire.

The gal vanometer wire was then shifted to the cup between la
and lb, and a balance and reading ri obtained by moving the
drum-contact.

The coils la and 16 were then transposed in position, and
a new balance and reading rj, very near to r,, were obtained*
The loose wire was then moved to the cup between la and 2, and
the balance and reading r, were observed.

This process of transposition and reading was repeated until
the coil la had been moved unit by unit trom the left-hand to the
right-hand extremity of the series.

As it is evident that the resistance of the length of wire
between pairs of readings such as ro, rj, and r^,r^i etc., bears the
same ratio to that of the whole wire to be calibrated as the resist-
ance of the coil la does to that of the series of which it forms part,
and that the latter ratio is a constant one, being independent
of the position in the series which the coil occupies, it follows that
every such length has the same resistance ; and the lengths being
expressed in terms of divisions of the drum circle, it is easy to
make a table showing the proper corrections.

FOR ELECTRICAL MEASUREMENTS

315

In practice, the coil la was further subdivided Four inter-
mediate points having been determined on the wire of which it
was composed, dividing its resistance into five equal parts, wires
were soldered to these points and were connected with small

IV.

DifFerenoes of

Percentage-

V.
Readings

Observed
readings

readings propor-
tional to Condno-

variation from
mean Gondno-

corrections for
Readings in

tivity

Col.V.

3-3

10091

1006-8

+ 1-015

1-015

1000

1012-9

2016-9

1004-0

+0-835

0-925

2000

2020-2

3021-0

1000-8

+0-513

0-788

3000

3025-6

4024-9

999-4

+0-373

0-684

4000

4028-6

5027-6

998-9

+ 0-322

0-612

5000

6031-1

6030-4

999-3

+0-363

0-670

6000

6034-2

7033-9

999-7

+ 0-403

0-546

7000

7039-8

8034-9

996-1

-0-069

0-471

8000

8038-7

9037-3

998-6

+0-292

0-451

9000

9041-8

ioa37-i

995-3

-0-039

0-402

10000

10040-7

110341

993-4

-0-230

0-344

11000

11038-0

12033-7

996-7

+0-001

0-316

12000

12033-6

13028-8

996-2

-0-049

0-288

13000

13032-3

14026-6

994-3

-0-140

0-267

14000

14030-4

15022-0

991-6

-0-411

0-213

16000

16027-6

16017-9

990-3

-0-541

0-166

16000

16022-4

17011-0

988-6

-0-712

0-114

17000

17014-9

18006-4

991-6

-0-421

0-084

18000

18010-0

19000-0

990-0

-0-571

0-049

19000

19003-7

19990-0

986-3

-0-943

0-000

20000

Mean

996-69

316 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

mercury cups sunk in the ebonite bridge-piece of the coil. The
case was then filled in with paraffin wax as usual. In calibrating
the drum-wire, therefore, besides the 20 pairs of readings corre-
sponding with the 20 transpositions of the coil, there were inter-
polated between each such pair four additional readings, thus
calibrating the wire into 100 parts.

The table on p. 315 gives the readings corresponding to the
twenty transpositions of the coil, their differences, which are pro-
portional to the conducting power of the wire between the points
at which the readings are taken, and the percentage-variation
from mean conducting power.

In Col. IV. are given the total percentage-corrections for the

readings in Col. V. They are obtained by taking from Col. III.

the algebraic sum of the percentage-variations for all observations

included in the reading, and dividing by the number of terms

summed ; thus the correction for drum-reading 3,000, and there-

, ^ . 1015 +0-835 + 0-513 ^^^^ 4. r ^u j-

abouts, 18 = 0-788 per cent, of the readmgs.

All these corrections happen to be subtractive.

It is of course unnecessary that the subdivisions of the 1-ohm
coil should be exactly equal, provided their ratio is known, and
this can easily be found by the drum-wire itself Thus the coil is
substituted for the series of 20 ohms, before referred to, and the
readings observed on the drum when galvanometer-connexion is
made consecutively at the six contact-points of the divided coiL
These observations will coincide very nearly with readings in the
foregoing table which are independent of the subdivisions of the
coil. Therefore, by applpng the proper tabular corrections, the
true ratio of these subdivisions is determined.

NINTH REPOET— SOUTHAMPTON, 1882.

The Committee have to report that Mr Taylor has continued
the experiments upon the temperature-coefficient of the resistance
of metals and alloys, the first results of which were communicated
at the York meeting. In consequence of Mr Taylor's absence fix)m
the country, the details of the further experiments cannot be com-
municated at present ; but it may be stated that they have shown
the possible influence of the process of annealing on the specific
resistance of wires and on the temperature-coefficient to be much
greater than has hitherto been commonly supposed. The following
are examples of some of the results obtained : —

Oerman-Silver, Wire drawn to be extremely hard and brittle.
The percentage- variation of resistance, for 1° between 13° and
100° C, was 0*0296. After annealing, the percentage-variation of
the same wire was 0*0421.

Steel. Wire, 0025 inch diameter, thoroughly hardened, and
then tempered in paraffin wax at 230° C. :

Percentage-variation of resistance for 1°, 0*267.

Same wire annealed ; percentage-variation for 1°, 0'316.

At 9° C, the ratio of the absolute resistance of this wire in the
hard state to that of the same wire when annealed was 1*229.

Platinum-Silver Alloy. A piece of wire made fix)m a particular
bar of the alloy was hardened by being drawn down through
a couple of holes of the draw-plate. In this state the variation of
resistance was 0*0255 per cent, per degree. After annealing in
the ordinary way, the variation of resistance per degree was
0*0258 per cent. The same wire was next placed in an iron tube,
which was filled up ynth sand and left all night in the fire. After
this treatment, the percentage-variation of resistance per degree
was 00344.

Plattnuni'Silver Alloy, another specimen. A wire from a
second bar of the alloy was annealed at a very high temperature

318 PRACTICAL STANDARDS FOB ELECTRICAL MEASUREMENTS

and left to soak in the fire and cool slowly, as in the last-men-
tioned experiment. The variation of resistance was now 0'095 per
cent, per degree, and the wire was as soft as pure silver and very
fi:ugile. After being heated to redness and quenched in water,
the corresponding variation of resistance of the same wire was
0*076 ; and when the wire had been drawn down through two or
three jewel-holes it was 00732.

These results indicate a connexion between the temperature-
coefficient of wires and their degree of hardness, and tend to
reopen the question as to the most trustworthy material for a
permanent standard of resistance. The Committee understand
that Mr Taylor will continue his experiments with the co-operation
of Dr Muirhead.

The Committee are pleased to be able to report that there
is a prospect that Lord Bayleigh may be able to organise, at the
Cavendish Laboratory, Cambridge, a system of testing resistance-
coils and issuing certificates of their correct value at a specified
temperature.

As stated in the Report presented last year, Dr Muirhead has
consented, at the request of the Committee, to issue standards of
capacity upon his own responsibility.

The Committee regret that they are not able to report any
progress towards the construction of a standard of Electromotive
Force.

They are unwilling to conclude without expressing their deep
sense of the loss which not only they, but all friends of physical
science, have suffered in the death of one of the most valued of
their colleagues, Mr Charles Hockin.

TENTH REPORT— SOUTHPORT, 1883.

The Committee report that, in accordance with suggestions made
at the last meeting of the British Association, arrangements have
now been completed for testing resistance coils at the Cavendish
Laboratory and issuing certificates of their value. These arrange-
ments have been made by Lord Rayleigh and Mr Qlazebrook, and
the report contains an account by the latter of the methods
employed and the conditions under which the testing is under-
taken, in order that those who use such coils may have a more
exact estimate of the value of the test.

The standards at the laboratory belonging to the Association,
the values of which have been recently tested, are all single units.
The best of these were all compared among themselves, originally
by Hockin {British Asaoddtion Report, 1867), and again by
Chrystal and Saunder (Report, 1876), and more recently, at
various temperatures between 0** C. and 25** C. by Mr Fleming in
1879 — 1881, and a chart has been constructed, from which the
resistance of any one coil at a given temperature between these
limits can be determined. On this chart a curve is drawn for
each coil ; the ordinates of the curve represent resistances, while
the abscissae give the temperatures. The temperatures at which
the various resistances were originally each one fi. A. unit are
known for the respective coils. For these temperatures the
ordinates of the curves drawn ought to be the same, and the
corresponding resistance one B. A. unit. Mr Fleming finds, how-
ever, that this is not the case. The resistances of the eight coils
examined at the temperatures at which they were originally said
to be correct are slightly different The greatest difference is
that between the coils marked C and Oy and amounts to O'OOll
mean B. A. unit.

The mean of all these resistanees at the respective temperatures

320

PRACTICAL STANDARDS

is taken as the mean B. A. unit, and is that to which the resist-
ance coils sent for testing are referred.

The coils examined are those marked as below in previous
reporta

A
2

B
3

C
58.

D
35

E
36

F
29

Q
43

Flat

1876

Flat

1867

In comparing the single unit coils the form of resistance bridge
devised by Mr Fleming and described by him (Proceedings of the
Physical Society, vol. ill.) is employed.

The bridge, with battery, keys and a suitable galvanometer, is
permanently fitted up in a ground-floor room with a north aspect.
The standard coils are kept in a case in the same room, and the
baths in which the coils are to be immersed are always ready filled
with water, which is thus at the temperature of the room.

When a coil is to be tested, a suitable standard is chosen, and
the two are placed in the water baths and left at least three or
four hours — more usually over night. The comparison is then
made in the ordinary manner by Professor Carey Foster s method*,
and the coils again left for some time without being removed from
the water. After this second interval another comparison is made.
The temperatures of the water baths are taken at each comparison,
and as a rule differ very slightly.

We thus have two values of the resistance of the coil to be
tested at two slightly different temperatures.

The mean of these will be the resistance of the coil in question
at the mean of the two temperatures.

We are thus able to issue a certificate in the following form: —
" This is to certify that the coil No. X has been compared with the
British Association Standards, and that its value at a temperature
of -4** C. is P B. A. units or P' R. ohms; 1 B. A. unit being
•9867 R. ohms." We further propose to stamp all coils in the
future with this monogram J^ and a reference number.

One single unit coil by Messrs Latimer Clark, Muirhead, & Co.^
three by Messrs Elliott Brothers, for Professor Mascart, and one by
Messrs Simmons & Co., have been tested.

It will be noticed that nothing is said about the temperature
coefficient of the coil or the temperature at which the coil is

* Journal of Soc* of Telegraph Enffineers, 1874.

FOR ELECTEICAL MEASUREMENTS 321

accurately 1 B. A. unit. To determine this exactly is a somewhat
long and troublesome operation, but at the same time it is one
which every electrician, if he knows the value of the coil at one
given temperature, can perform for himself with ordinary testing
apparatus. It does not require the use of the standards. For
many purposes the approximate value of the temperature co-
efficient obtained from a knowledge of the material of the coil will
suffice ; we may feel certain that anyone requiring greater accuracy
would be quite able, and would prefer, to make the measurement
himself We can state with the very highest exactness that the
resistance of the coil X at a temperature ^"^ C. is B. To obtain
the temperature coefficient accurately requires an amount of labour
which may be quite unnecessary for the purpose for which the coil
is to be used.

But it is requisite to have standards of higher value than one
unit, and part of the Association grant has been used in obtaining
coik of a resistance of 10, 100, 1000 and 10,000 units. Two of
each value have been purchased, so that by frequent comparison
of one with the other any accident to either may be checked.

It remains, therefore, to describe how these coils are to be
referred to the standards. For the 10 units two methods have
been adopted.

There are at the Cavendish Laboratory two 5-unit coils.
Each of these was compared with five single units placed in series,
using Fleming's bridge to make the comparison, and the 10-unit
coil was compared with these two in series.

The values obtained by two observers at a temperature of 12""
were: —

9-98360 Lord Rayleigh.

9-98393 R. T. G.

For the second method, suppose we have three coils each of
resistance about 3 units. Let these be 3 + a, 3 + y3 and 3 + 7,
then the resistance of the three in series is 9 + a + y3 + 7, and in
multiple arc, if we neglect terms like o? ^, etc., it is

i+4(« + y3 + 7),

thus neglecting terms such as a* ^, the resistance of the three in
series is just nine times that of the three in multiple arc.

But the three coils in multiple arc are very nearly one unit»
and can be compared with the standards. If then we combine in

B. A. 21

322

PRA.CTICAL STANDARDS

series with the same three one of the standards we have a resist-
ance of approximately ten units, whose value is very accurately
known, and with which any other 10-unit coil can be compared
by the aid of Fleming's bridge. Lord Rayleigh has devised an
arrangement of mercury cups, by means of which the changes
indicated can be easily performed.

The three 3-unit coils are wound on the same bobbin, and
inclosed in the same case. The six electrodes project in pairs,
and their ends lie in a plane. The figure represents a piece of
ebonite, through which holes are cut as indicated by the letters
a, by etc.

c' d

c e

h' d'

B h d f

On the under side of the ebonite, strong strips of copper, with
their faces well amalgamated, are screwed, forming with the holes
in the ebonite a series of cups, which are filled with mercury.

The copper strips are cut, as shown in tha figure, to make the
necessary connexions. The distances between the holes is such
that the electrodes of the three coils respectively fit into a b, c d,
and ef, or into a' i', c d\ and e'/'.

Connexion is made with the bridge by means of the cups A, B,
while the electrodes of the second single unit coil fit into g and A.
In the first position the three coils are in multiple arc, as will be
seen from the figure, and can be compared with a single unit,
while in the second they are in series with the other single unit,
and can be compared with the 10 units.

By this contrivance the 10 unit is referred to the single standard.

FOR ELECTRICAL MEASUREMENTS 323

To determine the value of a coil of 100 units, the three 3 units
can be replaced by three 30 units, and the single units by tens.

This, however, is not the most convenient method for the total
resistance if the wire of the Fleming bridge in use is only ^ of
a unit, thus affording too small a range for the ready comparison
of large resistances.

The following has been adopted : — Four coils are arranged as
in a Wheatstone's Bridge, one being the 100 units to be tested,
two of the others, in opposite arms, two known 10 units, and the
fourth a known single unit.

These coils are all arranged in the same circular trough of
water and their electrodes dip into four mercury cups.

If all the coils are correct no current will traverse the galvano-
meter. Of course in practice this condition is never realised.
Either one of the ten units or the single unit is too great. Let us
suppose it is the latter ; connect its two electrodes with the two
electrodes of a resistance box and take out plugs from this till
a balance is secured. Then if the resistance of the ten units be
Q and R, that of the single unit 8, and the shunt W, the resist-

W S
ance of the shunted arm is ^™ — -, and that of the 100 units is

Now, in practice, if Q, R, S are fedrly accurate, W will be
a large resistance, and an approximate knowledge of W will suffice.
W may thus, for all we require, be taken from a resistance box by
a good maker which has stood for some time in the room in which
the experiments are conducted, the temperature being taken as
that of the room. A box has been ordered from Messrs Elliott
Brothers, to be used for this and similar purposes.

The same firm have also supplied a high resistance galvano-
meter for the testing.

Of course if one of the 10-unit coils is too great, then the
shunt W must be put in with it.

In accordance with the resolution of the Committee, a fee of
£1. Is, has been charged for testing single units, and of £1. 11^. 6(2.
for others.

The only coils the testing of which is regularly undertaken are
single units and multiples of single units by some powers of 10.

But though this is so, two standard ohms have been ordered,

21—2

324 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

using for the value of the B. A. unit '9867 ohms, and when they
arrive and have been tested, it will be easy to determine the value
of coils which do not differ much from a real ohm. At present,
without these standards — the coils actually used in the recent
experiments at the Cavendish Laboratory have a resistance of
about '1, 24, and 168 ohms — the operation is troublesome. The
simplest accurate method seems to be to combine in multiple arc
the real ohm, and one of the 100 B. A. unit standards, and to com-
pare the combination with a single unit.

Dr Muirhead also reports the completion of three air con-
densers as standards of capacity.

The Committee are glad to learn that Lord Rayleigh is con-
tinuing his valuable researches at the Cavendish Laboratory with
the view of obtaining an absolute unit of current.

They would ask in conclusion that they may be reappointed
with the addition of the names of Mr H. Tomlinson and Professor
W. Gamett ; and that a further grant of £100 may be made to
meet the expense of procuring standards of resistance in terms of
the ohm.

ELEVENTH REPORT— MONTREAL, 1884.

The Committee report that during the year the construction
and testing of standards of electrical resistance has been proceeded
with. The coils of 10. 100, 1000 and 10,000 B. A. units, men-
tioned in the last Report have been compared with the standard
unit coils. An account of the comparison made by the Secretary
and Mr H. M. Elder, with a table of the values arrived at, is given
in Appendix L Further experiments on the temperature co-
efficients of these coils are in progresa During the year, twelve
coils have been compared with the B. A. standards, and certificates
of their values issued by the Secretary*.

At the Southport meeting of the Association a grant was made
to defray the expense of procuring standards of resistance in terms
of the ohm. At a meeting of the Committee held in March, 1884,
it was decided to defer the purchase of these till after the meeting
of the Paris Congress, and a resolution was passed to the effect
that " In the event of the Paris Congress adopting any definite
standard of resistance, standards be ordered for the Committee in
accordance with that value."

The Paris Congress adopted as a standard, to be called the
" legal ohm," the resistance at 0"* C. of a column of mercury
106 centimetres long, and one square millimetre in section. The
standard resistances at present in use being B. A. units, it became
necessary to assume a relation between the B. A. unit and the
legal ohm, in order to construct coils whose resistance should be
one legal ohm. This relation has been determined by various
observers with slightly different results, and a meeting of the
Committee was held on June 28 to consider the question. At
this meeting the following resolution, proposed by Professor W. Q.
Adams, seconded by Lord Rayleigh, was carried : — " That, for the

* In the original Beports tables were giyen of the valaee foand for ooile
submitted for test. Such resalts are not of general interest and they have,
therefore, been omitted here.

326 PRACTICAL STANDARDS

purpose of issuing practical standards of electrical resistance, the
number of B. A. units adopted as the resistance of a column of
mercury 100 cm. in length, 1 sq. mm. in section, at 0° C, be "QS^jO."
Taking this number, then

1 legal ohm = 1-0112 B. A. units.
1 B. A. unit = '9889 legal ohms.

Coils having respectively a resistance 1, 10, 100, 1000 and
10,000 legal ohms have been ordered, two of each value, so that,
by frequent comparison of one with the other, an accident to
either may be checked. These standards are to have their correct
values at temperatures near 15° C.

The two 1-ohm coils have been sent by the makers, and their
testing is being proceeded with. When this is complete the Com-
mittee will be in a position to test and certify to the values of
coils in terms of the legal ohm.

They propose that the certificate should run as follows : —

" This is to certify that the resistance coil X has been tested
by the Electrical Standards Committee, and that its value at a
temperature of 4° centigrade is P legal ohms.

" It has been assumed, for the purposes of this comparison, that
one legal ohm is equal to 1*0112 B. A. units."

The coils will be stamped with the monogram jfr^ and a refer-
ence number.

A portion of the grant has been expended in some additions to
the wire bridge belonging to the Committee, which have added
greatly to its utility, while two thermometers for the testing room
have been purchased.

The Committee would ask, in conclusion, that they may be
reappointed, with the addition of the name of Mr W. N. Shaw, in
order to continue the work of issuing standards of resistance.

FOR ELECTRICAL MEASUREMENTS

327

Appendix I.

On the values of the B, A, standards of resistance greater

than one B. A. unit

The coils of approximate value 10 B. A. units marked Elliott
Bros., No. 66, ^ 20, and No. 67, "^ 21, respectively, were
compared with the B. A. standards by the method described in the
last Report*, with the results given in the following table : —

Mark of coil

Date

Value found
in B.A.U.

Temperature

Elliott No. 66
^ No. 20

July 5
July 7

10-0065
10-0043

19-1'
18-3**

Elliott No. 67
^ No. 21

July 5
July 7

10-0060
10-0043

19-r

18-3*

Elliott No. 68
^ No. 22

July 24
August 11

100-038
100115

16-7'*
19-9°

Elliott No. 69
^ No. 23

July 24
August 11

100-024
100-097

16-7°
19-9'*

Elliott No. 70
^ No. 24

July 26
August 11

999-79
1000-78

15-8°
19-9'*

Elliott No. 71
^ No. 26

.1
. July 26

August 11

999-81
1000-79

16-8'
19-9'*

Elliott No. 72
^ No. 26

Elliott No. 73
;^ No. 27

■ ' 1
August 11

10006-2

19-8''
19-8*

1
August 11 10006-9

B. A. Beport 1888, p. 822.

328 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

The coils were immersed in the water bath, the temperature of
which remained constant during each observation, for some days
before the measurements were made.

The values thus found were used for the determination of the
coils of higher resistance, the methods of the last Report* being
employed in this case also. The insulation of the various parts of
the apparatus was tested carefully. Each result given in the table
is the mean of two or more determinations at the same tempera*
ture. The readings of the thermometer used were compared with
those of a standard instrument.

* B. A. Report 1883, p. 823.

TWELFTH REPORT— ABERDEEN, 1885.

The Committee report that during the year the standards of
resistance, in terms of the legal ohm referred to in the last Report,
have been constructed, and their values determined in accordance
with the resolution adopted on June 25, 1884.

The 1-ohm standards were generally referred to the original
B. A. units of the Association by combining in multiple arc with
the standard one of the 100 B. A. units, and determining by Carey
Foster's method the difference between the combination and a
B. A. unit, and then assuming, in accordance with the resolution
that 1 B. A unit = '9889 legal ohm.

The following values were thus found for the two standards : —

Resistance Coil, EllioU, No. 189, "^ 100.

Date

Temperature

Resistance

Nov. 24, 1884

11-4'

^ -99878

„ 26, „

11 -6"

•99890

„ 27, „

12-9"

•99916

„ 28, „

13-5"*

-99930

Dec. 5, „

13-5'

•99931

M 12, „

15'3"

-99979

July 30, 1885

17-2-

1-00027

»» 28, „

i8-r

1-00061

Mean value

Temperature coefficient

-999515, at 142" C.
•000266.

Resistance Coil, Elliott, No. 140, ^ 101.

Date

Nov. 24, 1884

„ 26,
Dec. 2,
Nov. 27,
Dec. 6,

„ 12,
July 30, 1885
„ 29, „

Temperature , Besistance

11 •4*'

•99826

11 -5°

-99827

12-8'

•99847

12 9*"

•99851

13-4"

•99865

15-4"

•99917

17-2'*

•99961

18-0°

•99983

Mean value

Temperature coefficient

•998845, at 14-1' C.
•000236.

330

PRACTICAL STANDARDS

The temperatures were taken by a thermometer graduated to
tenths of a degree centigrade, which had been compared with the
Eew standards.

The 10-ohm standards were then compared with the 1-ohm
by means of the arrangement suggested by Lord Rayleigh, and
described in the Report for 1883, and from these values were
obtained for the coils of higher resistance.

The results are contained below.

No. of Coil

Resistance

Temperature

No. 141, ^ No. 102

1000103

16-7"

No. 142, '^ No. 103

10-00169

16-75**

No. 143, '^ No. 104

99-9977

16-05'*

No. 144, ^ No. 105

100-0108

1605'

No. 145, ^ No. 106

1000-306

17-4°

No. 146, ^ No. 107

1000-276

17-4'

No. 147, ^ N<^108

10002-4

17-35°

. No. 148, ^ No. 109

10002-4

17-35'

These experiments were carried out at the Cavendish Laboratory
by the Secretary and Mr H. Wilson, of St John's College.

At the request of M. Mascart, the Secretary compared with
the legal ohms of the Association three mercury copies of a legal
ohm, constructed by M. J. R. Benoit, of Paris. A detailed account
of these experiments was laid before the Physical Society*. The
values found are given below.

No. of Tube

Value found by
M. J. R. Benoit

Value found by
R. T. G.

Diff.

37
38
39

1-00045

1-00066

-99954

-99990

1-00011

•99917

•00055
•00055
•00037

Mean ' 1-00022

-99973

•00049

♦ Phil. Mag, Oct. 1886.

FOR ELECTRICAL MEASUREMENTS 331

The work of testing resistance-coils has been continued.

The Committee hope that arrangements may be made for
issuing standards of electromotive force and constructing standards
of capacity. In conclusion, they would ask to be reappointed,
with the addition of the names of Professor J. J. Thomson and
Mr W, N. Shaw, with the renewal of the unexpended grant
of £50.

THIRTEENTH REPORT— BIRMINGHAM, 1886.

The Committee report that the work of testing resistance-coils
has been continued at the Cavendish Laboratory : —

Messrs Elliott Bros, called the attention of the Secretary,
during the spring of the current year, to the fact that in some
of the coils the paraffin used for insulation acquired in time a
greenish tinge, which is most marked round the interior of the
case and round the places at which the copper of the connecting
rods comes in contact with the paraffin. Careful examination
shows this green tinge in almost all the coils, and an analysis of
the paraffin made by Mr Robinson, of the Chemical Laboratory,
Cambridge, proved the colour to be due to a very slight trace of
copper. The insulation resistance of several of the standards was,
therefore, tested by passing the current from 24 Leclanch^ cells
through a high resistance galvanometer, and the coil from the
case through the paraffin to the wire. This resistance for most
of the coils tested was found to be from eight thousand to ten
thousand megohms. One coil in particular, sent by Messrs Elliott,
in which the green coloration was most marked, had a resistance of
5000 megohms. Thus it is clear that the resistance of the coils has
not hitherto been seriously aflfected by the presence of the copper in
the paraffin, but at the same time it becomes necessaiy to watch
closely for any changes which may occur, and to select very
carefully the material used. There appears to be great difficulty
in getting rid of all the acid employed in the manufacture of the
paraffin.

The only coil among those tested which showed an insulation
resistance, so low as to be serious, was the one known in the
Reports as Flat. When the galvanometer of 1700 ohms resistance
was shunted with 4 ohms a deflection of 80 divisions on the scale
was obtained. The same deflection was obtained when the
resistance in circuit was a megohm and the shunt was about

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 333

20 ohms. Thus the insulation resistance of Flat was only about
^ megohm, or 200,000 ohms.

Two coils of special interest have recently been sent to be
tested. One from Prof. Himstedt, of Freiburg, will connect his
determination of the ohm with those made in Cambridge ; while
the second is a coil of 10 B.A. units from the Johns Hopkins
University, which has been compared with the coils used in the
determination of the ohm there. The results of the observations
on these coils are, however, not yet completely worked out.

The Committee wish to express their sense of the great
desirability of establishing a National Standardising Laboratory
for Electrical Instruments on a permanent basis, and their willing-
ness to co-operate in the endeavour to secure the same.

The Committee have had under consideration the question
of the means to be taken to secure the general adoption of the
Resolutions of the Paris Congress.

The Committee have received by the kindness of the French
Grovemment a specimen of the platinum iridium wire, of which it
is proposed that the French National Standards of resistance
should be constructed. They hope shortly to make a series of ^
measurements of its specific resistance and temperature coefficient.

In conclusion they would ask to be reappointed, with the
addition of the name of Mr J. T. Bottomley and a grant of £50.

Appendix.

On the Values of some Standard Resistance Coils. By the
Secretary and T. C. Fitzpatrick.

In the last Report the values of the Standard Legal Ohm
Coils of the Association are given. For the 1-ohm coils the
temperatures range from 11^ to 18^, while the coils of higher
resistance were examined only at temperatures near l7^ It was
necessary in all cases to extend the range of temperatures in order
to determine the temperature coefficient. The observations were
made by the methods ab'eady described in the Reports, and the

334

PRACTICAL STANDARDS

values found are given in the following tables in which the previous
results are included : —

Resistance Coil, ^ 100.

Date

Temperature

Resistance

Nov. 24, 1884

11-4"

•99876

1, 26, „

11-6"

•99888

» 27, „

12-r

•99916

1, 28, „

13-5'

•99930

Dec. 5, „

13-5"

•99931

„ 12, „

15-3"

•99979

July 30, 1885

! 17-2'

1-00027

11 28, „

l&V

1-00061

Mean value

Temperature coefficient

•999610 at 14-18*.
•000271 per 1** C.

Date

Temperature

Resistance

Nov. 21, 1886

7**

•99753

>> 24, „

7-5''

•99770

,, 23, „

8-1'

•99787

Jan. 30, 1886

11-4"

•99876

Nov. 30, 1885

116**

•99878

Jan. 22, 1886

12^5''

-99906

Nov. 30, 1886

12-6"

-99911

Mean value

Temperature coefficient

Mean value of whole series ...
Temperature coefficient

•998401 at lO^lO*.
•000274 per 1* C.

•998770 at 12^28°.
•000272 per 1' C.

These results are represented graphically in Plate 8 by the
curve J^ 100, which is drawn through the means derived from
the two series, and represents within the limits of accuracy of the
experiments all the observations of the two series, the mean error
from the curve, omitting one observation, being about '00002.

In the diagram the circles indicate the observations of 1884-5,
the dots those of 1885-6.

'ixU/Ji

1 ■ A »» '.•»'< '.' I"

k »

jaV .ih: ^n T/^j^hD

/IQ .AniT t.;V JHT
r '. — -J— ^ T-— . -J. - ^ ,- . -J- J- — p I ■ I f 1 r"f"T ■*

*(• •^ ^ - •* ^

■i --^^

t ♦'

4 -t :

»i

f~* -♦

♦ i-

— ■ -t -

^*-f "

-i > -♦--

p I

■• t-

;. I.-,-

•- 'I

* fc —

, I I I I I ■

I r

' f '

- t

•» » ♦

I ...i . ^ . ^. 4

-t-

-^ ^ »- »* ^^ A» * • A« •4^- ^^

: r —

I ...

-« « B«

• - •- -

' . .^ ♦ .», 4 -4-

^ ..-.

♦*-r f ♦

- t A

+ » • •

-* - i - .. '

i i

. -4.

t ♦ ♦- f

] %.

♦. I- 4 - i- . -i

•C' »>»>«.•

'-»-►- I

t--t V

I
• - ♦

I .

•1;. ;

- -f —

• ■(-

♦ • t- •

' * t

• ♦

1 )■ •

- 1 *

» ♦ - ' ... _<{..-

, ■ . . . ■ i 4- • 4 /

^ .;—.-♦ |. . -• ■ .

■. i «.{ t - - - I

. . .» •««-, - ■♦ 1- f. i. .^— 4Ji.- » ., ^

I ' •

«, . . • i.i. i ^-,: i. i.-4 ...V

I ' r

*- ^-v l--i ^- 4.-1 -.^L.-4 4. — .

1-^.9*

■^1.

f r-

f *,.--*'

f— -♦ 4 .—

I. i . -

- i —

r 4 — i-

- r "

,-- f

i ~

I 1 1 4 . ;

- 4- •

• _ • * • 4

,. . t J- . .

If • ♦ r

"'•/."v.

»V'

■ t -»--

• r

.'A .^ %*«u

.. / •>.■

FOR ELECTRICAL MEASUREMENTS

335

Resistance of Coil, ^ 101.

Date

Temperature

Resistance

Nov. 24, 1884

11-4'

•99813

n 25, „

11-5"

•99816

Dec. 2, „

12-8'

•99847

Nov. 27, „

12-9"

•99851

Deo. 5, „

13-4**

•99865

., 12, „

15-4°

•99917

July 30, 1885

17-2"

•99961

»» 29, „

18"

•99983

Mean value -998816 at 1416°

Temperature coefficient ... -000259 per V C.

Date

Temperature

Resistance

Nov. 21, 1885

6-9*'

-99677

j» -^4, „

7.70

•99698

jj 23, „

7-9*'

•99704

Jan. 20, 1886

11 -3"

•99793

Nov. 30, 1883

11 -8"

•99803

Jan. 22, 1886

12-4''

•99821

Nov. 30, 1885

12-6''

•99834

Jan. 26, 1886

139*'

•99868

yy 28, „

14-3°

•99876

Mean value

•997860 at 10-98°.

Tempe

irature coefficient f

rom this series

•000272 I

)er 1* C.

On plotting these results it becomes clear at once that the
straight line joining the means of the two series will not represent
the results at all.

The first series is represented by the upper curve ^ 101 (1),

the second series by the lower curve J^ 101 (2).

Thus it would seem that between November 1884 and
November 1885 this coil had lost in resistance about '00015 ohm
at a temperature of 12° C. Again, the two curves are not parallel,
so that it would seem at first sight that the temperature coefficient
also has altered; but this inference is hardly justifiable, for the
experiments in series (1) cover the time froia November 1884 to
July 1885, the high temperature observations being made at the
later date ; if then during that period the coil was decreasing in

336 PRACTICAL STANDARDS

resistance the temperature coefficient would necessarily be too
low ; moreover we notice that the observations for July 1885 do
not lie very far from the curve which represents the results of the
second series.

We infer then that of the two coils of platinum silver made at

the same time — two years from the present date— one J^ 100 has
not changed since that date, and has a value of

•998770 legal ohm at 12-28'*

with a temperature coefficient of 000272, while the other has
changed by about '00015 ohm, and now has a value of '997860 at
10*98° and a temperature coefficient also of '000272.

The feet that the temperature coefficient of ^ 101 is the

same as that of '^ 100 would appear to show that it has now

reached its permanent state.

Messrs Elliott Bros, possess a standard ^ 63, made at the

same time as the above two coils which in August 1884 had a

resistance of

1'00027 legal ohms at 18•8^

while in April 1886 it was found to be -99928 at 16'6'* and

•99992 at 18'6°.

From this it follows that its value at 18-8** would be '99998,
indicating a fall of '00029 in a year and eight months.

This coil showed marked traces of the green coloration referred
to in the Report, but its insulation resistance was tested and found

to be 8000 megohms. Both the coils ^ 100 and 101 show slight
traces of the green colour ; their insulation, however, is remarkably
high. It would seem, then, that it is very necessary to avoid the
use of newly made coils in important researches, and to keep a
careful check on any secular changes by means of repeated

comparisons. I hope when the permanence of |^ 101 has been
certainly established to remove the paraffin and see if there is any
change in the coil visible to the eye which could account for this
fall in resistance.

The two lO'ohm coils ^ 102, 103 have also been compared
with the 1-ohm in the manner described in the Reports, and the
values are given in the tables below. These coils are stated by
Messrs Elliott Bros, to be made of " the same wire of platinum
silver *015 of an inch diameter and 3'52 metres long."

n rt
mambOOmptrnC,

W.rt^/i\ 4:i,;\iviOi>ii>.t. AftiVn'A

SiW^!» I

1 *

-.4 -

-4-4--T4-4--l.i-..

' III

~^-: --^^ " ^-p- i-^-i --^-^-4-,M-^-'4-t4^J

I ! 1 I I 1

.- U -K. 4 4. 4. J _i^ J ^ .i

*-«....-♦, ^ -^ V * ^- ♦ ♦■ ^- I

t 1 L :_ ,

»^ei?.

. — t ■»«- i--i— 4

^^^^,C

J ' I '

'^^i'. At^<li(A^-ft !.'<^\ £

■ ^•■^>ii

FOB SLECTBICAL MEASUREMENTS

387

Resistance of CM, ^ 102.

Date

Temperature

Value

July 1885

March 1886

» 11

^»" "
Nov. 1885

» >»

16-8'

16-7'

16-6'

15.6'

11-9'

11-8'

8-2'

7-5'

6-5'

10-00210

10-00222

10-00129

1000103

9-99833

9-98830

9-98797

9-97711

9-97512

9-97250

Mean value

Temperature coefficient

9-990597 legal ohms at 1283%
•00289.

This is represented by the straight line (drawn thus — • —
on the diagram) ^ 102, Plate 9.

Resigtwnce of Coil, ^ 103.

Date Temperature

Value

July 1886
»» »>

n >i

Majch 1886

>♦ »

>> ♦♦
Nov. 1885

»i »
1 ♦» »i

16-9'

16-8'

16-65'

16-6'

15-6'

12'

11-8'

8-3'

7-7'

6-6'

10-00202

10-00197

10-00130

10-001 42

9-99815

9-98767

9-98692

9-97479

9-97315

9-96975

Mean value

Temperature coefficient

9-989714 at 12•88^
•00312.

This is represented by the second line (drawn black on the

diagram) ^ 103, Plate 9.

This difference between the temperature coefficients has been
checked by determining the difference between the coils at different
temperatures directly, and the results of the comparison are quite
satis&ctory.

B. A.

838 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

The proportional errors of the individual observations are
somewhat larger in this case than they were for the single ohms,
amounting in one or two cases to OOOG, or 6 in 100,000, but the
accordance is perhaps as good as can be expected. The point of
interest lies in the fiEtct that the temperature coefficients of the two
coils differ so considerably as '00289 and '00312 per 1"" C. although
made at the same time from the same wire.

Similar observations have been made on the coils of 100, 1000,
and 10,000 ohms, but their number is not yet sufficient for the
construction of the curve of variation with temperature. These
we hope to lay before the Association on some future occasion.

FOURTEENTH REPORT— MANCHESTER, 1887.

The Committee report that the work of testing resistance coils
has been continued at the Cavendish Laboratoiy.

Shortly after the Birmingham meeting of the Association the
Secretary received a letter from the Board of Trade enclosing a
copy of the general bases of a convention proposed by the French
Government for the consideration of the Powers, with the object
of carrying out the resolutions of the Paris Congress with regard
to electrical standards.

The convention stipulates that a legal character is to be given
to (1) the legal ohm; (2) the ampere; (3) the volt; (4) the
coulomb; (5) the farad.

It charges the Bureau International des Poids et Mesures,
established by the Metric Commission, with the construction and
conservation of the international prototypes of the standard of
electrical resistance, the comparison and verification of national
standards and secondary standards.

These questions had, at the request of some of the English
delegates to the Congress of 1883, been considered by the
Committee at the Birmingham meeting, and the following series
of resolutions, which the Secretary was instructed to forward to
the British Government, had been agreed to on the motion of
Sir William Thomson, seconded by Professor W. G. Adams: —

(1) To adopt for a term of ten years the legal ohm of the
Paris Congress as a legalised standard sufficiently near to the
absolute ohm for commercial purposes.

(2) That at the end of the ten years' period the legal ohm
should be defined to a closer approximation to the absolute ohm.

(3) That the resolutions of the Paris Congress with respect
to the ampere, the volt, the coulomb, and the farad be adopted.

(4) That the resistance standards belonging to the Committee
of the British Association on electrical standards now deposited
at the Cavendish Laboratory at Cambridge be accepted as the

22—2

340 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

English legal standards conformable to the adopted definition of
the Paris Congress.

In reply, therefore, to the letter of the Board of Trade, the
Secretary forwarded a copy of the above resolutions, with a
statement of some of the reasons which had led to their adoption
by the Committee.

During the year the original standards of the Association have
again been compared by the Secretary. An account of this
comparison and of the very complete one made in the years
1879-80-81 by Dr Fleming, the details of which have not been
published previously, will be given shortly.

At the last meeting of the Committee it was resolved, on the
motion of Mr W. H. Preece, seconded by Sir William Thomson, to
recommend the adoption of the Watt as the unit of power.

The Watt is defined to be the work done per second by the
ampere passing between two points between which the difierence
of electrical potential is one volt.

The Committee were also of opinion that it is highly desirable
to proceed with the construction of an air-condenser as a standard
of capacity, and for this purpose they desire to be reappointed,
with the addition of the name of Mr Thomas Gray and a grant
of £100.

FIFTEENTH REPORT— BATH, 1888.

In conformity with the opinion expressed by the Committee
in their last Report some steps have been taken towards the
construction of an air-condenser.

A meeting was held in London and Dr Alex. Muirhead
exhibited an air-condenser of capacity about *0035 mf. This con-
denser consists of a series of concentric cylinders of brass insulated
from each other by glass rods.

Dr Muirhead expressed his willingness to lend this condenser
to the Committee with two others of similar construction, and it
was agreed that the Secretary should be requested to test them
and to make a determination of their absolute capacity. Some
delay in sending the instruments to Cambridge unavoidably took
place, but the experiments requisite are now in progress; so far,
of the £80 granted last year for the purpose only '£2. 10^. has been
expended.

During the year the original resistance standards of the
Association have been compared with each other by the Secretary
and Mr T. C. Fitzpatrick ; an account of the experiments is given
in an appendix to the Report, together with a chart giving the
values of their resistance between 10° and 20"*. The general
result of the comparison is that with two exceptions the relative
values of the standards between the temperatures of 10** and 20"*
have not seriously changed since they were constructed in 1865
until June 1888. A change of about *0002 B. A. unit has been
observed in the coil F since the end of June 1888.

The attention of the Committee has been directed by several
practical electricians to the desirability of a redetermination of
the value of the specific resistance of copper. It is known that
copper wire is now made having a resistance 3 or 4 per cent, less
than Matthiessen's standard.

342 PRACTICAL STANDARDS

In view of the importance of copper to electricians the
Committee have undertaken to make experiments on the specific
resistance of copper, and wish to thank the various gentlemen,
who have brought the matter forward, for their offers of help.

At the last meeting of the Committee it was resolved, on the
motion of Mr W. H. Preece, to adopt the name " Therm " for the
Gramme- Water Degree Centigrade Unit of Heat.

Thus one "Therm" is the quantity of heat required to
raise one gramme of water at its maximum density one degree
Centigrade.

It was also agreed to adopt the name ''Joule" for 10^ C.O.S.
units of wort Thus a Joule is equal to 10^ ergs. It is the work
done in one second by the power of one Watt, or again the work
done when a current of one Ampfere flows for one second between
two points between which the difference of potential is one Volt,
and hence a power of one Watt is one Joule per second.

Hence, also, if we take the value of the mechanical equivalent
of heat as 4*2 x 10^ ergs, we have

1 Therm = 4*2 Joules.

In accordance with a suggestion made at the Manchester
meeting the value of the resistance of mercury in terms of
the B. A. unit has been again determined by the Secretary and
Mr Fitzpatrick*.

They find that a column of mercuiy 1 metre long 1 sq. mm.
in section has at 0'' C. a resistance of '95352 B. A. unit, and that
the value of the ohm in centimetres of mercury is 106'29.

The Committee are of opinion that they should be reappointed^
with the addition of the name of Mr T. C. Fitzpatrick, to continue
the experiments already referred to; they ask for a grant of
£100. They propose that Professor G. Carey Foster should be
the Chairman and Mr R. T. Glazebrook the Secretary.

♦ Proc, Royal Soc, ▼ol. xliv. ; Phil. Trans. 1888.

FOR ELECTRICAL MEASUREMENTS

343

Appendix.

On the Pei^manence of the Original Standards of Resistance of the
British Association and of other Standard CoHs. By R. T,
Glazebrook and T. C. Fitzpatrick*.

The original standards were compared together by Messrs
Matthiessen and Hockin in 1865 and 1867, by Messrs Chrystal
and Saunder in 1876, by Dr Fleming in 1879-81, and by the
Secretary and Mr Fitzpatrick in 1887-88. The details of Dr
Fleming s observations have never been published, and we have
to thank him for having placed his note-book and papers at our
disposal.

The question of the permanence of wire standards has been
discussed recently by Professor Himstedt, Wied. Ann. xxxi. p. 617,
and it seemed desirable to bring together all the information
attainable as to the original coils of the Association and others
used by Messrs Matthiessen and Hockin in 1867.

The original coils of the Association are six in number, and
the temperatures at which each has a resistance of 1 B.A.U,
are given by Mr Hockin in the Report for 1867. In addition to
these six coils Messrs Chrjrstal and Saunder examined the coil
No. 29, marked F by them, and also a coil known as Flat, which
are not mentioned in Mr Hockin's Report. The results of these
two comparisons are given in the following table : —

Table I. — Table giving the results of comparisons

in 1867 and 1876.

Material of
Coil

No. in

Mark in

Temperature at

1867

1876

which Coil is

Report

1 B.A.U., 1867

IB.A.U., 1870 1

Ptir

1
161

Ptir

16-8

15-8

AuAg

15-3

15-3 !

15-7

16-8

PtAg

not given

(19-4?)

PtAg

15-2

18-2

* See also Report for 1908.

t In obtaining this oolamn it was assnmed that B remained nnohanged between
1667 and 1876.

344 PRACTICAL STANDARDS

It will be noticed that the coil 0 of Pt Ag is the only one for
which the table shows any marked alteration.

Now Matthiessen gives as the percentage increase of resistance
per 1° C. for Pt Ir the value '032. Our own experiments show it
to be lower than this, and the value found for G by Dr Fleming
after a most careful series of experiments is '0278. I can find no
record of the temperature at which Hockin actually worked. If
it were below 15" and the temperature at which the coil was right
was found by the use of the coefficient '032 the temperature so
found would be too low.

If we assume Hockin's measurements to have been made at
0** C. and take Fleming's value '028 for the coefficient we find the
temperature at which the coil was right to be 18'1°.

We have next to consider the very complete series of measure-
ments taken by Dr J. A. Fleming in 1879-81 ; the results of these
measurements were tabulated on a chart which has been kept
with these coils since that date. For the details of the experi-
ments we have to thank Dr Fleming, who placed at our disposal
his note-books. The principle of his observations was as follows.
If X, Y be two coils to be compared, one X, say, was kept at
0** C, while the temperature of Y was varied fix)m 0** to 20*. The
diflFerences between the resistance of X at 0* and Y at various
temperatures were measured by Carey Foster's method in terms
of the wire of the Fleming bridge. The values of this dilBFerence
were plotted as ordinates, the temperatures being the abscissae,
and thus a curve was obtained giving the variation of resistance
with temperature for the coil F. For the standard coil Flat this
curve is accurately a straight line. This coil was then kept at 0°,
and the temperature of X varied, and so on for all the coils.

Now, at the temperatures given in Table I., column 4, taken
from the 1867 Report the resistances of all the coils should be the
same. Fleming found that this was not quite strictly true. He
defines therefore as the Mean B. A. unit the mean of the values
of the coils at the temperatures at which they were originally
said to be equal. This value is shown on his diagram by a red
horizontal line.

For the coils of platinum silver alloy, which is now used for
standards, Fleming's results are accurately represented by straight
lines for the temperature curves. This, however, is not so strictly
the case for the coils A and B of platinum iridium alloy ; thus for

FOR ELBCTRICAL MEASUREMENTS

345

these two coils Fleming took observations in the neighbourhood
of 0°, 4®, 8*, 15°, and 21"*, numerous observations being made at
each temperature; the straight line on the chart joining the
means of the observations at 15'' and 21° passes considerably above
the observations at 0°, 4f°, and 8°. The same too is the case,
though in a less marked degree, for the platinum coils D and E,
In the chart as drawn by Fleming it has been assumed that the
temperature curves are straight lines, and these have been drawn
to represent all the observations as closely as possible, but the
differences are considerable.

If we draw curves instead of straight lines to represent
Flemings experiments these curves between 10° and 20° are in
all cases nearly straight, and the differences, at the two tempera-
tures, from Flat at 0° are given in bridge wire divisions in
Table II. #

Table II.

Temperature 10^

Temperature 20°

-88

205

-97

196

165

-280

338

-263

348

100

Flat

112

We could determine from this table the temperatures at
which the various coils are equal, and hence compare Flemings
results with those of previous observers ; it will be easier to do
this atler discussing our own observations.

During the past year and a half the coils have again been
examined by ourselves. We find that between the temperatures
of about 10° and 20° Centigrade the resistances of the coils,
including an eighth coil H (No. 6 of the Report of 1867), may be
represented by the formulae given in Table III.

In obtaining the table it has been assumed, in accordance
with the observations of Dr Fleming, confirmed by Lord Rayleigh
and ourselves, that the resistance of one division (about 1 mm.)
of our bridge wire at a temperature of 15° is '0000498 B.A.IT

846

PRACTICAL STANDARDS

The table gives in B. A. units the value of iJj — Flato, Rt being
the resistance of the coils in order at temperature f, Flato that of
Flat at 0°.

The results of these observations are given in the chart*,
Plate 10.

The vertical divisions are ten bridge divisions, and the hori-
zontal divisions 0'2° C. In the original chart, which is retained
with the standards, the vertical divisions were one bridge division,
or -0000498 B.A.U.

About eleven observations on each coil are recorded in the
chart, and in but few cases is the error between observation and
the corresponding straight line greater than that which would

Table III.

Coil

A
B

C
D
E
F
G
H
Flat

J?j-Hat^inB.A.U.

- -00386 +
-00431 +
+ •00057 +
-■01434 +
-•013;)0 +
+ O0227 +
+ •00192 +
+ •00202 +
+ •00279 +

•001426
•001436
■000710
•003078
•003023
•000286
•000274
•000281
•000279

(^-10)
(^-10)
(if -10)
(^-10)

(f-10)

(r-10)

(^-10)
{t-\0)
[t-\0)

arise from an error of one-tenth of a degree Centigrade in the
temperature of the coil.

If as above we adopt as the Mean B. A. unit the mean of the
value of the coils at the temperature at which each was said to be
originally correct we find that this mean lies on our chart at a
distance of 78"3 divisions above the value of Flat at 0°, so that

Flat at 0° = 1 - -00390 B. AU.

The value given on Fleming's chart is Flat at 0' = 1 - '00410 B.A,U.,
and the difference is within the errors of reading on his chart.

We have thus the data for finding the resistance of any of the
coils in Mean B.A.U. at any temperature between 10*" and 20°.

It remains to compare these results and those of previous

* KoTE. The smaller letters in brackets after some of the observations on th»
chart sive the initials of the obseryer.

>T' -^ **c*x r *" "'^^ ^ ' I "" ■ • " ^ * ~~ — '- » ■ - •^■•

. t

*■ 1

( '

I . ■

*- I

-r —

* 1 , ^
|- * • * ( ^ * .•

I
• 4 ^

f-J. . L

■•H

i. . .

\*>

I : i .
• ' * \ I ' ' ■ i

•—} •~i *-- ; •-' •> t- - , ---— * -'- - ♦- —

• ' 1 \ i ■ '

. , t * . * i _.\| .♦ J,, ,-4 ...»

"-^N-'J

-4 '4 : w 1-4-

4 • -f ^-^

* -f

. ► i '"^ >^. •

♦ '

' • f
4

. -4
i

- i -

i. * . • ^. • ^ "v : ■ * * f

« -r

•t

' 1 t

. i

• • • t i

- * «
I .'

> a
• t

» V- ♦ «

» ♦

1- t

1 • .

t I

I ' f ■ *

I ♦ « •. - • t^—^ •- ^ » i 4 « .^

• t ^
1- »

4 ,. . .- .-44

I ♦ -*- n — •- f- 4- r > • .i —

• f • ^ ♦■ f ♦ t -♦-

*■ r •'*♦*♦ <

• I t -h ♦- j • ^ » ; - t t |-|' ''i

i. i -

I T

f ' >

» t • - t '

''■^ ■ »i- y««.> I >■■ *■■■ «■ «

* r

- ♦

■r A

i ,1 :

"t • r ♦ .— |- M-^ t ^ "?■
-♦ ' '. i ■*- * • r4 -r -^ ♦ T t t
' ■ ' 'I

r I -♦-♦-'• • — I ■♦

t f-f ^-f t '

• i

*--4-

«^ * t 4

-- ♦

» 4-*-»'- » i r

*- V — ♦ r »

-* t L -♦ ^ j

^ . i 4 .

- f j • ♦ .

. -»- 4- * "^ -t

" ■• f *

t- ♦ ■ -

4- ,.^^

. *

« 1

• ♦ t
" * - «

-■ » .

«- V •*- t

\ T

F-

r '■

t
* -

1 .

• -f • I- -r .».-*-.. ^ . I _;

1 . 1 . . I !

- \ • f ♦-,-i ^* .«-»-4.,-..4-t-

•*-rt ,.-» -^.4»

I ♦ I

T '

-». <^m m •*

t -*

f t

1 '-

♦ • H » 4

■ -T

I I
-. i . .

t r

• r

• » -

»i ^ *'\ »< SK/r <.r, ,1. \/\\\ ,\^

I' GJiAC:^ATS .A .8 3HT 10 J0KAT2l33a 3HT aWIWOHC TflAHO '

£m O^A V8^r M

r
>

•f •♦ t •

I} • ; i • ■ .

.*» V ^. ■^'

* ' t

4- t- ■»

*♦■ ? — f "»

I '

I ...

;. M ! I

— t' *; ''

( 1.. 4

It- i

f - »■

J ..♦4 . ; ,. -

it-' ♦ t

! 1 t

t"" ' "t "1 ■ " •* ■ ■♦

■trt -I-

' i-4 . i

♦ - : '

.. I

• t 1

•■r\

< I

r ♦

* 'I

' I

— i- < • • •

••■♦■• r ♦ «

-! -^ -

I r

-^, . ..

-+ - --.

t <

T *■ ; • -

. t.

■\, ■

a *•- < 4

: 1

< • 4 \ » « ♦ *

— » ■« » ♦ f

4 I . . , ^

\. 4 •

I I I ' •

'f-f r r

i *-" ■

I I

'- r 1 •■-•--•.■♦' •■♦--:• * \ ' r : ■> -t- A- 4 — J- * *-4-^ ^. ^-^ j_«. -^ . ^ .

,. . ; ... ^ ■ j- I — I - ; \ V • , ♦ f I ..».. J -».'.* ^ I ♦- i t T -

-' - ♦ , -I

^ , * \. , J. t .4- 4-.t

1 ^

t I

4-j4 i »•-* ' 4--l-i ♦ k i-

s^-,- ^ -^4 -»-4 k. *-^^- + -. f ^j—

! ; i ; ■ ... •• ^ ^ ^ ' • I T 1 ^

i- ►., :t-44-^- -.-♦I -*4* A"ji»i\. \il|.j, .^^+l+.^'-,j
I ' I 1 ' . . 1 I '•• ' . ■ '

-*• * •- - - f - - -^ - I— f ~^~ * — ^ -m^ ^ ~ .mr ■■> - ^ -^ 'h"* -^——4- -4 il^ -». A-- ^ . i^ —

I r . I I '^ \ I

t- f -t -T * *- ' — • -f '* ' ^ 1 V- :■•♦•■

j I , • I ■ \ ' '

" I : \ \ «

._! .; . -. . . L 4 I. 1 .; « ;\, \% .

■ 1 ■ .
» « - *

■* 4.

■ •• -1- t i «- t * )

*•%■• \

4 4 » t * J-

^s\\ *.vj./.'^.^v% A.v.^v.\ /'VisV %^^viv«iV'5sV/V'\\c^v\^.'V^:ii^\\^\\y^5^

FOB ELECTRICAL MEASUREMENTS

347

observers. We will take Fleming's observations first, and for this
purpose have given in Table IV. the differences in bridge wire
divisions between the coils at temperatures of 10** and 20** and
Flat at 0**. For the sake of comparison Table 11. is repeated.

A comparison of the corresponding columns shows that the
differences except possibly in the case of A and B at the lower
temperature are probably not greater than the error of experiment.
It must be remembered that A and B change by 28 bridge
divisions for 1° Centigrade, while for D and E the change is about
60 divisions per degree; the temperature of the coils is hardly
certain to O'l** Centigrade and the differences are within that
error. As to the platinum silver coils it would seem possible that

Table IV.

Value of R

- Flat^ at 10°

' Value otR-

Flat^ at 20°

; CoU

1880

1888

1880

1688

-88

-77-6

205

209

-97

-86-5

196

202

11-5

156

164

-280

-288

338

330

-263

-267

348

340

45-5

100

103

38-6

93-5

—

40-5

Flat

112

F* has risen relatively to Flat by -0001 B.A.U. and that G has
£Bkllen by '00005; but these differences are almost too small to
make certain of With regard to the results for A and B at 10°
it may be remarked that Fleming's line for these coils is more
curved than for any of the others, and that his observations at
6*9° and at 3° lie distinctly above the line which seems to represent
best the observations at 0°, 9°, 15°, and higher temperatures. The
observations are not at sufficiently close intervals of temperature
to enable the curved line to be drawn with accuracy, and it is
clear when plotting them that the curve near 10° may be wrong
by as much as 5 or 6 bridge wire divisions.

We would conclude then that there is no certain evidence for

* The results of other experiments oonfixm this rise in the valae of F,

348

PRACTICAL STANDARDS

a change in the coils in the interval 1880 to 1888. A comparison
with the results of Hockin and Chrystal is not quite so easy. It
is clear from the chart that the coils are not exactly equal at the
temperatures originally stated, and any table of temperatures at
which they may be said to have the value of 1 B. A. unit will
depend on the assumption made as to possible changes in any of
the coils. Chrystal in 1876 found that the coils B and C were
equal at the temperatures at which they were originally stated to
be each 1 B. A. unit. He supposed these coils had not altered and
found on that assumption a table of standard temperatures which
agrees well with that of Hockin except for the coil G. According
to our observations the coils now marked as B and C are no longer
equal at the temperatures mentioned. We find, however, that D,
E, and 0 are practically equal at the temperatures given by
Chrystal, and if we suppose 0 has not altered we get the following
table of standard temperatures : —

Table V.

Coil

Standard Temperature,

1876

1888

16-1

15-7

15-8

153

161

15-8

19-4?

16-9

18-4

17-9

Flat

15-2

The change in (7, as shown in this table, is not large, probably
hardly greater than would be accounted for by experimental error,
while 2), E, and Q agree very closely.

The diflferences in the case of A, B, and F are important. To
take F first. It is a platinum silver coil. No. 29 of the original
report. Its temperature is, however, not given in the Report for
1867, p. 211, nor is it marked on the coil itself Chrystal says
that it was used in some of the experiments " because its variation
coeflScient was small, but otherwise we have not bestowed much
attention on it." In his first table he states that the results
given for F came from a single experiment, and he gives as its

FOR ELECTRICAL MEASUREMENTS 349

variation coefficient per l"" Centigrade 28 divisions of his bridge,
while Flat and &, also platinum silver coils, have coefficients of 34
and 35 divisions. Now the observations of Fleming and ourselves
show that without any doubt these coils Flat, F, and 0 have
practically the same coefficient, viz., *00028 B.A.U. per 1° C.
Taking Chrystars bridge wire as '075 ohm as stated by him, his
value for Flat and O comes to •00026 B.A.U., which is in fisdr
agreement; while for F we find '00021, a value which is now
undoubtedly too low. We must infer either that the value of F
has changed considerably or that there is some accidental error
in the one observation given in Chrystal's table. The change
necessary to account for the temperature difference recorded in
Table V. would be an increase in resistance of -00067 B.A.U.

Let us now examine the numbers for A and B. It will be
seen at once that they have altered appreciably, having, in fact,
just changed places. Their temperature coefficients are nearly the
same, and there is no doubt that throughout Chrystal's observations
the coil he called B was slightly higher in value than A^ while
throughout the observations of Fleming and ourselves the reverse
has been the case. The question naturally arises, have the coils
been interchanged ? Chrystal (B. A. Report, 1876, p. 1 7) states that,
though they have no proper labels, they are marked in some way
or other so as to be identifiable. At the present time they have
brass labels screwed on to the ebonite of the frame bearing the
stamp BAl. 76 A and B.A. 76 B respectively. These were placed
on at the time of Chrystal's observations, and there seems just the
possibility of an accidental interchange.

The coil H, No. 6, of the original Report is marked as correct
at 15'3°. It is now correct in the sense used above at 17*9®, and
here again we have apparently a large change. The resistance
would appear to have gone down by about '00070 B.A.U. in the
twenty-two years which have elapsed since it was made. This
corresponds closely to the change in 0 observed by Chrystal
between 1867 and 1876. Now we know that 0 has not changed
relatively to (7, D, and E, since 1876 — unfortunately H was not
examined by Chrystal — and we are led to ask whether the change
was a real one, or due in some way to the observations. The
suggestion already made in case of 0 applies again. The tem-
perature coefficient used by Matthiessen and Hockin is certainly
too high, -00032 instead of 00028. If his observations on the

350

PRACTICAL STANDARDS

platinum silver coils were made at low temperatures, and then
the value of the temperature at which the coil is correct was
found by the use of the temperature coefficient, the result would
be too low. It will be seen shortly that all the platinum silver
coils examined, not merely those already mentioned, appear to
have fallen appreciably in value relative to the others.

But we have another method of comparing the results.
Chrystal has given a table of the differences at lO"" between each
of the coils and Flat. Now we have seen reason to believe that
there is not much change in C, D, Ey and 0, Let us find from
Chrystars table the value of the difference between Q and the
various coils at 10^ and compare these with our results. In doing
this some uncertainty is introduced from the fact that the value
of the bridge wire in ChrystaFs observations was only determined
approximately as *075 B.A.U. In this way we get the following
Table VI.

Table VI.

Value of a - Z at 10° in B.A.U.

Differenoe

1876

1888

•00693

•00583

•00110

•00648

•00635

•00013

•00162

•00135

•00027

•01596

•01632

-•00036

•01506

•01527

-00021

-•00018

- •oooss

•00017

On examining these differences it would seem that A has
changed greatly, while B has remained unaltered. This is not in
accordance with the conclusions derived from Table V., and will
require further consideration. With regard to the other four coils,
the differences are almost within errors of observation and are
in fair agreement with Table V. Coil C appears to have risen
relative to 6 by '00027 ; thus, since its temperature coefficient is
'00071, this would correspond to an apparent fall in the temperature
at which it is right of about 0'3° Centigrade. Table V. shows that
there has been a fall in this temperature of 0*2''.

The temperature coefficients of D and E are about '00308, so
that the differences recorded for these coils would be accounted

FOB ELECTBICAL MEASUREMENTS 351

for by an error of O'V in the temperature, while the change in F
relative to & is so small as to be within the experimental errors.
We are thus led to infer that, while C may have risen slightly,
the others have not changed by any but a very small amount.
This conclusion as regards i^ is at variance with the one derived
from Table V. In fiwt, while at 10*" F is above 0 in value ; owing
to the small temperature coefficient used by Chrystal for F, its
curve of resistance crosses that of 0, and at temperatures near IS'',
at which 0 is about right, F is considerably below it.

If we take Chrystars value for F at 10"* and the temperature
coefficient 00026 instead of -00021 used by him, we find that F
would be right at about 17*6'' instead of at 19*4'', as given by
Chrystal. This is much closer to 16*9°, the value given by our
observations ; if we take it instead of the 19*4'' of Table I., the
results of this Table VI. and of Table V. would point to a rise in
the value of F of about -00017.

The conclusion then that would seem to follow from a com-
p£u4son of these two series of observations in 1876 and 1888 would
seem to be that, while considerable uncertainty attaches to the
coils A and B, changes in the other five coils, C, D, E, F, and
0, if they have occurred at all, are probably not so great as
•0002 B. A. unit. G and F may possibly have risen by this
amount, while jD, E, and 0 have not varied at all.

Professor ChrystaFs observations in 1876 are in accordance
with those of Messrs Matthiessen and Hockin in 1865 and 1867,
while the results of Dr Fleming's work in 1880 agree, as we have
seen, with our own at the present date.

The observations recorded and discussed above were made
mostly at temperatures between 10° and 20". A considerable
number more were made during the cold weather in January and
February of the present year at temperatures near 0°, and we must
now consider them.

At these low temperatures the observations are not nearly so
concordant as those already considered. The terminals of the coils
are stout rods of copper, and whenever the temperature of the
room is different from that of the bath in which the coils are
placed heat is conducted to them through the copper rods and the
temperature becomes uncertain; besides this it is difficult to
prevent the deposition of moisture on the paraffin with which
the cases are filled, and this again becomes a source of error

352

PRACTICAL STANDARDS

Table VII. gives a series of the differences observed between the
various coils and Flat. The coils were in a north room of which
the windows were open, and the temperature in the room was on
the average about 2" C. The differences are given in bridge wire
divisions.

Now from Tables UI. or IV. we can easily calculate what these
differences ought to be if we suppose that the temperature curves
are straight lines. In making a comparison of the results of this
calculation with the observed values given in Table VII. some
allowance must be made for the fact that the bridge wire referred
to in Table IV. was at a mean temperature of about 15^ while in
Table VII. the temperature was about 2*". Now the temperature
coefficient of the bridge wire — platinum iridium — ^is about *00143 ;

Table VII.

Feb. 22,

Feb. 24,

Feb. 25,

Coil

Mom-

After-

Feb. 23

After-

Morn-

After-

March

Mean

mjg

noon

ing

noon

433-9

422-4

424-4

428

422-7

422-6

425-9

459-9

452-4

454-7

465-8

452-0

451-5

454-7

136

—

134-6

135-5

134-3

913-8

922-4

927-4

921-0

—

921-7

894-2

895-4

905-8

907 0

—

901-1

6-4

5-3

7-5

7*8

9-6

7-3 :

14-1

16-8

16-7

18-4

16-5

15-7

16-7

15-4

16-1

thus the change in resistance for IS'' of temperature will be '0185
of the resistance at 2^, and we shall have to reduce each of our
observed values by this fraction of itself

We thus get the following Table VIII. of values of the difference
at O'' between Flat and the various coils.

On examining these it is at once clear that the supposition that
the temperature curves for A and B are straight lines is &lse.

The other coils, with perhaps the exception of F, would lie at
0° on the straight line which represents the observations between
10'' and 20° within the limits of the errors of experiments.

The numbers given in Table VIII. agree well with those found
by Fleming in 1880 with the exception of the coils A and B.

Some observations made at intermediate temperatures are in
agreement with the statements just made. Thus on March 2, the

FOR ELECTRICAL MEASUREMENTS

353

temperature of the room being 12°, we found that at 4'9® C. the
difference between A and Flat at 0° was 266 bridge wire divisions,
while for B at 4'8** the difference was 280 bridge wire divisions.

Thus in conclusion we infer that while the observations in
1880 and 1888 are in close accord for temperatures between 10**
and 20** there is a discrepancy between them at lower temperatures
for the two coils of platinum iridium A and B* The other coils,
however, do not show any marked evidence of change. For the
same two coils there is a discrepancy between our results and
those of Chrystal in 1876 and Hockin in 1867. For the other
coils the agreement between Chrystal and ourselves is as close as
can well be expected, and our results as well as those of Chrystal
agree with Hockin's for the gold silver coil C and the platinum

Table VIII.

Coil

Obserred Value of

Flat -X corrected

for temperature

of bridge

Value of Flat - X

at 0° obtained

from Table HI.

Difference

A
B
C
D
E
F
Q
H

4177

440-1

131-8

904-2

884-0

7-2

16-2

15-8

364
375
131
906
874

16-5

53-7

711

0-8

- 1-8
-10

- 4-8

- 0-3
0-2

coils D and E, According to both Chrystal and ourselves the
platinum silver coils have &llen in value relatively to the others
by something like '0006 B.A.U., corresponding to change in the
temperature at which they are correct of some 2** Centigrade. We
have seen, however, that (?, the only one of these coils which was
carefully examined by Chrystal in 1876, has not altered since.
In its case the whole fall, if it occurred at all, took place between
1867 and 1876, and we suggest that possibly the fall has not been
a real one, but merely apparent, owing to the use of the wrong
temperature coefficient by Hockin.

As has been said already, the value that has been assumed as
the Mean B. A. unit since Fleming constructed his chart in 1876
is the mean of the values of the six coils A, B, C, D, E, and Q
mentioned in the Report for 1867 at the temperatures at which

& A

854

PRACTICAL STANDARDS

they were then said to be correct. In terms of this unit we find
(Table IX.) the present values at the temperatures of 1867. We
also give in the last column but one the temperatures at which
these coils have the value 1 B. A. unit, and in the final column the
temperature coefficients per l"" Centigrade also in B. A. units.

There remains now for consideration the results of comparisons
which we have made on various other standard coils originally
issued by the Committee, and which have most kindly been put
at our disposal by their owners for the purposes of the Report.

Table IX.—Oiving the Values in 1888.

Coil

Original No. See
Report, 1867

Material

Temperature at
whioh ooil is
correct, 1867

Value of coil in

Mean B.A. units

at the temperature

gi?en in 1867

Temperature at

which coil is
1B.A. unit, 1888

Temperature

CoeflScient in

B.A. units

Ptir

16*

1-00076

16-4'*

•00143

4>tlr

15-8*

1-00010

15-7°

O0144

Au Ag

15-3'

1-00050

u-s**

-00071

16-7°

-99930

16-9°

•00308

Pfc

15-7*

100000

15-7'

-00302

PtAg

—

16-7*

-00028

PtAg

15-2°

-99940

17-3*'

■00028

PtAg

15-3'

16-8"

•00028

Flat

PtAg

^■~"

140'

O0028

Messrs Elliott Bros, have three coils. One, No. 41 of the
original set, was made by Matthiessen in 1864. A second, No. 56,
was first examined by Lord Bayleigh in 1882: these two are
B. A. units, while the third, Elliott, No. 117, is a legal ohm, first
tested by R. T. G. in 1884. These coils are all of platinum silver,
with a temperature coefficient of •00028. Table X. gives the
temperatures at which they were found correct at different dates.

The observations made in 1887 are separated from the others
by a double line because during 1886 it was observed that the
paraffin used in the insulation was becoming green, and it was
therefore removed and replaced by pure ozokerit. In consequence
of this some change may easily have taken place in the coils, and
the record after 1884 must be treated as a firesh one.

* This coil is not mentioned in the Report of 1867. The details given are from
the label.

FOR ELBCTBICAL MEASUREMENTS

355

In the first coil the most noticeable point is the drop of 2^
between 1864 and 1879 ; but since this drop is followed by a rise
of 1'' in the next twelve months one may feel uncertain as to
whether it is real or due to some error in 1879.

In the next five years there appears to be a gradual rise
in temperature corresponding to a £bl11 in resistance; the total
amount would correspond to a change in resistance of about
*0004 B. A. unit. The removal of the paraffin has seriously
affected No. 41.

Table X.

Coil and Mark on it
at present

No. 41 ^ No. 55
No. 56 ^ No. 56
No. 117 ^ Na 63

si
as

15-2

■i

8
H

13-2

'Si
n

14-2

14-5
141

•8

15-5

15-4

17-8

6-2

— 14-7

— 16-8

The next coil, also of platinum silver, is one belonging to

Professor Carey Foster. He writes as follows : — ** It professed to

be equal to

1 B. A. unit at 14-2' C.

I had it direct fi-om Matthiessen, who« I believe, adjusted it
specially for me from his standards." On comparing it with F
in May 1887 we find that it has a resistance of '99983 Mean
B. A. unit at 16•2^ It would therefore be right at 16*8.

This, of course, shows a considerable change, corresponding
apparently to a fall in its resistance of about *00073 B.A. unit.
It will be noticed that this fall is just about the same as that
observed in the platinum silver units of this Committee — F, 0,
and H, We shall refer to it again in connection with the next
series of observations.

But by far the most important series of coils are a set
belonging to Mr H. A. Taylor. With regard to them he writes : —
** Most of my coils belonged to Hockin long before I knew him,
and at his death they were given to me by his £ftther." ''The

23—2

356

PRACTICAL STANDARDS

early history of these coils is lost, unless it can be found in
Matthiessen's note-books. I am informed, however, that the one

unit coil I sent you both last year and this, ^ No. 68, was copied
by Hockin from the B. A. coils you now have at Cambridge at
the time when he had regular access to them. Whether from
a particular standard or from the mean of several, I do not know ;
but he considered it to be at 15*5° C. less than the B. A. unit by
'0003. I presume the Au Ag coils, Nos. 19 and 34, were verified
by Matthiessen and Hockin, as they have the formal B. A. stamp.
With regard to the tens, one, I think, belonged to Hockin and the
other was purchased by Messrs H. C. Forde and Fleeming Jenkin
of the Committee in the usual manner."

Table XI. — Assuming a Coil (Hockins Standard) tested by
Electrical Standards Committee ( '^ No, 68) to be, as stated
by Hockiny smaller than 1 B,A. unit by 3 7oo {three-hundredths
per cent) at 155^ C. The Table shows the Resistance in terms
of 1 B.A. unit of other Standards IS'S"* C, at the dates given.

1 (C. F. T.) copy called rightl
at 16*l^Centiffrade. J

1 (No. 19) B. A. coil issnedl
as right at 15*6* C. /

1 (No. 84) B. A. coil iasnedl
as right at UST 0. /

10 (C. P. T.) copy called right!
at 15-6* C. J

10 (No. 3) B.A. ooU issuedl
as right at (?) /

10 (No. 4) B. A. coil issued!
as right at 16-0° C. /

PtAg

Au Ag

AuAg

PtAg

2-i

2*6
6-5
6*9
2*6
8*1
81

•OQQCLl
WVW

1-OOOOT

lOHxns

10-0009

Sid

'98997

1-00014
10*0001
10*0021

:1S

lis

'99965
*999o0
1*00028
9*9992
10*0018
0*9995

agS

*9e9e4

*99979
1*00023
9*9991
10*0013

9<QQnQ

"it.

*99965
iW9D3
1*00016
0*9091

1*00020

10*0012 1 10-0011

tfiXkO/J Q'QQQI

Where the temperature of observation differs from Ib'b'' C. the reductions
to that temperature are made by the temperature coefficients given.

Table XL gives Mr H. A. Taylor's observations on his coils on
the assumption that Hockin's stcuidard has not changed.

The evidence of a change is very small. The observations
have lasted over 14 years. For the first coil there would seem

FOR ELECTRICAL MEASUREMENTS

367

possibly to have been a drop of about *0(X)1 between 1875 and
1879. The next coil may have risen by as much and &llen again,
while the third coil would seem to have risen by *00015. The
results for the 10-ohm coils are much the same. From the results
of six coils, some of platinum silver, some of gold silver, we conclude
that there is certain evidence of no change greater than 1 in 10,000
during the last fourteen years.

The next Table enables us to compare these coils with the
standards at Cambridge. It will be noticed at once that relatively
to the Cambridge standards the coils have all fallen.

Table XIL

Goila

Col. I

Col. II 1

Glaiebrook'8
Determination

Col. Ill

Col. IV

OoL V

Col. VI

Nominal Value
aa iMued

Coefficient
(Hockin's)

Temperature
Coefficient
(Tajlor's)

Prraent BeBist-
anoe of the Coils
at the Tempera-
tures Riven in
Col. I

No. 19

Cent.
I'OOOOO at 15*5*

I'OOOOO at IS'8*

-99970 at 15'6*

10-00000 at 15*5*

Cent.
'99969 at 16*5*

100030 at 16T
( '99986 at 16*6* (1888)
( *99885 at 18'25* (1887)

10*00809 at 18-r

Hun-
dredths

6*5
6*9

Per
cent,
perl*

6-97
7'4

Prom Prom
Cols. II Cols. II
and III and IV

'9989i -99889
'99968 *90963

No. 84

1H;^68 ...

10 No. 8«;^ 69

S'l
81

Assu
8'

2*8
. 2*8

med

-99698 -99899
-99900 *99906

9*99408

Let us take first Hockin's standard |^ 68. Using Taylor's
temperature coefficient we find as its present value — the mean of
the two given in the last column — at 15*5°, '99901. It has there-
fore fallen relatively to the Mean B. A. unit by '00069, practically
the same fall as that found for all the other platinum silver coils
examined. The coil C.F.T. (the first coil in Table XI.) Mrill also
clearly have fallen by the same amount. Similarly with the ten

unit platinum silver coil |^ 69, it has fallen from 10 to 9'9940,
or by '006, nearly the same percentage ; and since, according to
Table XI., the coils have not changed relatively to each other and
to the gold silver coils by more than one-sixth of this amount since
1874, there is some probability that the change, if it has taken
place at all, occurred between 1867 and 1874. It will be re-
membered that we arrived at a similar conclusion with regard

368 PRACTICAL STANDARDS

to G, The difference between the values of ^ 68, found by
myself in 1887 and 1882 as recorded in the two last lines of
Table XII., arises from the fact that in 1887 the coil was
compared with F, and in 1888 with Flat and 0, In making the
calculations it was assumed that the values of F, Flat, and Q
in terms of the Mean B. A. unit had remained unchanged
since Fleming's time. The results of our comparisons given in
Tables IV., V., etc., would, as has been said, point to a slight rise
in f of possibly as much as '0001, and this would reconcile the

two values for 1 H ^ 68. As regards the gold silver coils
Nos. 19 and 34, if we take the value as issued, the one has fallen
by 00111, the other by -00087. We must remember that the
temperature coefficients for these coils are much greater than for
the platinum silver coils.

If, however, we compare the values as issued with those found
by Taylor in 1875 — Table XI., column five — we find that while
No. 19 was then -99969 at 155, showing a fall of -00031, No. 34
was 1*00014, showing a rise of -00014. Since this date No. 19 has
fallen therefore by '0008, and No. 34 by -00051, and these numbers
are within the limits of error of the fall of '00065 found for the
platinum silver coils. We would infer then that while apparently
there was a serious change in these coils relatively to the platinum
silver standards between the date of issue and 1875, since that
date there has been no change. On referring to Mr Taylor's letter
on p. 355 it will be noticed that the history of these coils previous
to 1875 is uncertain; all that is known is that they have the
formal B.A. stamp, and it is stated in the Fourth Report of the
Committee, 1866, that all the coils issued are correct to -0001 at
the temperatures stated.

There is still another coil of some interest. This is now

marked J^ 54. It was made in accordance with the sugges-
tion of Chrystal in 1876, with a thermoelectric junction
attached. Fleming compared it with his standards in 1879
and 1880. In 1884 it was again compared by us and found
to have the value 99658 B.A.U. at 8*3°, with a temperature
coefficient of -000295. It was then sent to Professor Eohlrausch
at Wurzburg for comparison with some mercury units con-
structed by Strecker, and was returned by him at the end of his
experiments.

In 1888 it was again compared and found to be '99653 B.A.n.

FOR ELEGTRIGAL MEASUREMENTS 359

at 8-3°, with a coeflScient of -000290. It will be seen that the
change is '00005, which is within the temperature errors.

Thus we conclude, from this general account of the condition of
the coils at present, that with the exception of the platinum iridium
coils A and B there is no evidence of any change of as much as
•0001 B.A.U. since the years 1874 or 1876, but that all the platinpm
silver coils and the two gold silver coils belonging to Mr Taylor
changed apparently by about '0007 B.A.U. between the time of
their construction and the time at which they were examined by
Chrystal and by Taylor respectively. This change may of course be
a real one ; we incline, however, to suppose that it is apparent only,
and offer the following explanation, already several times referred to.

Hockin says in a note to his Table of Temperatures, British
Association Report, 1867, which gives the temperatures for the
standard coils of the Association : " The values given in the above
Table are deduced from the german-silver coil called B* used in
your Committee's experiments in 1864."

He does not seem to have compared among themselves the
standards of various materials, but to have referred each to B,
Now we are ignorant of the temperature at which the comparison
was made, but we know he used the coefiBcient '00032. This at
present is too high by '00004. If we suppose that Hockin made
his determinations with the coils in ice, then this error in the
temperature coefficient would lead him to a value for the coil at
15°, which would be too high by -0006.

Having once got a platinum silver coil supposed to be known,
it would be natural to use it as a standard rather than any of the
others, because of its low temperature coefficient, and the error
made in the original determination of G would thus be perpetuated.
This conclusion is borne out by the observations on Messrs Elliott's
coil No. 41, Table X. Its standard temperature fell apparently by
2" between the time of its issue by Matthiessen in 1864 and
Hockin's comparison in 1879, and then rose again between 1879
and 1882. This would be accounted for if we suppose that Hockin's
platinum silver standard was too low.

P.S. — November 1888. — Since the experiments detailed above
were completed a considerable change has taken place in F. It is
now almost exactly equal to Flat, that is, it has risen in value by
'00048 B. A. unit. Further investigations as to the cause of this
must be left till the next Report.

* Thifl is not the same as oar B.

SIXTEENTH EEPORT—
NEWCASTLE-UPON-TYNE, 1889.

Further steps have been taken towards the construction of
an air condenser. As stated in the last Report, Dr Alexander
Muirhead kindly placed at the disposal of the Committee, for
the purpose of experiment, three such condensers which he had
constructed. A series of tests of these condensers was carried
out by the Secretary, and laid before a meeting of the Committee
in London on April 15 th. It was then decided to adopt
Dr Muirhead's form of condenser for the new instruments of
the Committee, and two condensers, each having a capacity of
about '01 microfarad, have been ordered from the Cambridge
Scientific Instrument Company. It was hoped that these would
have been completed early this summer, but great difficulties have
been met with in obtaining the brass tubes required for their
construction, and, though well advanced, they are not yet finished.
A detailed description of their design is therefore left to the next
Report.

A second subject of investigation has been the specific re-
sistance of copper. During the year Mr T. C. Fitzpatrick has
made a large series of experiments to determine this, and the
Committee desire to thank cordially those manufacturers and
others who have given him assistance in this research. They
would specially mention the firms of Messrs Thomas Bolton and
Sons, of Cheadle, and Messrs Frederick Smith and Co., of Halifax.

Before publishing the results of this investigation, Mr Fitz-
patrick is desirous of experimenting on some copper which is
being prepared for him by chemical means — all which has been
used hitherto has been electrically deposited — and of attempting
still further to purify some of the copper already in his possession.

Two members of the Committee, Sir William Thomson and
Mr Preece, were present at the recent Electrical Congress in
Paris. They report that the following resolutions, several of

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 361

which have already been agreed to by the Committee, were
unanimously adopted.

(1) L'unit^ pratique de travail est le joule. II est ^gal k 10^
unites C.G.S. de travail. C'est I'^nergie d^pens^e pendant une
seconde par un ampere dans un ohm.

(2) L'unit^ pratique de puissance est le watt. II est ^gal k
10^ unit^ C.G.S. de puissance. Le watt est ^gal k un joule par
seconde.

Dans la pratique industrielle, on exprimera la puissance des
machines en kilowatts, au lieu de Texprimer en chevaux-vapeur.

(3) Four ^valuer Tintensit^ d'une lampe en bougies, on
prendra comme unit^ pratique, sous le nom de bougie decimale*,
la vingti^me partie de Tetalon absolu de lumi^re d^fini par la
Conference intemationale de 1884.

(4) L'unite pratique de coefficient d'induction est le quadrant.

1 quadrant = 10* centimetres.

(5) La p^riode d'un courant altematif est la dur^ d'une
oscillation complete.

(6) La frequence est le nombre de p^riodes par seconde.

(7) L'intensit^ moyenne est d^finie par la relation

Imcy = -^\ Idt

(8) L'intensite efficace est la racine carr^ du carr^ moyen de
rintensite du courant.

(9) La force 61ectromotrice efficace est la racine carr^e du
carr^e moyen de la force ^lectromotrice.

(10) La resistance apparente est le facteur par lequel il faut
multiplier Tintensit^ efficace pour avoir la force ^lectromotrice
efficace.

(11) Dans un accumulateur, la plaque positive est celle qui
est reliee au pdle positif de la machine pendant la charge, et qui
est le pdle positif pendant la d^harge.

(12) Le CJongr^s recommande comme moyen de determiner
le degr^ d'incandescence d'une lampe, la m^thode propos^e par
M. Crova.

Ces diverses propositions sont adoptees k Tunanimit^.

* La bougie d^imale ainsi d^flnie, se tronve dtre irte Bensiblement ^gale k la
boogie anglaise (Candle itandard) et aa dixi^me de la Carcel.

362 PBACnCAL STANDARDS

As an English equivalent of the above the Committee have
adopted the following resolutions, which they hope will meet with
general acceptance.

(1) The name of the practical unit of work shall be the Joule.
The Joule is equivalent to 10^ C.G.S. units of work. It is the
energy expended during 1 second by a current of 1 ampere when
traversing a resistance of 1 ohm.

(2) The name of the practical unit of power shall be the
Watt. The Watt is the rate of working of a machine performing
1 joule per 1 second. The power of a machine would naturally be
expressed in kilo- watts instead of in horse-power.

(3) The name of the practical unit of light intensity shall be
the Candle*. The Candle is equal to the twentieth part of the
absolute standard of light as defined by the International Conference
of 1884.

(4) The name of the practical unit of induction shall be the
" Quadrant." One Quadrant is equal to 10* centimetres.

(5) The " Period " of an alternating current is the duration of
a complete oscillation.

(6) The " Frequency " of an alternating current is the number
of complete oscillations per second.

(7) The "Mean Current" through a circuit is the time

1 1'^.
average of the current and is defined by mean current = jp I idt,

i being the current at each instant of the time T.

(8) The "EflFective Current" is the square root of the
time average of the square of the current. Thus, effective

current = ^ \ jp, | i^dt > .

(9) The "Effective Electromotive Force" is the square root
of the time average of the square of the electromotive force. Thus,

effective electromotive force= /<^l e^dt-, e being the actual

electromotive force at each instant of the time T,

* It will be seen that the Committee recommend the names ** Candle" and
** Impedance" as the equivalents for the French terms "Bougie d6cimale" and
*'B^8iBtance apparente" respectively. With regard to the latter, they are of
opinion that it is desirable to restrict the term '* Resistance" to actions purely
dissipative.

The candle is also very approximately equivalent to the English standard
candle and to one-tenth of the Carcel.

FOR ELECTRICAL MEASUREMENTS 363

(10) The "Impedance" is the factor by which the eflPective
current must be multiplied to give the effective electromotive
force. Thus, in the case of a circuit of resistance R ohms, -and
self-induction L quadrants, in which a simple harmonic electro-
motive force of frequency, n/27r is acting. Impedance = V^^ + i*n'}.

(11) In an Accumulator the positive pole is that which is
connected with the positive pole of the machine when charging,
and from which the current passes into the external circuit when
discharging.

Of the £100 voted to. the Committee last year, £75 has been
drawn from the treasurer ; £60 towards defraying the cost of the
air-condensers and £15 for some resistance coils and thermometers
required for testing.

The Committee are of opinion that they should be reappointed,
with the addition of the name of Prof. J. Viriamu Jones, and with
a grant of £50 to continue the experiments which are now in
progress.

They propose that Prof. G. Carey Foster should be the Chairman
and Mr R. T. Qlazebrook the Secretary.

SEVENTEENTH REPORT— LEEDS, 1890.

The original standards of the Association have agcun been
several times compared among themselves.

The results of the comparisons appear to show that while the
coils A, B, Gy D, E, and Flat have remained constant relative to
each other, the three platinum silver coils F, 0, and H have
changed.

The change in F was referred to at the end of the Report in
1888, and is now very large. The coil has increased in resistance
by about '0006 B. A. unit; 6r, on the other hand, has fallen by
about -0002 B. A. unit, and H by about '0001 unit. The evidence
for these various statements is given in an appendix to the Report
by the Secretary.

It is perhaps worth remark that in each case the change either
took place during the time that the coil was immersed in ice or
was found to have happened when the coil was next measured
after its removal from the ice.

The legal ohm coils have not varied relative to Flat.

The investigations into the resistance of copper have been
continued by Mr Fitzpatrick. The Committee desire again to
thank the gentlemen who have rendered him assistance.

Mr Fitzpatrick has examined various specimens of copper
supplied him as wire. He has also examined copper prepared
for him as pure by Messrs Sutton, as well as some which he
prepared himself electrolytically from carefully purified copper
sulphate. These last two specimens lead to practically the same
value as that obtained by Matthiessen for the specific resistance
of copper — viz., 1767 x 10~* B. A. units at 18° ; the specific gravity
of these specimens is about 8*90. Two wires supplied to him
have, however, a distinctly lower resistance: the value for one
being 1731 x 10-», and for the other 1724 x 10"^; a diflerence in
the one case of 2 and in the other of 2*4 per cent. The specific
gravity of the first of these wires is 8*940 and of the other 8*946,

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 365

Fig. 1.

aad Mr Fitzpatrick assigns the increased conductivity to increased
density rather than to greater purity.

Matthiessen gives his results for the resistance of copper at 0°.
The observations were, however, made mostly at a temperature
of 18** or 20°, and reduced to O*' by the use of a temperature
coefficient; so that the value at IS"* found from that at 0° by the
same coefficient will probably represent the result of Matthiessen's
work more accurately than the one he gives himself Various
other points of importance are discussed in Mr Fitzpatrick's
appendix. He hopes to be able to give the results for some
copper prepared by chemical
means by Mr Skinner and him-
self He has also made a number
of measurements on silver, but
these are not yet complete. i

Dr Muirhead and the Secre-
tary have both been working
independently at the construction
and measurement of a standard
air condenser.

Two such condensers have
been made for the Committee
by the Cambridge Scientific
Instrument Company, on a plan
suggested by Dr Muirhead, and

mentioned in the last Report. The capacity of each of these is
about *02 microfarad. Some slight alterations are required to one
of these, the other is completely satisfEustory. Its capacity has
been repeatedly found, and remains constant to at least within
1 in 2000 which is about the limit of accuracy attained. Its
insulation resistance is good, the loss by leakage being about
1 in 1000 of the total charge per 1 minute. It has been found
possible to compare readily with this standard various mica
condensers having capacities of 1, '5, '1, and '05 microfarad. The
accuracy of these determinations is about 1 in 2000. A full
account of the construction of the condensers and of the method
of making the various tests is given in an appendix by the
Secretary, while Dr Muirhead has contributed some notes on his
own condensers, and tests.

Another appendix contains an account of a very careful and

-.._v„ u

866 PRACTTICAL STANDARDS

interesting comparison between the standard mercury thermo-
meters of the Association and a platinum resistance thermometer
constructed by Mr E. H. QriflSths. The resistance thermometer was
graduated by means of Begnault's numbers for the vapour pressure
of water at various temperatures between 0"* and lOO"".

The curve of corrections obtained in this way is exactly parallel
to that given by the Eew comparisons; there is throughout the
range a constant difference of 0*02'' between them. This amount
is within the limits of error on the mercury thermometer.

The question of the best value to adopt for the dimensions of
a mercury column having a resistance of 1 ohm has been raised by
some members of the Committee during the year. There is no
doubt that the column of 106 centimetres adopted by the Paris
Conference in 1884 is too short.

After a discussion of the results of the most recent observations,
the following resolutions were adopted by the Committee : —

1. The Committee recommend for adoption as a standard of
resistance sufficiently near to the absolute ohm for practical
purposes the resistance of a column of mercury 106*3 cm. in
length 1 square mm. in section at a temperature of 0"" C.

2. That for the purpose of issuing practical standards of
resistance the number *9866 be adopted as the ratio of the B. A.
unit to the ohm.

Thus the new unit may be obtained from the B. A. unit by
increasing it in the ratio unity to '9866 ; or, to put it differently,
the specific resistance of mercury, in B. A. units, is taken as
'9535 X 10~*, and the length of a column of mercury which has a
resistance of 1 B. A. unit as 104*87 cm. The specific resistance of
mercury in ohms is *9407 x 10~*, while the ohm is 1*0136 B. A,
units.

In conclusion, the Committee wish to ask for reappointment,
to enable them to continue the work of constructing and issuing
standard instruments. Of the grant of £50 made at Newcastle
only £12. 17«. has been drawn. In order to check any further
change in the values of the B. A. units and to render it lees
necessary to employ the original standards in all the comparisons
which are made, it is desirable that the Committee should possess
three or four copies of the B. A. unit ; while, to enable comparisons
to be made between the new air condensers and condensers of

FOR ELECTRICAL MEASUREMENTS

367

capacity comparable with a microfarad, a resistance box going up
to several hundred thousand ohms is required.

The Committee are of opinion that they should be in a position
to purchase these resistances ; they therefore recommend that they
be reappointed, with a grant of £100, that Professor Carey Foster
be the Chairman and Mr B. T. Qlazebrook the Secretary.

Appendix I.

On the Values of certain Standard Resistance Coils,
By B. T. Glazebrook, F.RS,

The B, A, unit Standards.

The Standard B. A. units of the Association have during the
year been several times compared together both by the Secretary
and by Mr Fitzpatrick. Table I. gives the results of two sets of
comparisons made in August 1890 — the differences between the

Table I. Resistance of the B.A. Standards, August 1890.

Difference between each

Coil

Tempera-

ooil and Flat in bridge-
wire diTiflions

Difference
observed —

Change of
resistance

tare

calculated

per 1° in
b.w.d.

Observed

From chart,

Aug. 15, 1890

1888

17-2

27-8

33-0

-5-2

28-6

17-4

30-5

30-6

28-8

17-6

22*2

23-0

-0-8

14-2

i D

17-26

61-2

63-6

-2-3

61-7

; E

17-3

79-2

79-6

- -3

60-7

17-3

3*2

- 9-6

12-7

6.7

17-5

-220

-18-0

-40

6-6

17-4

-17-0
August 19

-15-0

-2-0

5-6

18-8

67-5

69-5

-2-0

28-6

17-8

eo-6

62-0

-1-4

28-8

19-2

31-6

36-0

4-4

14-2

18-8

146-7

16-1

6-3

61-7

19-0

170-6

17-3

2-4

60-7

18-9

2-9

- 9-6

12-4

6-7

19-0

-21-8

-18-0

-3-8

6-6

19-0

-17-7

-160

-2-7

6-6

368 PBACTIGAL STANDARDS

various coils and the platinum silver standard Flat are given in
the third column in bridge-wire divisions. One bridge-wire
division is very nearly '00005 B. A. unit.

In the fourth column are given the corresponding differences
obtained from the chart made in 1888. In the next column will
be found the differences between the observed values and those
given by the chart, while the sixth column gives the change in
resistance for 1** C. for the various coils. It will be seen that for
the first five coils the differences between observation and the
chart are such as would be readily accounted for by a small error
in the temperature, and we may say that there is no evidence of
a change in the resistance of these coils relative to Flat. This
conclusion is borne out by the results of a series of observations
made in January and February by Mr Fitzpatrick. But when
we come to the three platinum silver standards, F, 6, H, the
results are at once seen to be quite different. Thus F would
appear to have risen relatively to Flat by about 12'5 bridge-
wire divisions, while 0 and H have fallen by 4 and 2*5 divisions
respectively.

Since these are the most important standards, their temperature
coefficients being all very small, it was necessary to examine their
history with some care. A change in F had been noted in a
postscript to the Report for 1888. The general conclusions of that
Report were that up to the summer of 1888 there had been no
change in the value of the coils. It was shown that all the original
platinum silver coils examined then — those of Messrs Elliott,
H. A. Taylor, and others, as well as those belonging to the Com-
mittee— had apparently fallen in value relatively to the Mean
B. A. unit by about '0007 B. A. U. since 1687, but evidence was
adduced to show that the fall was only apparent, due to an
error in the temperature coefficient used at that date. A single
observation of Chrystal in 1876 pointed to the possibility of a
change in F, but that change was not confirmed by other evidence;
while so for as the platinum silver coils were concerned, the
observations of Dr Fleming in 1881, and myself in 1888, agreed
closely.

Since 1888, however, changes have shown themselves.

These are evidenced by the three following Tables II., III.,
and IV., which give the differences Flat— jF, Flat — 0 and Flat— if
respectively.

FOR ELECTRICAL MEASUREMENTS

369

Table II. Vaiue of Flat—F.

Date

Temperature

1
Valae

10-0

1
10-5 1

Chart 1888 \

16-0

9-5 j

20-0

14-8

8-5 1

May 16, 1888

9-0

July 2, „

3-0

II *'i i>

14-8

3-8

»» !•% «

14-2

4-2

»» 1^1 n

14-6

3-3

II * ^« II

14-7

3-3

11 28, „

16-7

, 4-2

1 Jan. 1890 ...

10-0

' -40

' Aiay „ ...

14-4

-3-5

Aug. „ ...

16-9

-3-2

[ II II

16-7

-30

Table III

. Value of Flat~G.

Date

Teinperatare

Value

(

10-0

17-6

Chart 1888 \

15-0

18-0

20O

18-5

July, 1888 ...

14-6

16-6

Jan. 27, 1890

10-0

16-9

II 29, „

4-5

16-7

Feb. 4, „

6-0

16-6

May 31, „

14-4

21-6

June 10, „

16-0

21-4

»» ll» 11

16-0

22-2

II 1*1 II

16-0

22-2

II 13, „

160

22-2

Aug. 9, „

19-0

21-8

,1 16, „

17-0

22*3

» 2^1 II

16-5

22-6

II 29, „

16-5

22-5

The first three lines in each table give the difFei-ences, at the
temperature shown, taken fix)m the chart drawn in 1888; the
remaining lines give the differences actually observed, with the
dates and temperatures. Thus, taking the various coils, it is clear

a A. 24

sVo

PRACTICAL STANDARDS

Table IV. Value of Flat—H.

Date

Temperature
10-0

Value

15-5

Chart 1888 \

15-0

15-6

20-0

16-6

July, 18>8 ...

14-6

141

Jan. 27, 1890

10-0

17-6

»» 29, „

4-5

17-6

Feb. 4, „

6-0

16-5

Maj 31, „

141

18-3

JuDe 10, „

16-0

18-1

»» ^ *•» »♦

16-0

17-7

»» 12, „

16-0

16-4

»» 1^> »

160

16-8

Aug. 9, „

19-0

17-7

11 1«^> i>

17-4

17-0

»» 28, „

17-0

17-8

n 29, „

16-4

18-2

that while up to May 1888 the difference between Flat and F
remained the same as shown by the chart and observations up
to that date, a change took place during the low temperature
observations in July 1888, while by the time the coils were again
examined in January 1890 a further change had manifested itself
This continued up to the present date, so that now at a temperature
of about 15'' the coil F has increased in resistance relatively to
Flat by about 12*7 bridge-wire divisions. This, assuming the
whole change to be in F, will correspond to a rise of resistance
of '00063 B. A. unit, or in other words the temperature at which
the coil is right has fallen by about 2*3^ In January 1890 the
coils were again exposed to a low temperature, and it seems
probable that the changes took place when the coils were in ice.

From the values in Table III., which gives the values of
Flat — 0 we see there is no evidence of change till May 1890.
The observations in July 1888 and January and February 1890
are quite in accordance with the chart, but in May 1890 it is clear
that 0 has fallen relatively to Flat.

The value of the difference at a temperature of 16° is
22*1 b.w.d. as against 18*1 given by the chart. Thus 0 has
fallen relatively to Flat by 4 b.w.d., or -0002 B.A. unit. This

FOR ELECTRICAL MEASUREMENTS 371

change was first observed after the coils had been exposed to a low
temperature.

With regard to H the change first showed itself during the
low temperature observations in January and February 1890, and
Table IV. indicates that the difference between Flat and H is
now 17'5 divisions as against 15*5 in 1888, or in other words, that

0 has fallen by -0001 B. A. unit. Also since Flat — F changed in
1888, while Flat— (? and Flat— F did not, we infer that the
change at that date was in F, not in Flat ; while since Flat — H
changed in January 1890 without a change in Flat — 0, it appears
that the change was in H, not in Flat; and finally, from the
observations in May 1890, which show a change in Flat — 0, but
never in Flat — H and Flat — Fy we infer a change in 0.

As to the cause of these changes, we can say but little. We
hope to investigate them more completely by the aid of the coils
lent by Mr H. A. Taylor and others, and referred to in the 1888
Report ; but it seems possible that they are due to strains set up
in the wire by the great contractions and expansions produced by
cooling and heating in the paraffin in which the coils are embedded.
The coil Flat is of a different shape to the others and little or no
paraffin has been used in its construction. The other coils, F, 0,
H, are embedded in paraffin in the usual way. On cooling down
to 0°, this shrinks greatly, and it is quite conceivable that this
shrinkage may have strained the coils and so caused the change.
We hope to test this by having coils made free from paraffin and
investigating with them the effects of repeated heating and cooling.
The fall of H and 0 would be accounted for by a loss of insulation
causing a slight leak either from the wire to the case or across the
surface of the paraffin. The insulation resistance for F, 0, H was
therefore testc^l and found in each case to be several thousand
megohms, while the surface of the paraffin which had become
dirty with time was scraped, but without producing any change
in the resistance. A leak, of course, would not produce the rise
found in F.

Observations of the coils at 0^ have always been unsatisfactory
and attended with considerable difficulty. This is mainly due,

1 believe, to the fact that the temperature of the room in which
the observations have been made has usually been above zero, and
that heat is conducted into the coils by the thick copper connecting
rods. It would seem possible, however, that part of the difficulty

24—2

372

PRACTICAL STANDARDS

(see Report of the Committee for 1888, Table VII., p. 352) may have
been due to real changes in the resistance arising from strains set
up by the cooling.

The Legal Ohm Standards.

The results of observations on the legal ohm standards of the
Association are given in the Report for 1886. Experiments made
on these between July 1884 and January 1886 showed that while

one coil, ^ 100, had retained its value unchanged, the other,
^ 101, had varied. These observations have been continued,
and the results are shown in the following tables, which give the
value of each coil as found by direct comparison with the standard
B. A. units, and its value as given by the chart in 1886.

Table V. Residts for ^ 100.

Date

Standftrd used

Tem-

Value

Value on

Difference

m compariRon

peratare

Chart

Feb. 1887

16-3

1 -00009

1-00008

•OOOOl

Nov. 1889

(}

15-8

•99997

•99996

•00001

})

14-8

•99971

•99968

•00003

))

16-0

•99998

1-00000

•00002

Dec. 1889

Flat

14-4

•99962

•99959

•00003

14-8

•99969

•99968

■00001

»»

)}

13-2

•99926

•99924

•00001

6-2

•99744

•99735

•00009

5-7

•99729

•99720

O0009

Table VI. Results for ^ 101.

Date

Standard used

• •

Tem-

Value

Value ou
CLart in

Difference

in comparison

perature

found

1885, 1886

Feb. 1887

16-3

•99970

•99930

•00040

Nov. 1889

15^9

•99965

■99920

■00035

161

•99932

•99899

•00033

»)

16-0

•99956

•99922

■00033

Dec. 1889

Flat

14-4

•99909

•99880

•00029

))

«>

16-0

•99925

•99897

•00028

13-3

•99h79

•99860

•00029

7-6

■99725

•99695

•00030

6-6

•99701

•99668

■00033

FOR ELECTRICAL MEASUREMENTS 373

These tables show three facts conclusively: (1) That up to
December 1889 no appreciable change had taken place in the
relative values of ^ 100— the Legal Ohm Standard— and Flat
or 0\ (2) that between January 1886 and February 1887 ^ 101,
which had varied previously, changed by about '0004 ohm; and
(3) that the greater part of that change has remained permanent
up to December 1889. At present the diflference between ^ 100
and ^ 101 is about "0004; in 1886 it was about 0008. The
agreement between the observations in November and December
1889 — in one set of which Flat was the standard of comparison,
while in the other 0 was used — show that the relative change in
G and Flat took place after this date.

Appendix II.

On the Air Condensers of the British Association,
By R. T. Qlazbbrook (with a Note by Dr A. Muirhead).

The question of issuing certificates of capacity has from time
to time been discussed by the Committee. The following paper
gives an account of some experiments that have been in progress
during the past two years with this object in view.

In the Report for 1887 the Committee express the opinion
that it is desirable to proceed with the construction of an air
condenser. In conformity with this opinion a meeting was held
in London, at which Dr A. Muirhead exhibited an air condenser
consisting of a series of cimcentric brass cylinders insulated by
glass rods, which appeared to the Committee to possess great
merits ; and it was decided that the Secretary should test this and
two similar condensers which Dr Muirhead offered to lend, before
proceeding further with the construction of condensers for the
Association. The tests were carried out with satisfactory results.

The capacity of each condenser was determined repeatedly,
using the method of a vibrating commutator, due to Maxwell,
already employed by J. J. Thomson, PhU. Trans, 1883, and
Glazebrook, PkiL Ma^. August 1884. The values found were : —

Oi » 0030514 microfarad.
Cg« -0031258
C,= 0033288

374

PRACTICAL STANDARDS

It was found that the ca{)acities remained constant from day
to day, and that the accuracy of a single determination was about
1 in 1000, although the capacity to be measured was so small

Some mica condensers belonging to the Cavendish Laboratory
were compared with these — details of the method will be given
shortly — and it was found that when comparing a condenser of
1 microfarad with the three air condensers combined, having thus
a capacity of '009506 microfarad, so that the ratio of the two was
about 100 to 1, an accuracy of about 1 in 1000 was attained.
It was also shown that the capacity of the mica condensers as
thus found differed by nearly 2 per cent, from their values as
determined by the rapid commutator, thus proving that the
commutator method was unsuitable for a condenser showing
absorption.

Thus for three mica condensers the following values were
found : —

With
commutator

Bj slow method
of comparison

•9690
•4934
•09543

•9868
■4994
•09644

These results make the necessity for an air standard all the
more apparent. A report on the experiments made up to that
date was laid before the Committee at a meeting in London in
April 1889. It was then decided to adopt Dr Muirhead's form of
condenser, and to have two made on the same pattern for the
Association. These have been constructed by the Cambridge
Scientific Instrument Company, following Dr Muirhead's plan, but
on an enlarged scale. Each has a capacity of about 02 microfarad,
or about six times that of one of the original condensere.

Fig. 2 shows the arrangement.

The condensers consist of twenty-four concentric tubes; the
outer tube is about 2 feet 9 inches high and 6 inches in diameter.
Each succeeding tube diminishes in diameter by half an inch ;
the tubes are about ^nd inch in thickness, and the air space
between the inside of one tube and the outside of the next is
about ^nd inch, but it was found impossible to get all the tubes

Fig. 3.

FOR ELICrSIOAL MBASOEKHENTS 876

of exactly the same thicknees, so that in some casee the distance
between the tubes is less than the abova These tuhes are carried
by two conical brass castings ; the outside sur&ce of each casting
forms a series of twelve steps,
over which the sacceesive tubes
fit. Each tube is held in
position by screws. The upper
cone is supported by the out-
side casing of the condenser,
and twelve of the tubes bang
vertically from it The lower
cone ia carried by three ebonite
pillare,about3inches in height;
the twelve tubes which are at-
tached to it come respectively
between those which are sus-
pended firom the upper cone.
Thus the insulation depends
on the ebonite pillars, assum-
ing there is no leakage across
the air fix)m the edges of the
tubes. There is an opening in
the outer casing, closed by a
door, by means of which the
ebonite can be cleaned ; the
whole is dried by placing in-
side a small vessel of sulphuric
acid. In the centre of the
upper cone there is a hole
through which a rod passes.
The rod is connected with the
lower cone, and forms the elec-
trode for the insulated cylinders.
An ebonite plug, fitting tightly
round the rod, can be pushed
down so aa to close the hole
and prevent the ingress of dust
when the condensers are not in
use ; when they are being used
the plug is removed.

376 PRACTICAL STANDARDS

The coDdensers are placed in the testing room at the Cavendish
Laboratory and covered by a wood and canvas case to protect them
from dust. It is not intended that they should be movable.

After this description of the condensers we will proceed to an
account of the tests to which they have been subject. The first
test was for leakage.

One set of cylinders was put to the earth while the other was
connected with a gold-leaf electroscope. An attempt was then
made to charge them with an electrophorus or a small electrical
machine, but this failed entirely. The electricity either sparked
across at places where the tubes were very close together, or,
before the potential rose sufficiently to affect the electroscope,
small fibres or dust particles which adhered to the tubes formed
leaks across ; it was clear that the condenser could not be charged
to the potential of the machine. Tests were then applied for
leakage when the potential was lower. One set of tubes was
connected to one pole of a battery — about thirty-six storage cells
were generally employed, having an e.m.f. of 75 volts — the other
set being in connexion with an insulated key ; the second pole of
the battery was connected through a galvanometer to the key
and the condenser charged. After an interval, usually about five
minutes, contact was again made at the key; the deflection of
the galvanometer needle — assuming the B.M.F. of the battery not
to have changed — was a measure of the quantity of electricity
which had leaked from the condensers in the five minutea

The amount of leakage was veiy different in the two con-
densers and depended greatly on the dryness of the air and
ebonite pillars. Thus on March 11, when strong acid had been
enclosed for some time, for condenser I. the leak per minute
amounted to about *1 per cent of the whole charge, while with
condenser II. it was about ten times as great.

The sulphuric acid was removed during the Easter vacation
and replaced by calcium chloride, and after this the leak in I. rose
to about 1 per cent, per minute or ten times its former value,
while that in II. was from 3 to 4 per cent of the charge. With
the calcium chloride inside the leak was never reduced to less
than about '8 per cent per minute. In August, the condensers
having been closed since June with calcium chloride, there was a
leak in L of about 3 per cent per minute, while in the same time
IL lost about 8 per cent of its charge.

FOR ELECTRICAL MEASUREMENTS

377

On August 14, immediately after this test, the calcium chloride
was replaced by sulphuric acid, and the leak was quickly reduced
to about 1 per cent, per minute for I. For II. no improvement
showed itself at once. The next day the leak in I. was about
*4 per cent, per minute ; that in II. had not been greatly reduced.
On August 16 the ebonite was therefore well cleaned, and air was
blown through the tubes of II. and the whole closed for about two
hours ; the leak had then fallen to about 2 per cent, per minute.
By August 18 the leaks were still more reduced, that in I.
being '2 per cent, per minute, while that in II. was '6 per cent,
per minute.

By the afternoon of this day, the upper parts of the condensers
having been open to the air of the laboratory for some six hours
during other tests, the leaks had appreciably increased, but they
had £Ekllen again the next day when the condensers were left closed
during the night.

Fig. 4,

Thus, during the observations in August, with the exception of
those on August 14, the condenser I. was losing its charge at the
rate of about ^^th part per minute, while the leakage in II.
was some five or six times as great, being about x^th part of the
charge per minute.

As will be seen later, several mica condensers were compared
with I. and II.; the leaks in them were all small, and did not
exceed more than ^th per minute.

We come now to the experiments for determining the capacities
of the two condensers. Of these, three independent series were
made, viz. in December 1889, May and June 1890, and August 1890.

378 FBACnCAL STANOABDB

The method ahettdy lefiened to was ined. Fig. 3 gives a
diagram of the method; in fig. 4 the ooimezioiis actually
emfdoyed are shown. With the notation employed, PkiL Mag.
Aagnst 1884, we have, if (7 be the capacity of the condenser, n the
number of times it is diazged per seoraid.

aC=

f, o^ ]

.( ab ) (^ . ag ]

In most of the experiments aboat to be described, we had the
following values in legal ohms : —

a = 10, d = 1000,

6 = 18, SF = 17,600.

while c, which was the adjustable arm, varied firom 6000 to
15,000.

With these values, the only correction which need be included
is the last bctor in the denominator, and we may write

wC=

The resistances were taken from a legal ohm box belonging

to the laboratory; the various coils in this box were carefully

compared with each other by Mr Searle, and found to be

consistent with each other, at any rate to within 1 in 10,000.

The coils were also compared with the standards of the Association,

and it was found that at 16^ they were greater than legal ohms in

the ratio of 1*0011 to 1. The standard temperature adopted in

the experiments was 17"", and since the coefficient of increase of

resistance of the box is about "0003 per 1"* C, the resistances

require to be multiplied by 1-0014, to reduce them to legal ohms.

In some cases, in the vaiue of c, coils from a B. A. unit box,

containing coils of ten, twenty, thirty, and forty thousand, B. A.

units were employed.

The values found for these coils by myself in terms of the legal
ohm box showed that they were very consistent with each other,
and that the nominal 10,000 B. A. unit was equal to 9880 legal
ohms as measured by the legal ohm box.

FOR ELECTRICAL MEASUREMENTS

379

In the comparisons of two condensers certain coils from a
megohm box were used; the value of each of these was also
determined. They were as follows: —

1 ... 98,731 Legal ohms of standard box.

98,626

98,698

98,735

98,725

98,776

In the experiments on Dr Muirhead's condensers, the vibrat-
ing commutator described in Professor Thomson's paper, Phil.
Trans, 1883, or in my paper, Phil, Mag, 1884, was used, with
complete success. In the experiments about to be described, this

Fig. 5.

was replaced by a rotating commutator which had been fitted up
by Professor Thomson and Mr Searle for their experiments on
the value of "v," and which possesses certain advantages over
the other form. Dr Muirhead and Dr Fleming have also used
a somewhat similar arrangement of apparatus. Fig. 5 shows the
arrangement. The split ring commutator is carried on the axle
HK, which is driven by a water motor. Two wire springs, Q, i2,
are in contact with the two halves of the commutator respectively,
and as it rotates, the brush P, made of very fine brass wire, is
brought into communication alternately with Q and R, The disc
LM was of iron, and its mass helped to steady the motion. On
one face of the disc a series of circles were drawn forming a

380 PRACTICAL STANDARDS

number of annuli. The successive annuli were divided each into
a different number of divisions by radial marks. Thus in the
innermost annulus there were four, on the next five, and so on.
The disc as it rotated was watched in the usual stroboscopic
manner through two slits on two -pieces of thin metal carried
by the prongs of a tuning-fork, which made about 64 vibrations
per second.

When the frequencies of the disc and of the fork were in
certain simple ratios to each other, the corresponding pattern on
the disc was seen in a steady position. The driving pulley of the
motor carried a second band, which passed over an idle pulley
near the observer at the tuning-fork, and the speed of the motor,
and hence of the disc, was adjusted partly by varying the flow of
water, partly by friction on this band, until the desired pattern
was seen in the steady position. This position was easily main-
tained by var3ang the friction on the string. The tuning-fork
drove a second fork an octave above itself in frequency. This fork
was mounted near the standard fork of the laboratory, and the
beats between the two were counted. The frequency of the
standard fork was determined by Professor Thomson and Mr Searle
for their experiments on "v," recently communicated to the Royal
Society. They found that it had changed slightly since it was
determined by Lord Rayleigh, and give as the result of their
experiments

Frequency at temperature f = 128* 105 {1 - (< - 16) -00011}.

The driven fork was always adjusted to a slightly lower frequency
than that of the standard, so that there were about 20 beats to
the minute between the two. During each series of observations
the beats were repeatedly counted, but they rarely varied during
the series sufficiently to affect the result. The commutator was
designed and partly constructed by Mr Searle, who observed at
the tuning-fork throughout. A little attention was required to
secure good contact between the springs Q, R and the rotating
parts, and also to adjust the brush P, but with moderate care in
the adjustments the apparatus worked perfectly.

The galvanometer was one constructed in the laboratory ; it
had a resistance of 17,600 ohms, with a long silk fibre suspension
— a quartz fibre would have been an improvement.

Its sensitiveness was such that 1 scale division corresponded

FOR ELECTRICAL MEASUREMENTS

381

to '83 X 10~" C.O.S. units of current ; the time of swing was
7*2 seconds, so that the sudden discharge through the galvano-
meter of 10~" C.G.S. units of electricity produced a throw of
1 division; or, in other words, the quantity which, when dis-
charged suddenly through, gave a throw of 7 divisions was
7 X 10~'®. This was determined by discharging through the
galvanometer a condenser of capacity *1 microfarad ; when charged
to 1 volt, the throw observed was 100 divisions, while the steady
current due to an E.M.F. of '001 volt produced a deflection of
72 divisions.

The observations were made by varying c. There was a
commutator in the battery circuit. In each position of this
commutator two values of c were taken and the corresponding
resting points of the spot on the scale observed. From these the
value of c, which corresponded to the zero position of the spot, was
obtained by interpolation.

These observations were made twice for each position of the
commutator and the mean taken.

We will give one series as an example : —

August 27, 1890. — Temperature of standard fork, IS'S"*.

„ „ Beats „ „ 20 in 65*4 seoonds.

„ 20 in 66-2 „

Condenser No. I.
Frequency, 80 approximately.

Position of
Commutator

Zero
Reading

Resistance

Resting
Point

J5890
J5880

47
61

)5880
(5890

46 ,
49

)5890
(5880

48
51

S5880
(5890

46
50

Temperature of ooils, 17*5^
Beats, 20 in 64*8 sees, at 19-3^

It will be seen that between the third and fourth series the
galvanometer zero has shifted slightly.

382 PRACTICAL STANDARDS

From these we get as the four values of c the following : —

6887-6
6886*6
5888-3
5887-5

Mean, 5887-5 at 17-6''
Correction to 17% ]9

Value of c « 6888-4 at 17",

while the beats are 20 in 66 seconds at 19**, or '307 per 1 second ;
at 19"* the frequency of the standard is 128*066 ; thus the frequency
of the driven fork is 128-066 -'307, %.e., 127-769. Thus for the
driving fork we have the octave below this, or 63*879, while the
frequency of the commutator is 6/4 of this.
Hence in this series : —

n = 79-849, c = 6888-4.

The accuracy attained in this series is a taiv specimen of the
whole. With these explanations we proceed to give the results
in tabular form, showing the date, the values of n and c, and the
resulting value of (7. The wire by which the condenser was
connected to the commutator, together with the commutator itself,
had a certain capacity which was determined in the same way,
merely disconnecting the wire from the condenser. In the
observations in December and June we found: —

a =10, d = 98,730, c = 28,460, n = 63-9,

whence the capacity of the wires is -0000625 microfieirad, while in
August, after the apparatus had been set up afresh in a different
position with new connecting wires, the value of c was 22,200 and
the capacity -0000799 microfarad ; for the wires the values of c
could be determined to about 1 per cent.

In the table the value of C has been corrected for the capacity
of the wires.

Taking the air condensers first, the tables show that, at any
rate for frequencies between 32 and 80 per second, the time of
charging has no effect on the capacity, while the individual
observations in each series are within 1 in 2000 of each other.

For condenser I. the results at frequency 64 are in all the series
the least, but this is not the case with condenser 11.

The capacity of condenser I. shows no change between
December 1889 and June 1890. The observations in August
1890 are all rather greater than those in the earlier series, but

FOR ELECTRICAL MEASUREMENTS

383

the increase, about 1 in 2000, is almost within the error of the
experiments. With regard to condenser II. there is an indication
of a rise in its capacity all through. It will be remembered that
we have already shown that the insulation resistance of 11. is
considerably less than that of I., but it is easy to see that this
leak was not sufficient to account for the change, for if i2 be the
resistance of the leak then our approximate formula becomes

n(7 + -^ = -J , instead of n(7 = ~ .
K ca ca

Table I. Condenser L

Date

Valae of
c

Value of

C, in
microfarads

Mean of
Series

Dec. 31, 1889...-

14762-5
7372-3
5894-3

31-96
63-90
79-876

•021025)
•021016 .
■021019J

•021020

May 20, 1890...-

14772-9
7376-6
5896-4

31-93
63-86
79-825

•021023)
•021017 1
•021025J

H)21022

June 16, 1890...

7375-0

63-86

•021022

Aug. 27, 1890... ,

14746-9
7364-8

6888-4

31-939
63-879
79-849

•021038)
•021027^
•021030)

•02ia32

_ - _

Mean of the whole, *021024 microfarad.

Table II. Condenser II.

Date

Value of
c

13967-4
6963-6
55751

Value of
n

C, in
microfarac

Mean of
Series

Dec. 31, 1889... i

(

31-96
63-90
79-875

-022238]

•022249

•022225.

-022237

May 20, 1890...]

13945-3
6967-4
5568-2

31-93
63-86 ,
79-826

•022271,

-022283 ^
•022266J

-022273

June 16, 1890...

6963-4

63-86

-022296

Aug. 27, lf^90...-

13774-6

6878-6
5500-4

31-939
63-879
79-849

-022623
-022616
•022518.

•022619

Aug. 28, 1890...

6878-6

63-881

•022515

•022516

384

PRACTICAL STANDARDS

Table III. Oiving the Capacity of two Mica Condensers
for various Frequencies of Charge,

Frequency

June 12

June 14

June 16

Mean

Condenser A

•04885

•04886

■04886

•04883

•04884

•04868

•04864

•04867

•04859

Condenser B

•09642

■09642

•09642

•09634

•09642

•09638

Now, the current through the condenser when leaking most
was about 0002 EC, where E is the E.M.F. to which it is charged
and C the capacity of the condenser.

Thus the resistance of the leak is ^^^- — >^, or '26 x 10*"

•0002 X C

C.G.S. units, since the value of C is '02 x 10""". This resistance is

250,000 megohms.

Hence the correction to the capacity = IjnR = '0002 x (7/n, and
this is far too small to affect the result.

There is no doubt, then, that the capacity of II. altered during
the experiments by about 1 per cent., and it will be necessaiy to
take it to pieces and set it up again.

It will be remembered that in the early part of August the
leak in II. was very great, and it seems probable that the steps
taken to discover the cause of the leak have produced a change in
capacity. The experiments on II., then, serve merely to show
that the capacity can be found by the rotating commutator method
to a high degree of accuracy, while those on I. prove that an air
condenser, of *02 microfarad capacity, has been constructed which
has retained its capacity unaltered for the eight months between
January 1890 and August 1890.

The values of c are given in terms of the coils of the legal ohm
box at 17"*. Hence the capacity found needs to be divided by
1*0014 to reduce it to legal microfarads, and it then becomes
■020996.

FOR ELECTRICAL MEASUREMENTS 385

Moreover, since 1 legal ohm = 1 "01 124 B.A.U., and 1 B.A.U.
= '9866 X 10* cm. per sec, we have

1 legal ohm = *9977 x W cm. per sec.

And the absolute electro-magnetic measure of the capacity of the

condenser I. is

-021043 X 10-" sec* cmr\

The effect of the leak in condenser II. was still further
investigated on August 28. The plates of II. were connected by
a resistance of 30 megohms. Hence the correction to C, which is

D becomes - 000620 x 10-» when » = 64

The value of C found with the leak in was -023013 x 10-»

Hence making the correction (7^*02249 micro&rad, which is
sufficiently close to the value found without the artificial leak.

Table III. shows that with mica condensers not very much
greater in separate capacity than the air condensers a change in
the frequency of the charge from 21 to 80 produces an appreciable
change in the capacity. This, of course, is in consequence of the
absorption. With large condensers, as we have already seen, the
effect is more marked.

It remains, then, to give an account of the experiments under-
taken for the purpose of comparing mica or paraffin condensers as
ordinarily used with the air condensers, and of investigating some
of the effects of absorption.

The two well-known methods of De Sauty and Sir William
Thomson have both been employed.

The arrangements are shown in figs. 6 and 7.

The first of these is not really suitable for use in cases in
which there is absorption, though, with care, a &irly accurate
measure of the instantaneous capacity can be found. The re-
sistances /2], i2s can always be arranged so that the effect of the
charge rushing into the air condenser shows itself as a sharp kick
of the spot of light — to the left, say — followed by a slower
deflection in the other direction, due to the absorption charge
soaking into the mica or paraffin. The resistance for which this
sharp kick practically disappears is fairly definite, and from it the
instantaneous capacity can be found, while an observation of the
resulting kick due to the absorption enables us to calculate the
increase of capacity which arises from that cause. This can be

B A. 25

PRACTICAL STANDARDS

done in ratious ways. The simplest, perhaps, ia to disconoect the
condensers from the circuit, and, replacing the mica condenser

by a variable condenser of small capacity, observe the kick this
produces in the galvanometer when charged with the same battery.
From this the capacity to which the absorption is equivalent can
be approximately calculated.

Fig. 7.

Thus a condenser of about '1 micro&rad was compared with
Dr Muirhead's three condensers combined. Taking (7„ R, to refer
to the air condenser, we had

O, = -009506, iZ, = 898650 ohms ;

FOR ELECTRICAL MEASUREMENTS 387

and with i2i=s 89300 there was a slight tremor to the left and
a movement of three divisions to the right. On changing Ri by
100 ohms the change in the motion of the spot was marked.

This gives for the instantaneous capacity Ci ^ '09550 ; the value
found by the commutator at frequency 64 was '09543 micro&rad.

To evaluate the five divisions the air condenser was dis-
connected and the mica condenser replaced by one of capacity
'001 microfarad; the kick observed was 4'8 divisions, while with
*002 microfarad it was 9 divisions. Thus a kick of 5 divisions
corresponds to about *0011 microfarad capacity. Hence the
capacity of the mica condenser, including the full effect of
absorption, is '0966 microfarad.

The second method, about to be described, in which the
absorption effect is included, gave '0965 micro&rad.

Let us now consider the second method. The current from a
battery flows through BiPB^ (fig. 7), a large resistance of amount
jRi + JB^. One plate of each condenser is in contact with Bi and
B2 respectively ; let Vi, F, be the potentials at these points. The
other plates Ai, A^ are insulated and connected together and to
the galvanometer G\ the other pole of the galvanometer can be
connected to P through the insulated key K^. The galvanometer
can be replaced by an electrometer. Let Ri be the resistance
P5, ; iia the resistance PBt, Suppose the point P be put to
earth, the rest of the circuit being insulated ; then if (7i, C, be the
capacities, it is easy to see that there will be no current through
the galvanometer on making the key K^, if C^Ri =» C^R^.

Now, in the case of a mica or parafiin condenser the capacity
is a function of the immediate past history of the condenser, and
different values will be found for the resistances Riy R%, according
to the time the charging has lasted. Dr Muirhead, however,
who uses the method largely, has shown how to obtain the
instantaneous capacity from the observations. His method is
described in the following extract from a letter to myself*. Li
the method as described one pole of the battery is to earth instead
of the point P of fig. 7.

Dr Muirhead writes: ''I have '05 microfarad nearly in air
condensers, and a series of mica condensers of % % '3, '331
(original 1/3), and *498 (original *5) mf. capacity, all enclosed in
a double air-tight box, to keep the temperature as uniform as

* See also EUctrician^ September 5, 1S90.

25—2

388

PRACTICAL STANDARDS

possible. The capacity of these standards is determined periodi-
cally by both the tuning-fork method (using a revolving
commutator instead of the tuning-fork) and by the ballistic
galvanometer method. One can make comparisons of these
condensers among themselves, and with other condensers by the
method I adopt, to an accuracy of 4 in 10,000. The temperature
coefficient of shellacked mica condensers is about *018 per degree
Centigrade, and of paraffined mica -034 per cent.
" Let Si be the capacity of the air condensers ;

„ S^ „ „ „ condenser to be compared with

air condensera

Pig. 9.

Fig. s.

" After making battery contact, supposing the charging of the

condensers to be instantaneous and the absorption nil, then we

have

t;flf, = (F-t;)flf,.

where v is the potential of the junction of the two condensers.
Should there be any delay in obtaining the balance, the position
of V on the slides will vary — say to v, ; then the charges on the two
condensers will be

Vi8i and (F-t;i)(fif, + a)

respectively, where a is the apparent increase of capacity of S^
due to absorption or soaking in of charge. On disconnecting the-
armature of S^ from the slides and putting it to earth, the potential
falls from F to 0, and immediately afterwards the potential of the
junction of the two condensers becomes, say, Vj, so that

FOR ELECTRICAL MEASUREMENTS 389

Hftnnp

Fand t;i are known, and v^ is indicated at once on an electrometer ;
or when a galvanometer is used it can be measured quickly thus :
— As soon as v, has been observed, break the galvanometer contact
and move the index of the slides down to 0 ; then directly after
bringing the armature of £^ f|:om the full potential of the slides
to zero, close the galvanometer circuit and observe the throw, a,
which is a measure of v^, the potential of the junction of the two
condensers."

In my own experiments, which were made after consultation
with Dr Muirhead, I adopted a method practically the same as
his ; but before describing it, it will be better to consider rather
more the effects of absorption. Let us suppose, at first, that
the leakage from either condenser is inappreciable. If there be
no absorption, each condenser is charged to its full potential
practically instantaneously; and it does not matter when or in
what order the keys, K^, K^, are put down, the position of P on
the slide is not affected.

Suppose now that Ci shows absorption, the capacity increases
with the time of charging. We can get the instantaneous
capacity by depressing, first, the key Kj and then Kt, but in this
case we are troubled with the effect of the slow after-charging as
in the other method. Still the resistance, for which the kick due
to the initial charging is zerd, is, with the condensers I employed,
Mrly marked, and a value for the instantaneous capacity can be
thus fairly accurately obtained.

If, now, Ki be made for 1 second and then Ki depressed,
a different position will be found for P. With this interval
of charge the apparent capacity differs appreciably from its
instantaneous value, and the after-effects of the absorption can
still be observed. The same is true for intervals of 2, 3, or
4 seconds — the value obtained for the capacity increases, and the
after-effect is still noticeable ; but with the condensers and battery
I used, if the time of charging was prolonged to 5 seconds, the
after-effect was inapprieciable, and the position of P on the slide,
and hence the apparent value of the capacity, was hardly affected

390 PRACTICAL STANDARDS

by further increasing the time of charge. In the experiments on
a cable recorded in Dr Muirhead's paper already referred to, the
absorption eifects continue much longer. In the observations
recorded below, then, unless the contrary is stated, the key K^
was held down for 5 seconds, and then, K^ being depressed, the
position of P determined, for which the galvanometer remained
una£Pected. The value of the capacity deduced then is the full
capacity for the potential to which the condenser is charged.
It is of course possible, though further experiments would be
wanted to prove it, that the full effect of absorption is not merely
to increase by a definite amount, independent of the potential,
the apparent instantaneous capacity, but that the increase may
depend on the potential to which in each case the condenser is
being charged. It will of course depend on the purposes for
which the condenser is to be used whether the instantaneous
capacity or the fiill capacity is required, and it probably will be
best, when issuing certificates, to state both the instantaneous
capacity and the maximum increase due to absorption — mentioning
at the same time the difference of potential used in the experiments
for determining this correction, and also the time of charging in
which this maximum increase is practically attained.

The method I employed in determining the correction due to
absorption was the following: — Suppose the plates, A^, A^.X^he
at potential zero and uncharged. Make the battery key, /fg, and
after keeping it made for some little time break it again. If
there be no absorption, Ai and A^ will still be at zero potential
and uncharged ; but let there be absorption in one of the two, J.,,
and let B^ be the positive pole of the battery, then, while the
battery is on, negative electricity is being absorbed by the
dielectric near 4,, and positive electricity is left firee over the
plates, il], A^, and the wires connecting them. When the battery
is broken the negative electricity begins to soak out, but the pro-
cess takes time. Hence, if immediately on breaking the battery
key, ifj, the galvanometer key, Ki, is made for an instant, there
is a throw of the galvanometer needle indicating the passage to the
earth of the positive electricity set firee by the absorption. If, after a
time, the galvanometer key be again depressed, there is an equal
throw in the opposite direction, caused by the passage of the
negative electricity which has again soaked out of the condenser.
The required correction is obtained firom either of these throws.

FOR ELECTRICAL MEASUREMENTS 391

For, let % be the current between Bi and B^; let Ci be the
instantaneous capacity of the one condenser and d of the other ;
and let Q be the quantity of electricity absorbed. Then the
quantity of negative electricity on the plate, Ai, is CiRii-^Q, and
the quantity of positive electricity on the plate, 4,, is (7,i2,i, if we
assume the potential of these plates to be still zero.

Therefore, C,R,% -^-Q^G^ R,i;

... ^^-^ Q

Then, neglecting the battery resistance, if E be the E.M.F. of the

battery,

Now, we have seen that with the galvanometer as I used it,
if 7 is the throw produced by the passage of a quantity Q, then
Q = 7 X 10-»«.

The battery consisted of 36 small storage cells, which, when
fully charged, had an b.m.f. of about 75 volts, so that

^ = 76xlO».

Also, C, = 021 micro&rad

« 21 X 10-»

Hence, with these numbers,

C^'R, 1675 V iJi/'
or, writing it as a correction to Ci,

«-^-f5 (•-!)"'-"■

Examples of the method of applying this correction will be
given shortly.

It will be noticed that a leak in one of the condensers may be
corrected for in the same way. For, suppose the mica condenser
to leak, then a quantity Qf of positive electricity passes through
to the plate, Ai, while the battery current is on, and the condition
that the galvanometer should not be deflected is

the same equation as previously.

392

PRAOTICAL STANDARDS

There will, however, be this diflference : on depressing the key,
Ky after breaking the battery circuit, a positive charge virill in
both cases pass from A Ui B through the galvanometer; if this
charge be due to absorption, there will, when the key is again
depressed after an interval, be a current through the galvanometer
in the opposite direction ; while if the first charge be due efitirely
to a leak, there will be no effect when the key is the second time
depressed. In practice, the leak and the absorption may exist
together either in the same or different condensers. In the second
case the leak will tend to produce opposite effects to those caused
by the absorption ; the quantity Q', however, increases nearly in
the ratio of the time of charging, while Q increases for the first few
seconds, but soon reaches a maximum and then remains constant.

These considerations are illustrated by some experiments in
which the condensers I. and II. were compared with various mica
condensers. The battery key was in each case made for 30 seconds ;
it was then broken, and the galvanometer key was made for an
instant. The resulting throw was the sum of those due to (1) the
leak in the mica condenser, (7), say; (2) the absorption in that
condenser, (a), say ; and (3) the leak in the air condenser, which
produces an effect in the opposite direction, — 7*, say.

After about 30 seconds more the key was again depressed ; the
resulting throw is due to the absorbed electricity which has again
leaked out, and will give us — a.

The following table gives the results ; each observation entered
is the mean of three or four.

Condenser

compared with

standard

—

II.

•05

— a

2-3
-3

-7-3
-2-6

•1

X+a-Xi
— a

2-2
-3

-9
-3

•1

X + o-X*
— a

2-2
-2-2

-7
-3-5

•5

X + a-X>
— 0

3-3
-3-3

-4
-2-6

•1

X + a-X*
— 0

4
-5

-3-2
-5-7

FOR ELECTRICAL MEASUREMENTS

393

If we take the comparisons with condenser I. first, it appears
that throughout \ — V is small. For the '06 and '1 microfieurad it
may be about — '5 division, while a is about 3 divisions ; for the
•6 microfarad, a is rather larger, being about 3*3, and \ — X* is
zero, while for the 1 microfarad a the absorption effect is distinctly
larger* being 6 divisions, and \ — V is about — 1. All this is,
of course, quite consistent with the £9u;t that condenser I. and
the mica condensers insulate well while there is absorption by
the mica.

When, however, we come to the condenser II. the results are
quite different. While the absorption effects are comparable, as
of course they ought to be, with those obtained in the comparison
with I., the leakage effects are very large.

The values of X — V in order are as follows : — 9, — 12, — lOS,
— 6*5, — 8. Now, we know that the mica condenser shows very
little leak effect; the above leaks are therefore almost entirely
in the air condenser II. If we suppose the total leak to be
proportional to the time, then for the 5 second charges used in
the experiments the corresponding values of 7 in the corrections
to be introduced for leakage will be one-sixth of the above, and
thus we get the following results : —

Con.
densers

Value of
7

Correction for the
Leak to Capacity '
in Miorofarads ;

Con-
densers

Value of
7

Correction for the

Loik to Capacity

in Microfarads

-05
•I

1-5 1 -00007
2 1 -00016

•5
1-5

•0003
•0007

It is clear that the corrections are in all cases small, being not
much over 1 in 1000, but they serve to illustrate the method.
The above corrections are only those for the leak ; the correction
for absorption could be found in the same way.

With a view to testing the method in a case in which a leak
only existed without absorption, a number of comparisons of I.
and II. were made.

In these experiments the resistance with I. was 296,240. The
resistances with II., and the deflections due to the leak obtained
by breaking the battery and then making the galvanometer, are

394

PRACTICAL STANDARDS

given below, together with the ratio of the two capacities corrected
for the leak.

Interval between

Leak in

Scale

DiviBions

Btittery and

Oalvanometer

Contacts

Resistance

Correction

0 seconds

275,980

1-0734

10734

6 „

275,180

2-5

1-0765

-•0032

1-0733

30 „

271,380

14-5

1-0916

- -0184

1-0732

60 „

267,180

22-5

1-1088

--0286

1-0802

5 »

91,370

4-3223

•0168

4-3391

30 „

92,670

4-2617

•0743

4-3360

eo „

94,170

41938

-1415

4-3353

The last three lines of the table give the results of a series of
comparisons between II., which had a leak, and a condenser of
*1 microfarad, which showed absorption. The resistance with II.
was 394,930 ohms.

In the first four lines the corrections are negative, for the
capacity of the leaky condenser is being found in terms of the
standard. In the next three lines they are positive, for the
ratio of the mica condenser to the leaky standard II. is being
found.

A comparison of the fourth and sixth columns shows the
results of the correction. In the fourth line it is clear that
the correction is not large enough. This probably arises from the
difficulty of making contact with the galvanometer circuit
sufficiently soon after the battery is broken to insure that the
whole of the charge accumulated by the leak should pass through
the galvanometer.

The leak correction was also tested with similar results by
putting an artificial leak in I.

We will now give some specimens of the observations made to
compare I. with a mica condenser in order to show the accuracy
attained. Condenser I. compared with '1 microfarad; resistance
with L, 493,560 ohms ; resistance with *! micro&rad, 105,800
+ a variable resistance given below.

In the table in which the effect of the galvanometer is shown
by the letters 12, L in the last column, R means there was a
deflection to the right, L to the left.

Thus in this case the effect of an alteration of 100 in the

FOR ELECTRICAL MEASUREMENTS

395

resistance, i.e,, ^^ of the whole, is very marked, and we may
take the following values for R : —

5 seconds' interval 105,800 + 500

2 „ „ 105,800+ 650

0 ;, „ 105,800 + 1300

! Literval between
Galvanometer and
Battery Contact

Variable BeslRtanoe
to be added above

Effect on Galvanometer

5 seconds

; 0 "

(700
-^400
1500

(400

hoo

1600
(1200

haoo

11400

very small R

L
R
L

Tremor Z, then swing to R
R

Other series of observations showed that the resistance for
10 seconds' interval was the same as for 5 seconds' ; if the interval
was prolonged to 30 seconds, a very small increase in capacity
was noticeable. Thus the effect of absorption is to increase the
capacity of the '1 microfarad by about 8 in 1000, or '008 of
the whole; of this '0065 shows itself in the first 2 seconds of
charging and '0015 afterwards, the increase after 5 seconds, if any,
being extremely small.

When comparing I. with '5 microfarad the resistances used
were 592,290 and 24,900 respectively. In this case an alteration
in the latter resistance of 10 ohms, or tkW» ^^ easily seen. The
following are the results : —

Interval

Resistance

Interval

Resistance

10 seconds
5 „

24,900
24,900

2 seconds
0 „

24,930
25,060

These again show that the absorption effect disappears after
5 seconds, and that the effect of absorption in 2 seconds is about
'0052, and in 5 seconds about '0064 of the whole capacity*

396

PRACrriCAL STANDARDS

When comparing with 1 microfarad, the resiatances were
592,290 and 12,580, the last number being accurate to about
5 ohms, or about the same proportion as before.

The results of the various observations are given in the
following table ; the observations made with IL have been
corrected for the leak, as already explained.

Table giving tiie Capacities of Certain Mica Condensers as
compared with the Air Condensers,

Date

August 19 ...
„ 23 . . .

Value from I.

•04934
•04934

Valae from II.

Value found by

Commutator at

frequeDcj 64

. .. a

•04938
04936

•04867

June 17 ...
August 14 ...

„ 18 ...

« 21 ...

•09772
•09751
•09773
•09773

•09780

•09786
•09781

•09638

August 18 (h)

„ 18 (A)

>» 21 ...

•5005
•5007
•5006

•5008

•5009

1 5010

August 18 ...
i» 21 ...

•9910
•9913

1 -9912
•9912

It will be noticed that, for either condenser L or IL, the
results are in very close accordance ; with the exception of one
observation, on August 14, the differences are barely as great as
1 in 5000, and the method is clearly capable of giving the value of
a mica condenser, in terms of the air condenser, to this accuracy.

The reason for the low result on August 14 is to be found in
the fisust that on that day the leak was considerable, being, as we
have seen, over 1 per cent, per minute. Full observations for the
correction were not taken; it would, however, amount to about
*0002, judged by the correction required to observations on IL,
when leaking at a similar rate.

The results from IL are equally consistent among themselves,
but all slightly greater than those from I. This would indicate
that the correction applied for the leak in IL is rather too larga

The capacities given in the table are those found with a
5 seconds' interval, by which time, as we have seen, the absorption

FOR ELECTRICAL MEASUREMENTS 397

on the mica condensers used is practically complete. We have
already discussed the method of determining the instantaneous
capacity, and a table of the corresponding values could easily be
given.

For our present purpose it is hardly necessary to do this, and
indeed for many purposes for which condensers are employed a
knowledge of the fiill capacity ia more useful than one of the
instantaneous one. In the last column the values of the capacities
found by the commutator method are given; the differences in
both cases amount to about 1*3 per cent, of the capacity.

During the forthcoming year condenser II. will be again set
up and tested, and the permanent arrangements for rapidly
comparing condensers and for issuing certificates vrill, I hope, be
completed

Appendix III.
On the Specific Resistance of Copper. By T. C. Fitzpatrick.

All the values given in tables for the specific resistance of the
metals are directly or indirectly obtained from the values given by
Matthiessen in his series of papers published in the Traneaxstions
of the Royal Society for the years 1860-1864, and in the Reports
of the British Association for the same years.

In the Transactions* for the year 1860 is a paper by Matthiessen
on the conductivity of pure copper, and on the effects of impurities
in it ; no alloy of copper having as high a conductivity as the pure
metal His results are expressed in terms of the conductivity of a
hard-drawn silver wire (100 at 0** C). He gives the following values
for samples of copper carefully prepared by himself: —

(1) 93-00 at 18-6°\ ^. . , ^

(2) 93-46 „ 20r ^rio^ToV \
3 9302,; 18-44 «308 at 189^ as the

(4) 92-76 „ 19-3« conductivity of pure

(5) 92-99 „ 17-5^ '^P^'"'

Numbers are given showing the effect on the conductivity of
small quantities of oxide, and he states that he found it necessary

♦ Phil. Trant. 1860, p. 86.

398 PRACTICAL STANDARDS

to pass hydrogen through the molten metal for some time for
entire reduction. In the Transactions for 1862, Dr Matthiessen
has a paper on the influence of temperature on the conductivity
of metala He again expresses his results in terms of a hard-
drawn silver vrire. On page 8 of that paper will be found the
results of his experiments on copper : the lowest temperature at
which measurements were made was 12° or 16° ; he there shows
how the results for pure copper measured at 18° may be reduced
to 0° C. ; but no measurement was actually made at 0° for any of
the metals experimented with.

He expresses the influence of temperature on a hard-drawn
copper wire, the mean result of a number of determinations, by

the equation

X = 100 - -38701^ ■{- -0009009^,

where 100 is the conductivity of copper at 0° C, so that a hard-
drawn silver and copper wire have the same conductivity at 0"^ C.

The values obtained by comparison with a hard-drawn silver
wire are then largely the source of the tables of specific
resistances ; but at the end of his appendix to the Report of the
Electrical Standards Committee for 1864, Matthiessen gives values
for hard-drawn silver and copper wires in terms of the new
B.A. unit, expressed as the resistance of a wire one metre long,
weighing one gramme.

These values are : —

Copper ... ... ... ... '1469

Silver -1682

The same table of values is given in the Philosophical
Magazine for 1865, where also is given a table of specific
resistances for wires one metre long, and one millimetre diameter,
expressed in terms of the B. A. unit, and calculated fix>m the value
of the known conducting power of gold-silver alloy in terms of
hard-drawn silver, and also in terms of the B. A. unit.

The values thus obtained do not agree at all well vrith the
results calculated for the resistances of the gramme metre by the
specific gravities of the elements furnished by tables.

Thus :—

Calculated Obsezred

Silver 02048 -02103

Copper -02090 -02104

FOR ELECTRICAL MEASUREMENTS 399

Matthiessen states that be omitted to determine the specific
gravity of the copper used in his experiments ; he probably would
not have obtained any very accurate results, as the weight of
copper he used varied from 1*5 to 4 grammes.

The accuracy of Matthiessen's results seems to depend, there-
fore, on the accuracy of his determination of the resistance in
terms of the B. A. unit of a hard-drawn silver wire ; in considering,
therefore, the question of the preparation of samples of copper
of higher conductivities than Matthiessen obtained, it may be
suggested that the cause of the difference is not explained by
the fact that Matthiessen did not prepare pure copper, but by
an error in the value of the standard with which the comparison
was made.

I have, therefore, made a series of experiments on the
resistance of pure silver wires; and, as a general result, have
obtained a value identical with that of Matthiessen ; the difference
is not due, therefore, to an error in the standard employed, as &r
as my experiments go.

Matthiessen does not give anywhere the details of his measure-
ments of the specific resistances of the metals in terms of the
B. A. unit ; in the B. A. Report he simply mentions that an
approximate table is subjoined, not even stating the fact that the
values are for a temperature of 0° C. I conclude, therefore, that
these values are calculated out from the former, of which an
account is given in the same B. A. Report, and which were
performed at a temperature of 20"^ C.

I have, therefore, on this account, as well as for other reasons
stated later, made my measurements at the temperature of the
air, and believe that as his values were reduced by a temperature
coefficient to values at 0° C, I shall, by using the same tempera-
ture coefficient, obtain results directly comparable with my own
measurements.

For the measurement of the resistance of the specimens of
wire a Wheatstone's bridge arrangement was employed. Two of
the arms of the bridge were formed by a 10 and 1 standard
B. A. unit, namely, 66 and 0 ; these were so nearly 10 to 1, that
they were taken to be in that ratio.

The third arm was ^ of a B. A. unit, and in the fourth arm
was the wire to be measured ; this was stretched on a flat board,
and soldered at the ends to copper plates, to which connecting

400 PRACTICAL STANDARDS

wires were also soldered ; the length of wire used was generally
a little less than two metres, and the wires were, approximately.
No. 18 B.W.G. The board had scales screwed to it at the two
ends. The board and wire were placed in a long bath made of
zinc, and filled with paraffin. Wires which were left in the bath
for some days, and, in more than one case, several weeks, were not
found to have been acted on by the oil.

One end of the wire, Pj, Q„ was connected by a binding screw,
through an adjustable resistance, r (^ metre of copper wire), to
the mercury cup, d, in which was one of the legs of the ^ coil,
and also to a reversing key in the battery circuit. The ^ and the
10-ohm coils were connected up together through an adjustable
resistance, Pii/i, one leg of each of the coils 10 and 1 was in the
same mercury cup, L ; and the other end of the 1 B. A. unit was
connected with the other end of the wire, P^Qs-

A single Leclanch^ cell was connected with the reversing key,
and the fourth point of this key was connected with the mercury
cup Z, into which the legs of 10 and 1 dipped. In this circuit
there was also a touch key. The galvanometer circuit was always
made, and thus there was no thermo-electric effect in the galvano-
meter circuit. To each of the mercury cups Qi, Pj, if,, M^ were
connected two thick wires with separate binding screws: one of
these wires was welded to the copper plate at the bottom of the
mercury cup. Each of these latter wires was connected with two-
way ke]rs ; those in P, and Q, to the key k^ ; those in M^ to the
key Ki ; those in if , to the key K^.

The base points of the keys Kx and K^ were connected with a
delicate reflecting galvanometer, that employed for the comparison
of the standards on the Fleming bridge. The base of the key k^
was connected with the third point on the key K^y and the third
point on the key Ki was connected to the base point of a fourth
key, A^a, the two other points on this key being connected with
riders, with which contact could be made with two points on the
wire PsQs^ the riders had straight edges, and thus their position
on the scales could be easily determined. In performing an
experiment, the keys Ki and K^ were so connected that the
mercury cups, and so the ends of the coils 10 and 1, were in
circuit with the galvanometer. The resistance, PiAT,, was then
varied till, on making the battery circuit, no deflection resulted.
The ends of the 10 and 1 were then at the same potential, and as

FOR ELECTRICAL MEASUREMENTS

401

the other ends of these coils were connected with the same pole of
the battery, there was the same &11 of potential on the two lines.

The keys Ki and K^ were then reversed, and by the keys
ki and k^ one end of the ^ coil and one point on the wire P^Q^
were connected through the galvanometer, and afterwards the two
other ends. The riders were adjusted till there was no deflection
of the galvanometer. The length of wire between the two riders
had then a resistance of -^ that of the ^ B. A. unit coil.

By means of the series of keys it was easy to repeat the
observations, and to connect either end of the ^ coil with the wire.
The resistance PiAfi, did not often change during the experiments.

Fig. 10.

as the room was at a constant temperature ; any change in it only
caused a shifting of the position of the riders. In each experi-
ment, after all the adjustments, the bath was well stirred, and
everything left for half an hour. It was generally found that
the riders did not require any re-adjustment. The battery was
reversed, and all the coils moved. The latter never caused any
effect ; sometimes the reversal of the battery caused a shifting of
the two riders a millimetre or two in the same direction. Another
reading was taken three or four hours after.

The coib, i, 10, and 1, were in water baths, and their tem-
perature remained the same for hours together. The temperature

& A.

402 PRACTICAL STANDARDS

of the paraffin bath was not so constant; it was kept well
stirred, and a thermometer divided to 0*2*^ C. never showed
any difference in the temperature at the different ends of the
bath when the readings were taken. The thermometer employed
was Kew corrected; and the corrections given were verified by
recent comparison with a platinum thermometer by Mr Griffiths.

Since the two standard coib employed were accurately in the
ratio of 10 to 1, the accuracy of the resistance measurement
depended entirely on the value of the ^ 6. A. unit. This was first
made as nearly as possible ^, but it was found that for the size of
the wires measured (18 B.W.G.) this was too high a resistance ; it
had therefore to be reduced. For the determination of its value
there was cut out in a block of paraffin wax a large central
mercury cup, and outside this a circular channel; thick copper
plates were cut to fit them, and both plates were well amalgamated.
By means of this cup arrangement the three B. A. units {H, (?,
and Flat) were connected in multiple arc, and by means of stout
copper rods the multiple-arc arrangement was connected with the
mercury cups on the Fleming's bridge, and so compared with the
^ B. A. unit. The following observations were taken : —

July 12, 1889 : ^ (18-4°) + 986-6 (b.w.d.) = M.A. + 24*6 (b.w.d.),

July 22. 1889 : i (17^) + 986 (b.w.d.) = M.A. + 24*1 (b.w.d.),

August 26, 1890 : ^ (16-8°) + 9861 (b.w.d.) = M.A. + 23-9 (b.w.d.).

The value of a bridge- wire division (b.w.d.) is "0000498 B. A.
unit at IS"", and the wire has a temperature coefficient of *00143.

It is evident from these series of values that the ^ has not
changed in resistance during the period of the experiments.

This comparison, however, introduced a possible error, as the
temperature of the bridge wire at the time of experiment was not
accurately known, and this is important when nearly the whole of
the bridge wire is employed. To eliminate this possible error the
\ was compared with four B. A. units in multiple ara In this case
a large number of bridge-wire divisions had to be subtracted from
the value of the |, and the whole number of bridge-wire divisions
entering into the calculation for the values of the ^ was largely
reduced. The four coils in multiple arc were (^, Q, H^ and
Flat) :—

Aug. 25, 1890 : i, 168° + 157 (b.w.d.) = M.A. + 85205 (b.w.d.),

Aug. 26, 1890 : J, 168° + 157*5 (b.w.d.) = M.A. + 851*9 (b.w.d.).

FOR ELECTRICAL MEASUREMENTS 408

All the four coils were at the same temperature (Id'S""). Their
values are taken from the B. A. Report, 1888 : —

Flat 1-000448

F' ... ... ... ... 1*000028

\jt ... ... ..• ... vvvQQ

XX ... ... •.. ••. v«7t/Oi/

They give for the two multiple-arc arrangements the values
'33330 and '24998. The connecting rods have a resistance of
•00042, and the value of the \ at le*"* is 28537 B. A. unit. Its
temperature coefficient is '0001 per l** C.

To measure the lengths of the wires two microscopes with
scales and verniers reading to '1 of a millimetre were set up and
firmly clamped in position; the distance between them was
determined by means of a beam compass and the aid of a third
microscope: the distance between this and the other two being
directly read off on the beam compass for set positions of the
verniers. The wires were cut with a fine fi:et-saw at the points
corresponding to the position of the riders in the resistance
measurements. Before weighing the wires were carefiiUy cleaned
with methylated spirit. The balance employed was the one used
by Mr Glazebrook for our determination of the specific resistance
of mercury, the weights were balanced against one another, and in
all cases double weighings were taken.

The specific gravity of most of the wires was determined ; for
this purpose distilled water was boiled and cooled rapidly, the coil
of wire immersed, and the beaker and its contents placed under
the receiver of an air-pump, which was connected with a water-
pump; this was left running for two or three hours till all
air-bubbles had disappeared; the weight of the wire in water
was determined, and a second reading taken some hours later.
As the weight of wire used was fix>m 16 to 20 grammes, &irly
accurate values for the specific gravity of the several wires were
obtained, and thus the value for each wire in terms of the B. A.
unit for the resistance to conduction between the opposite faces of
a cube of the material was found.

The first object of these experiments was to test directly in
comparison with the B.A. standards samples of copper wire of
high conductivities, with the view of comparing them with
Matthiessen's standard. Application was therefore made to

26—2

404

PRACTICAL STANDARDS

several firms for high-conductivity copper wirea My thanks
are due to those who sent samples.

A table of results for all the specimens tested is given, and it
shows the variation in resistance of high-conductivity wires.

Resistance of Various Specimens of Wire.

Wire

Date

BesiBtance of a wire

snch that 1 metre

weighs 1 gramme at

IS"* C. in B.A. units

Specific
gravity

Specific lesistance

percc. at IS'^C. in

B.A. units xlO-«

Hard-
drawn

An-
nealed

Hard-
drawn

Annealed

July 22, 1889
Nov. 6, 1889

1649
1650

8-86
8-87

1743
1745

July 22, 1889
Dec. 2, 1889

—

1645
1546

8-88
8-89

1741
1742

III.

Dec. 3, 1889

1713

8-87

1922

IV.

July 10, 1889
Aug. 1, 1889

1578
1578

—

8-89
8-89

1776
1776

IV.'

Nov. 1, 1889

1511

8-885

1724

July 31, 1889
Oct. 30, 1889

1673
1572

8-89
8-89

1770
1770

V.'

July 20, 1889
Aug. 2, 1889
Aug. 8, 1889

—

1526
1626
1627

8-89
8-89
8*89

1712
1713
1716

Aug. 10, 1889
Oct. 18, 1889
July 10, 1890
July 14, 1890

1546
1549
1549
1548

—

8-94

notobsrvd.

1730
1732
1731

VL'

Aug. 8, 1889
Oct. 11, 1889

—

1508
1609

8-94
8*94

1688
1688

VIL

Nov. 4, 1889
July 16, 1890

1543
1543

8-946

1724

VIIL

Oct 23, 1889
Oct. 28, 1889

1700
1702

—

8-96

1903

IX,

Aug. 5, 1889
Aug. 18, 1890

1672
1672

—

8-90
8-90

1766
1766

—

Aug. 5, 1890
Aug. 26, 1890

1573
1569

—

8-91
8-92

1767
1751

—

Aug. 27, 1890

1569

—

8-93

1750

—

Matthi<
due
his

3fl8en'8 value re-|
!ed to 18% using j-
own coefficient j

1671

—

not given

1766-6

As calcu-
lated by
Fleeming
Jenkin and
Fitzpatrick

FOR ELECTRICAL MEASUREMENTS 405

IV. and IV'. are the same copper, but IV. is hard-drawn, IV'.
is annealed ; they were measured just as they were sent from the
manufacturers ; the same is true of V. and V'., VI. and Vr.

It will be noticed that VI. and VI'., which are of considerably
less resistance than the other wires, are of higher specific gravity :
the firm that sent them thus wrote of them, '* It is only occasion-
ally we come across copper as high as this or high enough to be
called the highest (in conductivity) we can produce. This copper
has been produced electrolytically by our ordinary process." How
this copper was treated after electro-deposition I do not know.

1 am inclined to think fix>m my own experience that this difference
in density is due rather to the condition of the copper than to its
relative purity. Matthiessen found that veiy small quantities of
impurities reduced the conductivity 20 or 30 per cent, and a
sufficient amount of impurities to cause this decrease in density
from 8'94 to 8*90 must make a larger increase in the resistance of
the copper.

The temperature coefficient is stated to be different for various
specimens of metal, according to their purity. Matthiessen him-
self seems to have been of this opinion ; but the mere difference
in density of the metal might be expected to affect the alteration
of conductivity with the same change in temperatura I have
not been able to find any experiments bearing on this question.
It is quite easy to obtain samples of wire of different density by
varying the process of drawing, and the temperature coefficients of
such wires might be found to be different

Comparing V. and Y\ with VI. and VI'. it is seen that with
this increase of density there is a distinct diminution in the effect
of annealing.

IV.-IV'. = -O0677^

v.- v. = 00577 i.

VI.-VI'.= 004 j

I thought it might be possible that VI'. was not completely
annealed, so, for a direct comparison, two specimens of VI., which
had been measured hard-drawn on July 10 and 14, 1890, were
annealed ; for this purpose a fiat copper vessel was made of about

2 cm. height and 18 in diameter, with a closely fitting lid; the
wire was packed in this between sheet asbestos, which had been
previously heated; the vessel was filled up vrith lampblack, and

406 PRACTICAL STANDARDS

heated over a big bunsen burner and gradually cooled ; the process
generally took about twenty-four hours ; the wire was found not to
be oxidised after the process was over.

WiM

Hard-drawn

Annealed

Difference

1549

1510

•0039,

II.

1548

15U9

•0039.

The difference Matthiessen obtained was '0038.

The above method of annealing was found very effective.
Silver wires, which on annealing decrease 10 per cent in
resistance, gave the same value after a second annealing as they
did on the first occasion.

Wire VII. was a wire sent me by Mr H. A, Taylor, and had to
be drawn down before it could be measured ; another piece of the
same wire drawn down on a different occasion gave the same
value ; this wire has the lowest resistance of any I have obtained ;
it has, too, the highest specific gravity. Mr Taylor says of it
'' that it has a higher temperature coefficient than that given by
Matthiessen."

Vin. was a sample of wire obtained from Germany, and said
to be electrolytically prepared ; its high resistance is, I think, due
to the presence of oxide, as I fused some of it in hydrogen, and
when measured partially annealed it gave the value *1566 at 18""
for the wire, 1 metre weighing 1 gramme.

IX., X., and XL are wires of my own preparatioiL Pure
copper was prepared electrolytically by Messrs Sutton, of Norwich,
and supplied me in thin sheet, and this was fused in a porcelain
tube 18 centimetres in length and 1 centimetre in diameter; the
tube was fitted up in a small furnace made of sheet iron, and
lined with ganister ; this was heated rapidly in a blast flame led
in at the bottom. Some difficulty was experienced in obtaining
the copper in a solid cylinder. In the early experiments hydrogen
was passed into the tube while the copper was being fused, and
was made to bubble through the molten copper ; on breaking the
tube the copper was found to be full of small holes ; the copper
had absorbed the hydrogen at the high temperature and given it
off again on cooling; on another occasion the copper was fused
down in hydrogen, and the tube was connected with a water-pump
and exhausted and the copper allowed to cool in a vacuum ; this
gave a more continuous cylinder. It was found best to fuse the

FOR ELBCTRICAL MEASUBEMENTS

407

copper under borax, after previous reduction ; a good cylinder of
the metal was thus obtained.

I was unfortunately not able to draw down the copper for
myself; this was very kindly done for me by Messrs Smith, of
Halifax, and Messrs Johnson and Matthey. The porcelain tubes
had been prepared of such a size that the cylinder of copper could
be drawn without further heating ; the copper, therefore, was not
fused after it left my hands.

Two sheets of the electrolytically prepared copper were iused
on different days, and one cylinder was sent to Messrs Smith to be
drawn, and the other to Messrs Johnson and Matthey.

Wires IX. were drawn by Messrs Smith, wires X. by Messrs
Johnson and Matthey.

Wire XL was drawn by Messrs Johnson and Matthey from a
sample of copper which I prepared by electrolysis from a pure solu-
tion of copper sulphate ; the copper was deposited on a plate of copper,
which had had its sur&ce rubbed over with graphite ; by this means
the deposited copper was easily stripped off* the plate ; the other
plate was of platinum. After a time the solution was changed ;
the deposition was very slow, as it was thought that there would
be less likelihood of copper sulphate getting in between the layers
of copper. The deposit was boiled with dilute sulphuric acid and
then in water, and v. as afterwards ftised as above described.

Wires IX. were measured as received; this accounts for the
close agreement between the two determinations. Wires X. and
XL I had to draw down further to measure them on my bridge.

Wires X. (2) and XL were drawn down with great care and
not so much as X. (1).

Below is a table of the measurements made for the deter-
mination of their specific resistances : —

Wire

Valoe of
1/8

Temp.

Weight

wire

Length of
wire for

determina-
tion of

resistance

Length
cut and
weighed

Resistance of

gramme per

metre

IX. (1)

•28547

n-o**

20388

1921

192-5

1574

18-3'

« (2)

•28541

17-4^

20-153

192-4

190-45

1569

17-5'

X. (1)

•28560

18-2''

19-708

189-3

188-8

1577

18-6"

„ (2)

•28536

les"

20-252

192-39

192-34

1561

17-r

•28535

16-7"

20-262

192-11

192-51

1563

n^''

408

PRA.CTICAL STANDARDS

1572

Mean value

1573

•1571

1569

B.A. unit

1569

These values reduced to a common temperature of IS'' are :-

IX. (1)
IX. (2)

X.(l)

X.(2)
XL

Thus '1571 B. A. unit is the resistance at IS'' of a metre of
hard-drawn copper wire weighing 1 gramma

Matthiessen in the B. A. Report* gives as the resistance of a
gramme metre at 0" '1469 B.A. unit.

I have calculated from this the value at IS'', using the

temperature coefficient that he gives in his paper on the influence

of temperature on the conducting power of metals. I have

taken no account of the terms in t' as they practically cancel one

another.

R. IS" = R*^ (1 + 00387010,

R. IS" =1571.

This is the value that I have obtained as the mean of my own
observations.

All my observations were taken at the temperature of the
room, and in the table above the values for the different wires are
given at the observed temperature, and then all reduced to a
common temperature of IS"" C. Most observations of this character
are taken at the temperature of 0° C, but on the whole it seemed
more satisfactory to work at the temperature of the room. In the
comparison of the B. A. units I have found that with a difference
of temperature between coils which are connected by thick pieces
of copper there is always conduction of heat, and it is impossible
to tell accurately what is the real temperature of the coils.

My observations were made in the B. A. room at the Cavendish
Laboratory, which has a north aspect, and often the temperature
did not alter more than a few tenths of a degree, whilst the
temperature of the coil baths often remained perfectly steady
for several consecutive days. I cannot find any observations of
Matthiessen's at 0° C; certainly his observations on copper were
made at 1S°, and, consequently, if- the value given by him at
O'' C. has been obtained by the use of a temperature coefficient,

* B, A, Report, 1864, or PUl, Mag. 1865.

FOR ELECTRICAL MEASUREMENTS 409

my value might be expected to agree with his at IS"", the tem-
perature of his observations, supposing the samples of copper of the
same character.

Matthiessen's results are given in terms of a gramme per metre,
and for wires of metre length and 1 mm. in diameter.

In a paper in the Philosopkicai Magazine, Matthiessen gives
the value for hard-drawn copper in these terms as : —

•02104 B. A. unit.

From his value for the gramme metre, using the specific
gravity 8*95 given by tables, the same quantity was calculated,
but gave the result '0209 ; in a note added he states that had he
used the specific gravity 8'91 his results would have been more
nearly alike; but a specific gravity 8*90, I find, would give an
identical value.

This would show, then, that Matthiessen's own table, calculated
for values obtained by comparison with hard-drawn silver, is
accurate. I have tested silver wires, but have not had time to
draw up the results in tabular form ; and I obtained an almost
identical value for hard-drawn silver wire, as supplied me from
Messrs Johnson and Matthey, as is given by Matthiessen for the
resistance of a gramme per metre.

It will be observed that wires IX. have the specific gravity
8*90, and give a value in terms of B. A. units for a cubic centi-
metre of the material identical with Matthiessen's value; this
value is not given directly by Matthiessen, but is calculated fix)m
his results by Fleeming Jenkin, and given in his table in his
book Electricity and Magnetism, it is 1*652 microhms. I have
calculated it from Matthiessen's value, given in the Philosophical
Magazine, and get the number 1*653. Using the same temperature
coefficient as before, the resistance at 18"^ C. of a cubic centimetre
of hard-drawn copper is 1766*6 x 10~® B. A. units.

On comparing the values for wires IX., X., and XI. in these
terms, the results do not agree so well together as when expressed
in terms of the gramme metre ; there is a corresponding difference
in the values of the specific gravities ; these latter have been very
carefully determined, and the experiments repeated with the results
given.

Wires, therefore, of the same resistance expressed for grammes
per metre, may give a very different result, when expressed as per

410 PRACTICAL STANDARDS

cubic centimetre: attention has been drawn to this fact in the
discussion on the Elmore copper in the Electrician*, M. Roux, of
Paris, in a letter gives the following table for high-conductivity
wire from a paper of M. Hospitalier in UiJlectricien, 1887 ; this
paper I have, unfortunately, not been able to see.

Density 8897 9-32 9*6

Conductivity, equal volume 1024 1067 1108

Conductivity, equal weight 101*7 101*2 101 6

What is 100 in the conductivity units is not expressed.
M. Roux thinks that the former, i.e. for equal volume, is the
more rational method of expressing the result.

Matthiessen expressed all his results in terms of equal weight,
justifying it by the greater accuracy obtainable when working with
small weights of wires. Small errors in the value of the specific
gravity are easily made, and cause a similar error in the result
for equal volumes of diflFerent wires ; unless working with long
lengths of thick wire the weight of the wire is small. The weight
of the water displaced cannot be determined within '5 to 1 milli-
gramme, and that only with care : this error in '5 of a gramme
means only an accuracy of 1 in 500. The values given in my
table are probably correct to 1 in 1500 or 1 in 2000, as the
weight of water displaced was in all cases over 2 grammes.
Results, therefore, for resistances of wires of equal weight are
the most trustworthy, and, I think, also the most satisfactory
if used to express the resistance of a material and not of any
given wire.

Wires X. (1) and X. (2) are of the same copper, but drawn
down separately:. X. (1) was beginning to fray, and another
specimen of the same copper drawn down still further had on this
account to be rejected; this has affected the resistance value
expressed in both ways. Thus: —

X. (1) 1573 1767

X. (2) 1569 1751

but much more so when expressed for equal volumes. In both the
copper is of the same quality.

It will be noticed that with increase of specific gravity there is
a decrease of resistance, even when the results are expressed for
wires of equal weight. The resistance diminishes, therefore, more

* Electrician, December 7, 188S.

FOR ELECTRICAL MEASUREMENTS 411

rapidly than the density increases. Wires of the same quality
may, in consequence of a difference in drawing, have a different
density, and so the results expressed in terms of equal volume
will differ considerably, while those for equal weight are the same,
or approximately so.

The values obtained for IX., X., and XI. are so nearly identical
that it is not unfair to conclude that they are samples of pure
copper ; their value is identical with that obtained by Matthiessen
at, I believe; the same temperature. The greater difference obtained
at 0'' C. between Matthiessen's value and samples of copper tested
now at that temperature is probably due to the &ct that
Matthiessen's value was not determined at 0°, but reduced in
value for that temperature from observations, as stated above, at
about 20° C.

The higher conductivity or less resistance for the two samples
given in the table is due, not to increased purity in the preparation
of the copper, but to the difference in the process of preparation,
whereby a sample of greater density is obtained than results from
the working up of small quantities of copper in the laboratory.

A sample of copper hew been prepared by chemical means with
the help of my friend Mr Skinner, but has not yet been measured.

Appendix IV.

A Comparison of a Platinum Thermometer with some Mercury
Thermometers at Low Temperatures. By E. H. Griffiths,
M.A., Sidney College, Cambridge.

The following communication describes the mode of con-
structing an air-tight platinum thermometer for use at low
temperatures. The thermometer was graduated by means of
the freezing and boiling points of water, and as regards inter-
mediate points Regnault's determinations of the temperature
and pressure of aqueous vapour were adopted. The precautions
observed in the construction of the apparatus, and in the method
of observation, are described. The thermometer was tested by
comparison with a number of thermometers standardised at Kew.
The curves, showing the result of these determinations, are in
remarkably close agreement, and when the observations were

412 PRACTICAL STANDARDS

sufficiently numerous it appeared possible to calibrate the bore
as accurately as by the usual more laborious process. The further
advantage of this method is that thermometers can be compared
under the conditions in which they are to be used.

In a communication to the Royal Society read on June 19,
1890, I described a method of constructing and graduating
platinum thermometers, and gave a table of boiling and freezing
points for various substances lying between 100'' and 600°,
determined by means of these instruments.

Subsequent observations indicate that a slight change appears
to be taking place in the readings of these thermometera I
attribute this (1) to alterations in the glass, (2) to presence of
moisture in the tube — the asbestos roll on which the spiral was
wound being highly hygroscopic. I therefore decided to construct
a thermometer in which there should be no contact between the
glass and the platinum, and which should be thoroughly dry and
hermetically sealed.

I was unable to discover any suitable non-conductor capable of
resisting high temperatures; but in anthracene (melting-point
213'') I found a substance suitable in every respect for use at low
temperatures. I subjected a sample to severe tests and, up to
a temperature of about ISO"*, found it to be a better insulator than
paraffin.

The leads to the coil were constructed of silver, the inner one
a rod and the outer a tube. The resistance of these leads was
about '001 ohm and therefore any change in the external resistance,
caused by change of temperature, might be disregarded. The
silver lecuis approached to within about 1 inch of the spiral, and
were connected to it by moderately thick platinum wires ; thus a
flow of heat from the spiral to the silver was diminished. The
wire forming the coil was about 56 inches in length, and had a
diameter of *005 inch. The spiral was about 2 inches long, having
a resistance of about 135 ohms at 0° C, and the external diameter
of the covering tube was about *3 inch. The ends of the asbestos
roll were made of greater diameter than the portion on which the
spiral was wound, and thus there was no glass contact. The tube
and contents were heated up to a temperature of several hundred
degrees, and dried air passed through for some hours. It was then
exhausted and the open end placed under the surface of melted
anthracene, which was allowed to rise until nearly in contact with

FOR ELECTRICAL MEASUREMENTS

413

the coil. When cool, the whole of the thermometer, from the spiral
to the upper end (about 13 inches) was a solid mass, while the
spiral and asbestos roll were perfectly dry and in an almost vacuous
space. I have taken nearly 600 observations with this thermometer
and cannot detect any signs of change. When the lower part was
undergoing rapid changes in temperature, thermo-electric effects
showed themselves, but by reversing the battery and galvanometer
connexions during each reading these effects were eliminated.
A low-resistance galvanometer was used, and the current which
passed through the thermometer when determining its resistance
did not exceed one-hundredth of an ampere. To illustrate the
closeness of the agreement in the results obtained at different
times I give the following determinations of the resistance at a
temperature of lOO"" determined in the usual manner by means of
a hypsometer with manometer attached. Full corrections were
made in the barometric reading, and the results reduced to
lat. 46^

Date

Temperature

Resistance (after

corr. for temp.

of coils)

July 26
„ 27

Aug. 12
„ 13

100'' C.

>»

18-2029
18-2034
18-2025
18-2031

Mean ..

• • • • a • •

18-2030

The expression for the platinum temperature by this thermo-
meter was

4-6811 ^ ^^' *«*^ X " ^'^^'

almost exactly agreeing with the coefficient of the wire
Mr Callendar's air thermometer {Phil. Trans. A. 1887),

Mr G. M. Clark, B.A. (Sidney College, Cambridge), now joined
me in the investigation, and as we proposed to use this thermo-
meter for the calibration and graduation of mercury thermometers
between O"" and 100"", we decided to obtain intermediate tempera-
tures by means of Regnault's numbers connecting the temperature
and pressure of aqueous vapour. For this purpose we constructed

414

PRACTICAL STANDARDS

a large iron tank with two plate-glass sides, holding about
16 gallons of water, and through two holes bored in the bottom
inserted two barometer tubes, the upper 16 inches of each being
within the tank. One of these was used as a standard barometer,
and was prepared with great care, the distilled mercury with which
it was filled having been boiled in the tube for more than six hours.
The internal diameter of the tube was 14 mm., and the absence of
any meniscus wets very marked. If the level of the surface of the
water in the tank was below the top of the barometer, and the
water warmed, the sublimation of mercury in the vacuous space
was observable. The second barometer was made from the same
length of tubing as the first, and communicated at its upper
extremity with a small flask {A), in which was placed the platinum
thermometer.

Distilled water was boiled in vacuo for some hours, to expel all
traces of air. The flask and barometer tube were then exhausted
by means of an air-pump, and the lower end of the tube placed in

Fig. 11.

a flask (B) containing the previously boiled
water, which rushed up, filling the tube and
flask (A).

The water remaining in B was then boiled
until this flask and a bent tube passing from it
into a basin of mercury, 30 inches beneath, were
completely filled with steam, and, on cooling,
the height of mercury in the tube enabled us
to determine that the pressure on its surface
was that of aqueous vapour only. The water
in the upper flask was then boiled for many
hours, and only allowed to cool occasionally to
permit of the water in the lower flask being
boiled away. To prevent access of air the
steam was driven oflF through the mercury.
When the water in flask A was reduced to
about a tablespoonfiil, the boiling was stopped,
and the level of the mercury was raised until it
flowed back first into flask B and thence into
the barometer tube, as flask A cooled.

The open end of the barometer tube was then sealed, the
flask B replaced by a small cup of dry mercury, and the end of
the tube opened below the sur&ce. The water remaining on the

FOR ELECTRICAL MEASUREMENTS 415

top of the column was driven back into the flask by pouring hot
water over the tube.

During other experiments water occasionally collected on the
mercury, but by means of a concave mirror it was driven back
into the flask ; the mirror was of course removed some time before
an observation was taken.

The tank, filled with water, was maintained at any required
temperature by means of a gas regulator. The lower parts of the
barometer tubes were screened by sheets of asbestos, and the two
cups were connected by a small siphon. The glass sides of the
tank were covered with white paper to prevent radiation;
openings were left for observations, during which the water in
the tank was kept in a continual state of agitation by the
oscillation of a large paddle driven by a water motor. The
paddle, fixed in one corner of the lid, swept across the tank,
driving the water before it, and lifting it at the same time. We
have tried several forms of stirrers, and we believe this to be *a
more effective form than a screw or a plunger.

The difference in the height of the mercury in the two baro-
meter tubes was ascertained by the kathetometer G. 33, in the
Cavendish Laboratory, and by means of it readings could be
taken to '50 mm. Care was taken to bring both levels horizontal
before each observation.

As the coefficient of expansion of the kathetometer scale was
unknown and the temperature of the room usually about 20° C,
we decided to compare it with the standard scale R, whose
coefficient of expansion and scale errors had been determined by
the Standards Department of the Board of Trade*.

Twenty-one comparisons were made (greatest divergence from
the mean '10 mm.), and the result was as follows : — 300*35 mm.
on kathetometer scale at 20"* = 300-35489 of Board of Trade
Standard (S.S.) at 0^

Thus no scale correction was necessary.

The difference (D) of the mercury columns was corrected for
temperature, pressure of mercury vapour and latitude, and the
resulting length denoted by D^: the temperature corresponding
to Do was deduced from the very full table given in Part 3 of
Camelley's Melting and Boiling Point Tables.

The extremities of the curve (at 0** and 100°) having been

* Standard metre, verified Jane 1SS2, designated i2 in Mr Chanej's report.

416 PBACTICAL STANDARDS

determined, it was only necessary to get points between 30°
and 80°.

Ninety observations were taken, and although occasional
divergences presented themselves, the mean path gives a curve
which we believe to be within less than '02° of the true path
at all points. It agrees closely with the curve obtained by
Mr Callendar from the parabola,

'■"[{m)'-m]-

by measuring one-tenth of the ordinate along tfae abscissa*.

The following equation, however, represents its path more
accurately.

y = -018795(-0001991f +000,000,111-5P. The curve itself
is shown in fig. 12.

Fig. 12.

We proceeded to teat our conclusions by comparison with
thermometers standardised at Kew ; for this purpose a rotating
annular ring, through the centre of which the platinum thermo-
meter passed, was inserted in the lid of the tank, in such a
manner that the mercury thermometers, fixed in holes bored near
its circumference, could successively be brought into the field of
view of the kathetometer without any re-adjustment of the

* It mnst be remembered that CftI1eDdu''B dilTerenoe onrve gives the ocmiiBiion
between plaliDam sod air tbennomeler tempGratures. vhilrt Begnanlt nted a
mercury (beimometer (M.A.8. XXI.), and tbaa carve A givn the relation between
pUtinnm and mamuj thermometer tempetatnre.

FOR ELECTRICAL MEASUREMENTS

417

telescope; the thermometers were then read by one observer,
whilst the platinum resistances were taken by the other. The
freezing-points were not, however, determined by this method,
bnt by direct immersion in powdered ice, adopting the precautions
recommended by Guillaume in his Thermomdtrie de Precision.

The following curves were then drawn, which indicate the
result of the comparison of our platinum thermometer with those
standardised at Kew.

Curve

B
C
D
E

Thermometer,
Kew No.

1
Standardised

76148

75149

43762

8394

October 1888

May 1886

Dec. 1880, Jan. 1882, April 1888

All these thermometers were made by Hicks ; the first three
were kindly placed at our disposal by Mr R. T. Glazebrook ; the
last is one of those referred to by Mr W. N. Shaw in a communi-
cation to the B. A. during the Bath Meeting, the successive curves
of which, theu exhibited by him, he has kindly allowed us to copy.

In these diagrams the abscissae represent the temperature — in
the strong curves, that obtained by us, and in the faint, that
obtained by Kew : the ordinates in each case being the divergence
of the actual readings from these results. Where crosses occur at
almost identical temperatures they indicate observations sepcuuted
by a considerable interval of time; in no case did less than
20 minutes elapse, whilst in some several days.

Three only of our observations are unrecorded on these charts,
and in each case, owing to imperfect light, interruptions, etc.,
these experiments were regarded as doubtful before their results
were deduced.

The gradual rise of the zero point is clearly indicated ; apparent
discrepancies are probably due to the fact that the Kew deter-
minations are less frequent than ours, and as a consequence many
of the smaller deviations have escaped notice.

The results show : —

L That thermometers whose range does not include 0^ and
IW may have certain fixed points determined by this method.

B. A.

PEACnCAL STAHDA.BDS

FOR ELECTRICAL MEASUREMENTS 419

2. That an actual calibration of a mercury thermometer can
also be readily accomplished.

3. That the platinum thermometer, properly constructed, may
serve as a standard by which to trace the changes which may take
place in mercury thermometers.

4. That since the readings of the platinum thermometer
are independent of the extent of the stem-immersion, it can be
conveniently employed for the graduation of thermometers partially
immersed, as in ordinary use.

We have since calibrated about twenty thermometers by this
method, and we believe the results to be satisfisu;tory in all cases.

Appendix V.

On ihe Absolute Resistance of Mercury.
By R. T. Olazebrook, F.R.S.

The following' table gives the results of experiments made
since 1882 on the absolute resistance of mercury. The first eight
lines relate to experiments in which the resistance of a wire
has been found absolutely and then expressed in terms of the
resistance of mercury by direct observation. In the next four
lines the results of comparisons between certain coils of wire and
the resistance of mercury are given. It will be noticed that the
value found by Lord Rayleigh for the resistance of 100 cm. of
mercury in B. A. units is considerably in excess of the results of
other experimenters. If in obtaining from his value of the B. A.
unit expressed in ohms the value of the ohm in mercury we use
*9636 instead of '9541, Lord Bayleigh's values 106*24 and 106*21
become 106*30 and 106*27, and the mean result 106*28 is hereby
raised to 106-30.

The observers whose results are given .in the last seven lines,
with the exception of Lorenz, did not themselves directly compare
the results of their absolute determinations with the resistance of
mercury, but with coils usually of German-silver, the value of
which in mercury units was certified either by Siemens or
Strecker.

27—2

420

PBACTICAL STANDARDS

G404fpC5cOCO(NQ4

CS Cv CQ C*3

b* CD A 00 '^ CO CO

« rM Cp (^ 0< P <J>

lO CP lO tO CD CD O

oooo o

ft?

S o

•*» '^ BO

s|§3

"a a

> . Q

I I

-<^ Ob 00 04 iC

t^ ^ eo to ^
CO cd CO CO CO

aO iC lO aC lO
O) <^ Ob 0) (^

^ 09 09 "^

CO CO CO CO
lO iC to iC

od Oi 02 dd

• • • •

to r»

coco
00 ■

r-i "^ Q

'-I "^ ©

$'?

lO CD

CD CD © © CD
00 00 00 00
0d (^ Ob 0d

CD
tO
CD

• M •

a 4) O
0,^00

a S-Q ^
H ^ <D S

J III

•I.1J

-lag

c»

^ o

g ^ g

.s ®

III

fiS-

< cS O
c o ^

s a ^

sag

4 e8 o

00
00

(N CO -^ Tjt t^ t^—, O

OD 00 ^ 00 00 00^ Od

oOoOoOoooOoo s^
09

»o OC
X 00
OC X 00

^ to Oi CO to
00 00 OQ 00 OQ
OO OO 00 OO CO

e
>

^J3 5
bObOg

■S'2^ilt|-3

<*A|l4 08 CO

^-i09C0"^iOCDI>00

O O ph 04

CQ rf aO CD t^ 00 O)

FOR ELECTRICAL MEASUREMENTS 421

The value given by Salvioni in his paper {Rendic(mti deUa
R, Accademia dei Lincei, vol. v. fasc. 7) is *95404. Owing to a
mistake in calculation, in consequence of which a correction
was applied with the wrong sign, the value sent to him from
Cambridge for his B. A. standard was in error by *0005. When
this is corrected his value becomes *95354» thus agreeing very
closely with the others. Salvioni's value in line 11 is obtained
through a coil of Strecker's.

EIGHTEENTH REPORT— CAKDIFF, 1891.

Some fhrther experiments have been made with satis&ctory
results on the air-condensers of the Association. A megohm re-
sistance box has been purchased for use in comparisons of capacity.

With a view to testing the permanence of the resistance
standards it was thought desirable to compare them again with
the mercury standards. This was done in December and January
by the Secretary. The coil Flat was compared with two mercury
tubes constructed in 1884 by Mr J. R Benoit, which had been
filled at Cambridge early in the year 1885, and had remained full
since. An account of the comparison was read before the Physical
Society May 9, 1891, and appears in the Philosophical Magazine,
July, 1891.

The tubes were compared with the B.A. standards. If we take,
as was done in 1885, for the resistance in 6.A. units of a column
of mercury 100 cm. long 1 sq. mm. in section, the value '95412
B.A.U., we have the following results for the resistance of the tubes
in Legal Ohms.

No.

Value in 1885
found by R. T. G.

Value in 1891
found by R. T. G.

1
-99990 -99986
•99917 -99913

The differences are only '00004 Legal Ohms, which is too small
to feel really certain about. If we accept for the resistance of
mercury the value -95362 B.A.U., which (B.A. Report, 1890)
appears the best value, then we have: —

Value given by
Benoit 1885

1-00046
•99954

Value found by
R. T. G. in 1891

1-00033
99959

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 423

These comparisons were made with Flat, and lead to the con-
clusion that it has remained unchanged.

In November, 1890, the Association was invited by the
President of the Board of Trade to nominate two members to
represent the Association on a Committee " On Standards for the
Measurement of Electricity for use in Trade." A meeting of the
Electrical Standards Committee was held on December 2, and it
was agreed to suggest to the Council of the Association the names
of Professor Carey Foster and Mr R. T. Glazebrook as representa-
tives. These gentlemen were appointed by the Board of Trade
together with Mr Courtenay Boyle, C.B., Major Cardew, Mr E.
Graves, Mr W. H. Preece, Sir Wm. Thomson, Lord Rayleigh,
Dr Jno. Hopkinson, and Professor Ayrton.

This Committee after various meetings drew up a report, a
copy of which is printed as Appendix I. to this Report.

The standards of resistance constructed in accordance with
Resolution 6 of the Report are now in the hands of the Secretary,
and are being compared with the standards of the Association.

Numerous experiments on the methods of constructing Clark's
cells, and on the electromotive force of such cells, have been made
at the Cavendish Laboratory by Mr Wilberforce, Mr Skinner, and
the Secretary. These are still incomplete, but the experiments so
far as they have been finished lead to the value 1*434 volts at IS"" C.
for the E.M.F. of the cell. The value found by Lord Rayleigh was
1*435 at the same temperature.

Mr Fitzpatrick has continued his experiments on the re-
sistance of silver, and an account of these will be given in a future
Report

The Committee ask for reappointment with omission of the
names of Principal Garnett and Mr H. Tomlinson, and addition of
those of Dr G. Johnstone Stoney and Professor S. P. Thompson.
They recommend that Professor Carey Foster be Chairman, and
Mr R. T. Glazebrook Secretary. They further ask to be allowed
to retain an unexpended balance of last year's grant, amounting to
£17. 4«. 6(2., as well as for a new grant of £10.

424 PRACTICAL STANDARDS

APPENDIX I.

Report of the Electrical Standards Committee appointed

BY THE Board of Trade.

To the Right Honourable Sir Michael Hicks-Beach, Bart., M.P.,

'President of the Board of Trade.

In compliance with the instructions contained in your Minute
of the 16th December last, that we should consider and report
whether any, and, if so, what action should be taken by the
Board of Trade under section 6 of the Weights and Measures
Act, 1889, with a view to causing new denominations of standards
for the measurement of electricity for use for trade to be made
and duly verified, we have the honour to submit the following
report : —

1. Before coming to a decision as to the points referred to us,
we were anxious to obtain evidence as to the wishes and views of
those practically interested in the question, as well as of Local
Authorities who are concerned in the administration of the Weights
and Measures Acts.

2. With this view we prepared drafb resolutions embodying
the proposals which, subject to further consideration, appeared to
us desirable, and forwarded copies to the representatives of various
interests for criticism. Copies were also forwarded to the Press.
We also invited the following bodies to nominate witnesses to give
evidence before us : —

The Association of Chambers of Commerce of the United

Kingdom.
The Association of Municipal Corporations.
The London County Council.
The London Chamber of Commerce.

3. In response to this invitation the following gentlemen
attended and gave evidence : —

On behalf of the Association of Chambers of Commerce,
Mr Thomas Parker and Mr Hugh Erat Harrison.

On behalf of the London County Council, Professor Silvanus
Thompson.

FOR ELECTRICAL BCEASUREMENTS 425

On behalf of the London Chamber of Commerce, Mr R. E.

Crompton.
The Association of Municipal Corporations did not consider
it necessary to ofifer any oral evidence, but the follow-
ing resolution passed by the Law Committee of that
body was adopted by the Council of the Association: —
" The Committee are of opinion that, assuming that
the science of electricity has advanced so far that
it ' is now possible properly to define the three
units referred to in the Board of Trade letter/'
(%.e., the ohm, ampere, and volt) " and to construct
an instrument for the purpose of standard measure-
ment, the time has arrived for the Board of Trade
to take action thereon."

4. In addition to the witnesses above referred to the following
gentlemen were invited to give evidence, and we are indebted to
them for valuable information and assistance : —

Dr J. A. Fleming.

Dr Alexander Muirhead.

5. We also had the advantage of the experience and advice
of Mr H. J. Chaney, Superintendent of Weights and Measures,
who, at the request of our Chairman, was present at our meetings.

6. After a careful consideration of the questions submitted to
us, and the evidence given by the various witnesses, we have agreed
to the following resolutions : —

Resolutions.

L That it is desirable that new denominations of standards
for the measurement of electricity should be made and
approved by Her Majesty in Council as Board of Trade
standards.

2. That the magnitudes of these standards should be de-
termined on the electro-magnetic system of measure-
ment with reference to the centimetre as unit of
length, the gramme as unit of mass, and the second as
unit of time, and that by the terms centimetre and
gramme are meant the standards of those denomina-
tions deposited with the Board of Trade.

426 PRACTICAL STANDARDS

3. That the standard of electrical resistance should be

denominated the ohm, and should have the value
1,000,000,000 in terms of the centimetre and second.

4. That the resistance ofifered to an unvar3ang electric

current by a column of mercury of a constant cross
sectional area of one square millimetre, and of a length
of 106*3 centimetres at the temperature of melting ice,
may be adopted as one ohm.

5. That the value of the standard of resistance constructed

by a committee of the British Association for the
Advancement of Science in the years 1863 and 1864,
and known as the British Association unit, may be
taken as '9866 of the ohm.

6. That a material standard, constructed in solid metal, and

verified by comparison with the British Association
unit, should be adopted as the standard ohm.

7. That for the purpose of replacing the standard, if lost,

destroyed, or damaged, and for ordinary use, a limited
number of copies should be constructed, which should
be periodically compared with the standard ohm and
with the British Association unit

8. That resistances constructed in solid metal should be

adopted as Board of Trade standards for multiples
and submultiples of the ohm.

9. That the standard of electrical current should be

denominated the ampere, and should have the value
one- tenth (0*1) in terms of the centimetre, gramme,
and second.

10. That an unvarying current which, when passed through

a solution of nitrate of silver in water, in accordance
with the specification attached to this report, deposits
silver at the rate of 0001 118 of a gramme per second,
may be taken as a current of one ampere.

11. That an alternating current of one ampere shall mean

a current such that the square root of the time average
of the square of its strength at each instant in amp^s
is unity.

12. That instruments constructed on the principle of the

balance, in which, by the proper disposition of the con-
ductors, forces of attraction and repulsion are produced.

FOR ELECTRICAL MEASUREMENTS 427

which depend upon the amount of current passing, and
are balanced by known weights, should be adopted as
the Board of Trade standards for the measurement of
current whether unvarying or alternating.

13. That the standard of electrical pressure should be

denominated the volt, being the pressure which, if
steadily applied to a conductor whose resistance is one
ohm, will produce a current of one ampere.

14. That the electrical pressure at a temperature of 62° F.

between the poles or electrodes of the voltaic cell
known as Clark's cell, may be taken as not differing
from 1*433 volts by more than an amount which will
be determined by a sub-committee appointed to in-
vestigate the question, who will prepare a specification
for the construction and use of the cell.

15. That an alternating pressure of one volt shall mean a

pressure such that the square root of the time-average
of the square of its value at each instant in volts is
unity.

16. That instruments constructed on the principle of Sir

W. Thomson's Quadrant Electrometer used idiostati-
cally, and for high pressures instruments on the
principle of the balance, electrostatic forces being
balanced against a known weight, should be adopted
as Board of Trade standards for the measurement of
pressure, whether unvarying or alternating.

7. We have adopted the system of electrical units originally
defined by the British Association for the Advancement of Science ;
and we have found in its recent researches, as well as in the
deliberations of the International Congress on Electrical Units,
held in Paris, valuable guidance for determining the exact magni-
tude of the several units of electrical measurement, as well as for
the verification of the material standards.

8. We have stated the relation between the proposed
standard ohm and the unit of resistance originally determined
by the British Association, and have also stated its relation to the
mercurial standard adopted by the International Conference.

428 PRACTICAL STANDARDS

9. We find that considerations of practical importance make
it undesirable to adopt a mercurial standard ; we have, therefore,
preferred to adopt a material standard constructed in solid metal.

10. It appears to us to be necessary that in transactions
between buyer and seller a legal character should henceforth be
assigned to the units of electrical measurement now suggested,
and with this view, that the issue of an Order in Council should
be recommended, under the Weights and Measures Act, in the
form annexed to this report.

Specification referred to in Resolution 10.

In the following specification the term silver voltameter means
the arrangement of apparatus by means of which an electric
current is passed through a solution of nitrate of silver in water.
The silver voltameter measures the total electrical quantity which
has passed during the time of the experiment, and by noting this
time the time-average of the current, or, if the current has
remained constant, the current itself can be deduced.

In employing the silver voltameter to measure currents of about
1 ampere the following arrangements should be adopted. The
kathode on which the silver is to be deposited should take the
form of a platinum bowl not less than 10 cm. in diameter, and
from 4 to 5 cm. in depth.

The anode should be a plate of pure silver some 30 square cm.
in area and 2 or 3 millimetres in thickness.

This is supported horizontally in the liquid near the top of the
solution by a platinum wire passed through holes in the plate at
opposite comers. To prevent the disintegrated silver which is
formed on the anode fix>m falling on to the kathode, the anode
should be wrapped roimd with pure filter paper, secured at the
back with sealing wax.

The liquid should consist of a neutral solution of pure silver
nitrate, containing about 15 parts by weight of salt to 85 parts of
water.

The resistance of the voltameter changes somewhat as the
current passes. To prevent these changes having too great an
effect on the current, some resistance besides that of the volta-
meter should be inserted' in the circuit. The total metallic
resistance of the circuit should not be less than 10 ohms.

FOR ELECTRICAL MEASUREMENTS 429

Method of Making a Measurement

The platinum bowl is washed with nitric acid and distilled
water, dried by heat, and then left to cool in a desiccator. When
thoroughly dry it is weighed carefully.

It is nearly filled with the solution, and connected to the rest
of the circuit by being placed on a clean copper support, to
which a binding screw is attached. This copper support must be
insulated.

The anode is then immersed in the solution so as to be well
covered by it and supported in that position ; the connexions to
the rest of the circuit are made.

Contact is made at the key noting the time of contact. The
current is allowed to pass for not less than half an hour, and the
time at which contact is broken is observed. Ccure must be taken
that the clock used is keeping correct time during this interval.

The solution is now removed fix>m the bowl and the deposit is
washed with distilled water and left to soak for at least six hours.
It is then rinsed successively with distilled water and alcohol and
dried in a hot-air bath at a temperature of about 160"* C. After
cooling in a desiccator it is weighed again. The gain in weight
gives the silver deposited.

To find the current in amperes this weight, expressed in
grammes, must be divided by the number of seconds during which
the current has been passed, and by -001118.

The result will be the time-average of the current, if during
the interval the current has varied.

In determining by this method the constant of an instrument,
the current should be kept as nearly constant as possible, and the
readings of the instrument taken at frequent observed intervab
of time. These observations give a curve from which the reading
corresponding to the mean current (time-average of the current)
can be found. The current, as calculated by the voltameter,
corresponds to this reading.

430 PRACTICAL STANDARDS

Provisional Memorandum on the Preparation of the

Clark's Standard Cell.

Definition of the Cell,

The cell consists of zinc and mercury in a saturated solution
of zinc sulphate and mercurous sulphate in water, prepared with
mercurous sulphate in excess, and is conveniently contained in a
cylindrical glass vessel.

Preparation of the Materials.

1. The Mercury. — To secure purity it should be first treated
with acid in the usual manner, and subsequently distilled in
vacuo.

2. The Zinc. — Take a portion of a rod of pure zinc, solder
to one end a piece of copper wire, clean the whole with glass
paper, carefully removing any loose pieces of the zinc. Just
before making up the cell dip the zinc into dilute sulphuric
acid, wash with distilled water, and dry with a clean cloth or filter
paper.

3. The Zinc Sulphate Solution. — Prepare a saturated solution
of pure ("pure re-crystallised") zinc sulphate by mixing in a
flask distilled water with nearly twice its weight of crystals of
pure zinc sulphate, and adding a little zinc carbonate to neutralise
any free acid. The whole of the crystals should be dissolved with
the aid of gentle heat, i.e. not exceeding a temperature of SO"* C,
and the solution filtered, while still warm, into a stock bottle.
Crystals should form as it cools.

4. The Mercurous Sulphate. — Take mercurous sulphate, pur-
chased as pure, and wash it thoroughly with cold distilled water
by agitation in a bottle; drain ofif the water, and repeat the
process at least twice. After the last washing drain off as much
of the water as possible.

Mix the washed mercurous sulphate with the zinc sulphate
solution, adding sufiicient crystals of zinc sulphate from the stock
bottle to ensure saturation, and a small quantity of pure mercury.

FOR ELECTRICAL MEASUREBCENTS 431

Shake these up well together to form a paste of the coiisistence
of cream. Heat the paste sufficiently to dissolve the crystals, but
not above a temperature of 30^ C. Keep the paste for an hour at
this temperature, agitating it from time to time, then allow it to
cool. Crystals of zinc sulphate should then be distinctly visible
throughout the mass ; if this is not the case, add more crystals
from the stock bottle, and repeat the process.

This method insures the formation of a saturated solution of
zinc and mercurous sulphates in water.

The presence of the free mercury throughout the paste pre-
serves the basicity of the salt, and is of the utmost importance.

Contact is made with the mercury by means of a platinum
wire about No. 22 gauge. This is protected from contact with the
other materials of the cell by being sealed into a glass tube. The
ends of the wire project from the ends of the tube ; one end forms
the terminal, the other end and a portion of the glass tube dip
into the mercury.

To set up the CeU.

. The cell may conveniently be set up in a small test tube of
about 2 cm. diameter, and 6 or 7 cm. deep. Place the mercury in
the bottom of this tube, filling it to a depth of, say, 1*5 cm. Cut a
cork about '5 cm. thick to fit the tube ; at one side of the cork
bore a hole through which the zinc rod can pass tightly ; at the
other side bore another hole for the glass tube which covers the
platinum wire ; at the edge of the cork cut a nick through which
the air can pass when the cork is pushed into the tube. Pass the
zinc rod about 1 cm. through the cork.

Clean the glass tube and platinum wire carefully, then heat
the exposed end of the platinum red hot, and insert it in the
mercury in the test tube, taking care that the whole of the exposed
platinum is covered.

Shake up the paste and introduce it without contact with the
upper part of the walls of the test tube, filling the tube above the
mercury to a depth of rather more than 2 cm.

Then insert the cork and zinc rod, passing the glass tube
through the hole prepared for it. Push the cork gently down
until its lower surface is nearly in contact with the liquid. The
air will thus be nearly all expelled, and the cell should be left in

482 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

this condition for at least twenty-four hours before sealing, which
should be don^ as follows : —

Melt some marine glue until it is fluid enough to pour by its
own weight, and pour it into the test tube above the cork, using
sufficient to cover completely the zinc and soldering. The glaaa
tube should project above the top of the marine glue.

The cell thus set up may be mounted in any desirable
manner. It is convenient to arrange the mounting so that the
cell may be immersed in a water bath up to the level of, say, the
upper surface of the cork. Its temperature can then be de-
termined more accurately than is possible when the cell is in air.

NINETEENTH REPORT— EDINBURGH, 1892.

The work of testing resistance coils at the Cavendish La-
boratory has been continued. The Committee have ceased issuing
standards in terms of the legal ohm of the Paris Congress. Most
of the coils tested are ohms as defined by the resolutions of the
Committee and the Electrical Standards Committee of the Board
of Trade (see " B.A. Reports," 1890 and 1891). According to these

1 B.A. unit = -9866 ohm.,

1 ohm = 101358 B.A. unit.

Four ohm coils constructed by Messrs Elliott Bros, for the
Board of Trade, one of which is to be selected as the legal unit
of resistance for the United Kingdom, were tested very carefully.
Some fourteen or fifteen comparisons were made for each coil at
temperatures between 9° and 18° C. between June 1891 and January
1892. The coils were compared directly with the B.A. unit "Flat,"
the difference being expressed in terms of the bridge wire ; while
in another series of observations a coil of resistance, 100 ohms,
was put in multiple arc with the ohm standard and the difference
between "Flat" and the combination found ; in this case the length
of the bridge wire used was small, and the possible error arising
from uncertainty as to its exact temperature was avoided. That
this error was very small was proved by the fact that the two sets
of observations gave practically identical results. The following
are the results: —

Elliott, 261 3^ No. 300 1 + -000309 (/- 15-36)

EUiott, 262 3^ No. 301 1+ -OOOSIOC^- 15-36)

Elliott, 263 ^ No. 302 1 + 000300 {t - 15-4)

EUiott* 264 ^ No. 303 1 + -000312 (t - 15-4)

In the case of two of the coils, Nos. 261 and 263, there was
one observation for each which differed from the value given by
the above formula by '00015 ohm, and this was due to the fact
that the ends of the coils had got dirty and needed reamalga-
mation. None of the other errors in the sixty observations
exceeded *00008 ohm, and there were only eight which were as

B. A. 28

434 PRACTICAL STANDARDS

great as '00005. Thus the resistances of these coils are known in
terms of the B.A. standards to a very high degree of accuracy.

During the year Messrs Elliott Bros, have supplied the
Committee with two 1-ohm and two 10-ohm standards; the tests
of these are in progress; two 100-ohm standards are on order.
Messrs Nalder Bros, are also. constructing some standards. The
Fleming bridge belonging to the Association has been put into
thorough repair; the mercury had damaged some of the copper
connecting pieces.

The Secretary and Mr Skinner have continued during the year
their experiments with Clark cells. These have been communi-
cated to the Royal Society, and the paper is being printed in the
Phil, Trans,

They find that the E.M.F. of their standard cell is 1'4342 volt
at 15" C, while cells set up by Lord Eayleigh in 1883, 1884,
Mr Elder in 1886, Mr Callendar in 1886, Dr Muirhead in 1890,
Dr Eahle of Berlin, Dr Schuster, Mr Wilberforce, Mr Griffiths,
and themselves in 1891 and 1892 agree closely, the variations
among them all being very rarely greater than '0005 volt.

During the Edinburgh meeting the Committee were honoured
with the presence of Dr von Helmholtz, M. Guillaume of Paris,
Professor Carhart of the United States, Dr Lindeck and Dr Eahle
of the Berlin Reichsanstalt. These gentlemen came by invitation
to consider the question of establishing identical electric standards
in various countries, and two meetings of the Committee were
held, at which discussions took place. Major Cardew, of the Board
of Trade, was present, and took part in the discussion. Dr von
Helmholtz drew special attention to the need for having a unit of
resistance defined in terms of a specified column of mercury, and
pointed out that the difficulty arising fi-om the uncertainty of the
relation between the centimetre and the gramme might be avoided
by defining the mass of mercury column of given length, which
has a resistance of one ohm. After discussion the following reso-
lutions were agreed to : —

1. That the resistance of a specified column of mercury be
adopted as the practical unit of resistance.

2. That 144521 grammes of mercury in the form of a column
of uniform cross section 106*3 cm. in length at 0° C. be the specified
column.

3. That standards in mercury or solid metal having the same

FOR ELEOTBICAL MEASUREMENTS 435

resistance as this column be made and deposited as standards of
resistance for industrial purposes.

4. That such standards be periodically compared with each
other, and also that their values be redetermined at intervals in
terms of a freshly set up mercury column.

It was further agreed that these resolutions should be com-
municated to the Electrical Standards Committee of the Board
of Trade.

With regard to the unit of current and of electromotive force,
it was agreed that the number '001118 should be adopted as the
number of grammes of silver deposited per second from a neutral
solution of nitrate of silver by a current of 1 ampere, and the value
1*434 as the electromotive force in volts of a Clark cell at 15* C.

Dr von Helmholtz expressed his full concurrence in these
decisions, which are, as he informed the Committee, in accord with
the recommendations which have already been laid by the Cura-
torium of the Reichsanstalt, as well as by himself, before the
German Government.

The Committee rash to place on record their thanks to Dr von
Helmholtz and the other visitors for the help they have afforded
them in coming to so satisfactory a conclusion.

Dr Lindeck laid before the Committee some information as to
the properties of the manganese alloy used at the Reichsanstalt
for resistance coils {see Appendix IV.), and it was agreed that it
was desirable to obtain copies of the German standards in man-
ganese for further comparison with the standards of the Association.
The Committee therefore recommend that they be reappointed,
with the addition of the name of Mr George Forbes, and with a
grant of £25, including an unexpended balance of £10; that
Professor G. Carey Foster be Chairman, and Mr R. T. Glazebrook,
Secretary.

Appendix I.

Information circulated by the Secretary for the Meeting of the
Conimittee on August 4, 1892, with additional Notes.

The Report of the Electrical Standards Committee of the
Board of Trade is printed in the "B.A. Report" for 1891.
Further information as to the values of the units is given in the
"B.A. Report" for 1890. The following summary may be of

use: —

28—2

436

PRACTICAL STANDARDS

® is

ii t>» CO N (N Oi •-• '^ «-• !-•

C004C0P9CQ(^C0C0C0 CO

CO CO CO CO

t* 90 OB 00 "^
CO F-^ 00 OB (N

• • • • -

kO ^ (O lO

oooo

coeo

S§81SS

i-H

"^ "OB go <M lO i-i

r* rp w ^ lo "^

CO CO CO Co CO co

lO kfB lO iQ O to
OB OB OB <^ ^ OB

-^ CM N -^

CO CO

lO tft tfB lO

OB 9^ OB ^

•m "S CO

® 30
,2 •©

q « 9
'^ > is;

I
si's J

1-^ 1>- r-l

- r-l "^ Q O CD 't

lOr-r-i'^^coQOco

CDCOCOCOCOCDCDCO
aOaOQOQOQOOOOOaO

ob^^obob^ob^

00
OB

— S 2 S

J i^ a

£^ § a^

'S'?

.s »

meth(
meth

Lorenz
Lorenz

eC CD
lO

ffi 00 CD

o S >

s
o

i1 g

lis

a s ^

4 « O

**^ *•-• -S

bObOP

c a "^

sag

eS 4 o

00
00
00

« CO Tf t^ t^^

00 00 00 00 00 5
QOOOOOOCOO s^oooo

O) OB

00 oO ^

00 00 00

"^ »0 OB CO lO

00 00 op gc X
00 00 00 00 00

MJ3
bobO

a «

H <M » 5 C3

'H'H I ^'S 3*3 s

-f?^

s
o

a
-g > >

QQWOQai

»4 U

••H »

i-i©IC0'*O;Dt^000B

r-l (N ®1

tij S cq S Q ^ 3

CO '^ kC QD r- OC Ob

FOR ELECTRICAL MEASUREMENTS 437

The Board of Trade Committee recommended for adoption the
values '9866 and 106*3.

The specific resistance of mercury in ohms is thus

•9407 X 10-*.

Also 1 Siemens' unit = '9407 ohm

= •9535B.A.U.
1 ohm = 1-01358 B.A.U.

The results found by Lord Rayleigh were reduced by the use
of the value for the specific resistance of mercury in B.A. units
found by him. If instead we use the value '9535, the mean of
the best determinations, we have, as in the table, for the length
of the mercuiy column having a resistance of 1 ohm the values
106*27 and 106*31. At the meeting of the Committee M.
Ouillaume stated that a correction should be made to M. Wuilleu-
meier's value, which would raise it to 106*31. This arose from
some uncertainty as to the correction required to the resistance of
mercury for change of temperature on which M. Guillaume had
been working lately. Taking these corrections we arrive at 106*31
as the mean of the above.

2. The Electrolysis of^ Silver,

The following values have been found for the mass of silver
deposited firom a solution of silver nitrate in one second by a
current of one ampere : —

Mascarfc, J. de Physique, ili. 1884 0011156

Rayleigh, Phil. Trans, ii. 1884 0011179

Kohlrausch, WiscL Ann. uvii. 1886 0011183

T. Gray, PhU. Mag. xxii. 1886 about 001118

Potier et Pollat, J. de Physique, ix. 1890 ... 0011192

Dr Schuster has shown (Proc. R. 8. 1. 1892) that the amount
of silver deposited when the voltameter is in a vacuum is about
"04 per cent, greater than when it is in air.

3. Clark Cells.

The following values have been found for the E.M.F. of a Clark
cell at 15"* C. They have been reduced firom those given in the
original papers on the supposition that 1 B.A.n. a *9866 ohm, and

438 PRACTICAL STANDARDS

that the ma8s of silver deposited per second per ampere is '001118
gramme : —

Rayleigh, Phil. Trans, ii. 1884 1-4345 Volt

vycLrUCBAU ••( ••« ■•• •»• tea ••• ••• X ^xOtzv ••

Kahle, ZeUschrift fUr Iiistnimentenkuruie^ 1892 ... 1*4341 „
Glazebrook and Skinner, Proc, R. S. li. 1892 ... 1*4342 „

A comparison has been made between the standards of the
Association and the Berlin standards through a Clark cell and a
resistance coil belonging to Dr Schuster. By the kindness of
Dr von Helmholtz the Secretary is able to communicate the
results of a direct comparison between the two sets of standards
to the meeting (see Appendix VIII.).

Appendix II.

On the Change of Resistance of Mercury with Temperature.

By M. G. GUILLAUME.

At the meeting of the Committee M. Guillaume communicated
the results of his determination of the relation between the
resistance of mercury and the temperature. Great precautions
were taken with the view of ensuring that the whole of the
mercury in the tube should be at the temperature of the bath.
Two series of determinations with different arrangements in
the bridge were made. The results of these two series give for
the resistance of mercury in a glass tube in terms of the tempera-
ture the values —

(a) iJ, = Eo (1 + -00088023^ + -OOOOOlOOesr*) ;

(6) i2t = Eo (1 + -OOOSSISTT + •0000009909r»).
And for the specific resistance of mercury the values —

(a) pt = />o (1 + •000887457 + -000000181 1>) ;

(b) Pt = />o (1 + -000888767 + •0000010022r>).

In the formulae T is the temperature reckoned from freezing-point

by the air-thermometer. According to Mascart, de Nerville, and

Benoit —

Rt = iio (1 + 00086497+ -000001127*) ;

while according to Strecker —

Rt = Ro(l+ 0009007 + 0000004570.

FOR ELECTRICAL MEASUREMENTS 439

Appendix III.
On a Special Form of Clark Cell. By Professor H. J. Carhart.

Portability, — Standard cells must be portable in order to make
them serviceable for general, technical, and scientific purposes.
To secure portability I have adopted the following construction :
Into the bottom of a glass tube f in. x 2J in. is sealed a platinum
wire. In filling pure distilled mercury is first poured into the
tube. On this is placed the mercurous sulphate paste. A tightly
fitting cork diaphragm is then pushed down firmly upon the
paste. Some zinc sulphate solution is then poured in, and a zinc
rod is immersed in this solution, its lower end touching the cork.
The tube is then securely sealed. Such a cell is perfectly
portable, and may be sent by post without disturbance to its
contents.

Temperature'Coefficient — It is well known that an increase in
the density of the zinc sulphate solution decreases the E.M.F. of
a Clark cell. This effect is included in the temperature-coefficient
of a Clark cell containing crystals of zinc sulphate, since some of
the crystals dissolve when the temperature rises and the density
increases. When the temperature falls recrystallisation occurs.
To avoid the change in e.m.f. due to this change in density !• have
preferred to use a solution saturated at O^'C. Such a solution
I have found to have a specific gravity of 1'397 at 20*" C. The
E.M.F. of the cell is then 1*44 volt if the Clark cell as made by
Lord Rayleigh be taken as 1*434 volt.

The temperature-coefficient may be written as follows : —

. ^e = ^15 {1 - -000386 {t - 15) + -0000005 (t - 15)*},

or per degree the coefficient is

- 000387 + 000001 {t - 15).

This is almost exactly half the value usually obtained for the
Clark cell with crystals.

A cell made with such a solution has the advantage that it
reaches its equilibrium quickly after a change of temperature.
Cells made in the old way require time for the process of crystal-
lisation and for diffusion.

440 PRACTICAL STANDARDS

Appendix IV.

On Wire Standards of Electrical Resistance, By Dr St Lindeck,
Assistent bei der Physikalisch-Technischen Beichsanstalt,
Charlottenburg, Berlin.

It is well known that electrical resistances made out of the alloys
generally used for this purpose, as German-silver, nickelin, etc.,
change their value in the course of time, and this in a degree which
cannot be allowed in measurements meant to be at all accurate.
Such a variation is naturally the greater the more unstable the
material and the shorter the time elapsed since winding.

The following is a short account of the researches conducted
in the Physikalisch-Technischen Reichsanstalt in Charlottenburg
on this subject by Dr Feussner and myself. That investigation
had to be undertaken specially, as it is one of the duties of the
Electrotechnical Laboratory of this institution to secure that
reliable standards of resistance may be obtainable by electricians.

We found by preliminary experiments that by more or less
continued heating a coiled wire showed an interesting variation of
its resistance, which led to a systematic investigation of the whole
question*.

For different materials we determined: —

1. The chemical composition, the temperature-coefficient, and
the specific resistance of the material.

2. The variation of resistance through the strain produced
by winding.

3. The time-variation during the period subsequent to
winding.

4. The influence of heating to different temperatures.

A piece of the double silk-covered wire was wound on a
wooden bobbin and its ends soldered to thick copper bars. The
bobbin was placed directly in a petroleum bath, in order to deter-
mine exactly the temperature, and its resistance was accurately
measured by a Wheatstone's bridge arrangement. As regards
the variation of resistance through winding, it was observed that

* Some of the results here qnoted as to the influence of stress and of a moderate
rise of temperature were previously arrived at by Dr T. KlemenSid {Sitz.-Ber. Wien*
Ahad. 97, 188S).

FOR ELECTRICAL MEASUREMENTS 441

the resistance of all kinds of wire increased by winding, as would
be expected, the increase being more pronounced for a given
gauge of wire the less the bobbin's diameter. This increase is
due to a mechanical hardening of the wire by strain, and it is
well known that bhe resistance of any metal is less in the annealed
state than in the hardened condition. At the same time an elonga-
tion can hardly be avoided, especially with thin wires, also causing
an increase of resistance. As the gauge of the wires generally was
1 mm., this second cause was of secondary importance.

In the first place we investigated a German-silver alloy which
the firm of Siemens and Halske in Berlin used for standards at
that time. It appeared that the increase of resistance through
winding could amount to 1 per cent., and that the time-variation
during the following months was very considerable; the latter
showed itself always as an increase of resistance. Another re-
markable circumstance is the further increase of resistance
(amounting to a few tenths per cent.) by heating such a wire for
several hours at about 100° C.

It might be supposed that the wire would be annealed by the
effect of the high temperature, and that its resistance would
therefore decrease. But our extensive investigations gave the
important result that heating causes an increase of resistance in
all alloys containing zinc to any considerable amount. On the
other hand, all alloys examined containing no zinc show a decrease
of resistance under the same conditions. The increase of resistance
by winding is also much more pronounced with alloys containing
zinc than with those in which this metal does not occur. All this
seems to point out that in the former alloys changes of structure
go on, which are accelerated by any kind of stress or by variations
of temperature, and always tend to increase the resistance. These
changes of structure also become apparent by the time-variations,
which occur when the resistance coil is left to itself. The latter
observations are in perfect agreement with what was found by
former observers on the time-variation of German-silver. The
interesting result was then arrived at, that the time-variation
would be much accelerated by heating the resistance at a high
temperature, say 100° C, for a few hours. Within two months
after winding, the period in which German-silver varies most,
variations could not be shown within the errors of observation in
wires treated in the manner described. During longer periods.

442

PRACTICAL STANDARDS

say one or two years, variations would still occur, even with an-
nealed German-silver coils. But they hardly reach the tenth part
of those occurring when this process has not been gone through.

The following table shows the results of one of the experi-
ments with two wires of German-silver (60 per cent. Cu; 25"4
per cent. Zn ; 14'6 per cent. Ni), the specific resistance of this

material being 30 microhms per — ^ , and its temperature-coeffi-
cient 0*036 per cent, per degree Centigrade. In both cases the
wire (1 mm. in thickness) was wound on a bobbin of 10 mm.
diameter. The wire marked I. was left to itself after winding,
whereas the wire II. was annealed after winding by heating it to a
temperature of 90° Centigrade during three hours. The resistance
of each was measured at intervals of nearly two months from time

to time.

•

Table II. — Oerman'Silver.

1
[not annealed after winding) \ 11. (annealed after winding)

Resifltanoe

at 20° C.

(in Ohms)

1
1

flO g 1

Date

Bemarks Date

« So

m,m C^ ^^

Bemarks

■

1889

: 1889

13/11.

2-2460

Before winding 1 13/11.

2-2470 ! Before winding |

14/11.

2-2.594

After winding. Increase] 14/11.
through winding, 0*60 |

2-2666

After winding. Increase \
through winding, 0*87

per cent.

i ;

per cent <

15/11.

2-2597

1 15/11. 2-2733

After heating to 90" G.

i| i during three houra;!

increase through heat- !

1 ! ing, 0*29 per cent |

16/11.

598

16/11. 732 I

22/11.

603

22/11. 733 1

: 4/1II.

608

4/I1I.

734

. 19/III.

612

19/III.

729

Temperatiu^ changed

duringthe measurement

6/IV.

615

Time- variation in two 6/IV.

732

Time-variation in two

mouths, 0-09 per cent. ]

'' 1
;i 1

months, practically
nothing

The above table shows clearly that annealing a wire after

winding has a very good effect on the constancy of the resistance.

Quite analogous results were obtained with other alloys

FOB ELECTRICAL MEASUREMENTS

443

containing zinc, e,g.y nickelin, which has been much used for
standards in Germany. The less the percentage of zinc, the less
became the above-mentioned variations of resistance.

As these zinc-containing alloys showed themselves so unreliable^
we extended our investigations to other alloys.

A few years ago the firm of Siemens and Halske in Berlin
made use of an alloy on account of its comparatively low tempera-
ture-coefficient (002 per cent, per 1°C.), called patent-nickel.
This was tested in the Reichsanstalt in the same way as the other
alloys. It contains about 25 per cent, of nickel and 75 per cent,
of copper. The experiments gave the following results : —

1. The variations of resistance by winding are considerably
less for this material than with alloys containing an appreciable
amount of zinc.

2. Heating produces a decrease of resistance ; this decrease
is sometimes greater than the increase by winding, because the
hardening produced by drawing the wire is also diminished. There
is, however, not the slightest evidence for a change of molecular
structure.

Materials with such properties are evidently much more
appropriate for the construction of standard resistances. It was,
indeed, found, by comparison with mercury resistances, that coils
of ** patent-nickel," which had been, as we call it, artificially aged
by heating at about 140'' C, have remained constant for two years
within a few thousandths per cent. In the following table, for
instance, are stated the differences of two patent-nickel standards
of 1 ohm (No. 22 and No. 23), as observed at different times : —

Table III.

Diflerenee of

Difference of

Date

' No. 23 -No. 22

Date

No. 23 - No. 22

in Ohms

1S90

■

1891

21/VL

, 0-00012

9/V.

0-00009

M/VII.

11-6

30/VII.

26/XI.

11-5

1892

1891

l/III.

8-7

29/1.

2/V.

21/V.
19/VII.

9-6

9-4

444

PRACTICAL STANDARDS

On the other hand, from comparisons of the sum of No. 22 and
No. 23 with four diflFerent mercury standards (I., II., III., and IV.,
each of about 1 ohm), I proved that the absolute values of the two
standards had remained constant within the errors of observation,
as the following numbers show : —

Table IV.

Date

Values of No. 23 at 20° G. deduced from Comparisons of

No. 22 + No. 23 with the Mercury Standards ■

II. + III.

I. + 111.

III. + IV.

I. + IV.

I. + II. ;

Nov. 1890...
Feb. 1891...
June 1892...
July 1892...

0-99989
85
89

0-99990
88
88
89

0-99986

0-99988

1
t

0-99986 '
89

The " patent-nickel " would therefore be a material well fitted
for resistance-coils. A large number of alloys were also examined,
consisting of nothing but nickel and copper. An alloy containing
about equal amounts of each metal was found to have an ex-
tremely small temperature-coefficient, the latter amounting to
about 0*003 per cent, per degree Centigrade as against 0*02 per
cent, for patent-nickel. Unfortunately, however, the thermo-electric
effect of these alloys against copper is very high. For the alloy
just mentioned (consisting of 50 per cent. Cu and 50 per cent. Ni,
called " constantan ") it amounts to nearly forty microvolts per
degree Centigrade, considerably surpassing the thermo-electro-
motive force of most of the usual thermo-j unctions, like iron —
German-silver, for instance. This high thermo-electric effect
evidently constitutes a considerable drawback, as the connecting
pieces have to be made of copper.

On the whole our experience has led us to the conclusion that
for standards such alloys do best which, besides copper and nickel,
also contain manganese. A few years ago Mr Weston, of Newark,
U.S.A., discovered that alloys containing manganese possess a
very small temperature-coefficient, and that it is even possible
to obtain metals with negative temperature-coefficient in this
way. I am not aware how far this discovery has been practically
taken advantage of in the United States. After hearing of

FOR ELECTRICAL MEASUREMENTS

445

Weston's observations the further investigation of manganese
alloys was taken up at the Reichsanstalt, and we obtained very
interesting results.

The alloy, which is now being regularly manufactured and
brought out under the name of manganin, consists of 84 per cent,
of copper, 12 per cent, of manganese, and about 4 per cent, of
nickel. As the observations made by me for the last two years in
the Reichsanstalt have shown, this is a most appropriate material
for standard resistances.

The general character of the resistance-variations of manganin
with temperature may be best understood from the diagram (fig. 1),
in which temperatures are taken for abscissae, and the resistances
of a hundred-ohm standard are plotted as ordinates. In this case
up to 40'' C. the temperature-coefficient is positive, the absolute
value, however, being very small, as the following table of the
mean linear coefficients between the temperatures stated in the
first column shows: —

Table V.

Range of
Temperatare

Mean
Linear T.C.

Range of
Temperatare

Mean
Linear T.C.

10" to 20'
20 „ 30
30 „ 36
35 „ 40
40 „ 46

+ 26xlO-«
+ 14 „
+ 4 „
+ 3 „

+ 1 „

, 46'' to 50°
1 50 „ 65
66 „ 60
60 „ 65

-lxlO-«
-2 „
-4 „

-5 »

For most purposes the variability of resistance with tempera-
ture may now, indeed, be quite neglected. As a matter of fact,
very elaborate and sensitive methods are required to demonstrate
the existence of any temperature-coefficient at alL On raising
the temperature beyond 50*" C. the resistance attains a maximum,
thence to diminish again. In this latter part of the curve we
therefore actually have a negative temperature-coefficient.

In order to show that at the same temperature the resistance
always returns to the same value — in other words, that there is no
hysteresis in the relation between those two quantities, some
points of the curve were determined with temperatures descending
from 70'' C, whereas others were obtained with ascending tem-
perature. This process was repeated several times. The spots

446

PRACTICAL STANDARDS

marked ^ correspond to descending, the spots marked ^ to
ascending temperature, and the points belonging to the same
series of observations have the same sign. All points are ex-
tremely close to the same continuous curve, and it is quite obvious
that this curious behaviour is a constant ph}rsical quality of the
material. Of course such a resistance-coil must have been
artificially aged before the beginning of the observations ; it vras
indeed heated during five hours at a temperature of about 140*^ C.
Otherwise, as I mentioned before, a progressive process of decrease
of resistance through annealing would superpose upon the regular

Fig. 1.

f00,030

/mow

^ /00.000
3.%S30

f i

fO'

ZO'

30'

90-

JO'

eo'

70' e

Tcjnpei^aXujie'

variation of resistance according to the curve. It is true that this
maximum resistance-point does not always occur at exactly the
same temperature for wires of dififerent size ; it is well known that
the electrical constants of all resistance alloys change slightly with
the gauge of the wire. But it is also true that the maximum
resistance-point of manganin of a thicker size — say 1 mm. —
occurs, as a rule, at about 30° C, and so at ordinary temperatures
the temperature-coefficient is even less than for this particular
specimen of wire.

The material is very soft, and can be drawn to the finest
gauges; but it must not be annealed in free air, because the

FOB ELECTBICAL MEASUREMENTS 447

manganese then would oxidise, and the qualities of the material
would be altered. Thus it is not possible to buy, for instance, a
wire, say 1 mm, thick, and to draw it down to the required gauge
without taking proper precautions.

In concluding, I will very briefly refer to the construction of
our wire-standards — for instance, to a standard of one ohm
(fig. 2)*.

Fig. 2.

The double silk-covered wire is wound on a metal bobbin, b b,
which is previously covered with a thin piece of silk, coated with
shellac varnish and heated, in order to secure good insulation.
The bobbin can be screwed to the ebonite disc d, but it is not
fixed to it before the accurate adjustment of the resistance. The
resistance of the wire must be 1 — 2 per cent, larger than one ohm
to begin with ; then it is wound on the bobbin, heavily coated
with shellac vamiah, and heated in an air-bath at a temperature
of 140° C. during about five hours. By this procedure we obtain,

* See Dr K. FeaBBOer, Ztiltchrift JBr Inttrumrntenkuttde, 1890, p. 6.

448 PRACTICAL STANDARDS

as already stated, very constant resistances ; liirther, the shellac is

melted at this temperature, and becomes after cooling a hard,

highly insulating mass, which at

the same time protects the wire ^" '■

against any chemical action.

To the ends of the wire are
previously soldered with silver two
small copper rings. The exact
adjustment is made by means of
a fine wire-resistance,/, of 100 —
200 ohms put in multiple arc with
the thick wire. A comparatively
great length of this fine wire
corresponds to a veiy small change
of the whole resistance, and so it
may be easily adjusted to a few
thousandths per cent. Then the
small rings at the ends of the two
wires are screwed together and
soldered to the stout connecting
pieces, pp. A wide brass case,
c c, serves to protect the wire.

In taking observations the re-
sistance is put in an oil -bath
(fig. 3); the temperature of the
wire may then easily be deter-
mined, and besides that, there cannot exist any thermo-electric
force between the two solderiugs. It is a matter of &ct that the
thermo-electric force of manganin against copper is very small
indeed; it amounts only to l^S microvolt for 1°C. ; the corre-
sponding value for other resistance materials is generally 20 — 30
microvolts. We see that even in this respect the manganin is
much preferable.

The construction of standards of O'l ohm, and, on the other
hand, of 10, 100, 1000 ohms and more, is essentially the same as
described. Of course there is no multiple arc to those of 10 ohms
and more.

As to the constancy of manganin resistances I will quote a
few figures. Table VI. refers to a resistance which is used to
determine the electromotive force of the standard Clark cells with

FOR ELECTRICAL MEASUREMENTS

449

the silver voltameter. Thus very often (more than fifty times) a
current of about one -half of an ampere was passed through it for
one hour each time. At 18** C. I found the following values : —

Table VI.

Tx ^ Resistance in

Ohms

Date

ResistaDce in
Ohms

6/1. 1890
15/IV. „
12/n. 1891

2*9998
99

22/VII. 1891
9/11. 1892
17/VIL „

2-9996

98
96

Again, in the following table are stated in microhms the
differences in the resistance of four manganin standards (No. 148
to No. 151) of one ohm. The numbers marked * were observed
by Drs Ereichgauer and Jager, using Eohlrausch's differential
galvanometer method, the others by myself, using a Wheatstone's
bridge arrangement.

Table VI I.

Deo. 1891 •

Feb. 1892

July 1892 •

July 1892

Ohms

Ohm»

No. 148— No. 149

- 121 X 10-«

- 124 X 10-«

-117xlO-«

— .

No. 150

-135

-136

-129

No. 161

- 80

- 79

- 86

Na 149— No. 160

- 14

- 16

- 12

-16

—No. 161

+ 41

+ 39

+ 31

—

No. 16a-No. 151

+ 66

+ 63

+ 43
— «

+48

Measurements were also made of these standards shortly affcer
their construction in July 1891, but not with quite the same
accuracy as the later ones. Anyhow, they show, in connexion
with numerous comparisons of the four coils with other standards,
which were checked by mercury resistances, that the manganin
coils were constant for the space of one year within a few
thousandths per cent.

The patterns referred to are intended to be standards of
resistance. On the other hand, resistances of 0*01, 0001, and
even 0*0001 ohm are used for measuring large currents up to a few
thousand amperes by compensating the potential difference which

B. A.

460 PRACTICAL STANDARDS

the current itself produces in flowing through the resistance.
These resistances consist of manganin plates, which are soldered
with silver to stout copper bars. The dimensions of the plates
are chosen in such a manner that the value of the resistance is
too small to begin with, and the definitive adjustment is arrived
at by boring small holes in the plates ; the latter are again coated
with varnish in order to protect them against any chemical action
of the oil, and so on. For uncovered wires, as they are used, for
instance, in bridges, or in technical resistances, the manganin is
perhaps not so appropriate as the alloys commonly us^. For
all other resistances, however, we think it is the best alloy
hitherto known, because it facilitates the electrical measurements,
and brings them to a higher degree of accuracy than was formerly
attainable.

Appendix V.
On the Clark Cell*. By Dr Kahle.

In connexion with the Report of the Committee I beg to call

your attention to some researches I made, by order of the

Physico-Technical Institute at Berlin, on Clark cells. The time

tis too short for communicating my measuring methods and results

in full extent ; I can only give you a short summary of the chiet

^points.

I used in my researches Lord Rayleigh's H form, the positive
electrode being mercury once distilled, the negative an amalgam
containing ninety parts of mercury and ten parts of zinc. The
last was poured into the vessel as a hot liquid, and solidified on
the bottom. The paste, which covers the positive electrode, is
made by grinding together mercurous sulphate, mercury, and a
mixture of crystals and concentrated solution of zinc sulphate.
No heat was used in preparing this paste. The mercurous sulphate
was bought, and contained, according to chemical analysis, no
foreign ingredients. The zinc sulphate was made basic ^y
boiling with rods of metallic zinc; after cooling, the dissolved
oxide of zinc precipitates, and with it the oxides of the metals
more negative than zinc. For oxidising the ferrous sulphate,
which is always present in commercial zinc sulphate, a small

* For a detailed account see ZeiUchrift fUr Instrumentenkundef April 1892, and
Eleetroteeknische ZeiUchrift, Heft SO, 1892.

FOR ELECTRICAL MEASUREMENTS 451

■

current was sent between two platinum electrodes through the
boiling basic solution; the ferrous sulphate was changed by the
generated oxygen into ferric oxide, and fell out. The iT cells set up
with these materials showed a great agreement in their e.m.f/s.
I never found a difiference greater than x^^xyth of a volt between
the E.M.F. of any two of them.

The next point I studied was the influence of the impurities
in the different materials composing the cell on the E.M.F.,
because on the one hand it is well known that the smallest
impurity of the mercury alters very distinctly the E.M.F., and on
the other hand the mercurous sulphate I bought never contained
impurities of a remarkable amount, and different samples always
had the same qualities; I only investigated, as the most im-
portant matter, the impurities of the zinc and its sulphate. It
was found that the foreign ingredients of the zinc sulphate are of
very little importance, and that only the presence of free acid in
the above-described cleaning process, the result of boiling with
metallic zinc, alters the E.M.F. in a considerable degree. Among
the impurities of the zinc only those caused by metals more positive
than zinc are of importance ; the zinc may contain considerable
quantities of the negative metals without any alteration of the
E.M.F. I conclude that the impurities of the zinc are of greater
importance. If we use it in the form of rods amalgamated on the
surface, it seems to be a great advantage to dissolve the zinc in
mercury, using it then as a solid amalgam.

The following are the values I found by a great number of
observations for the temperature-coefficient of different forms of
cells, measuring between 10° and 30° C. in rising and decreasing
temperature. The figures here given are the mean values of some
cells of the same form, treated in the same manner : —

29—2

452

PRACTICAL STANDARDS

Table VIH.

Form of the cell

Temperatare-coefficient

Mean differ-
ence between

calculated
and observed

values of

S.M.F.

Mean differ-
ence between
thes.M.F.of
the different
cells and that

of the mean
of several fl*
cells used as

standards

The unit being rrvVvv^b
of a volt

H cell set up in Lord
Rayleigh's manner

0-000812 +0-000013 (<- 15)

+ 3

H cell, the paste cover-
ing both electrodes

OKX)0774 +0000020 {t - 15)

1
+ 7

A new form for re-
search purposes,
' the paste covering
, both electrodes

0-000791 +0-00001 7 (f - 16)

+9 1

The cell issued hither-
to bv the German
Reichsanstalt

0O00806+0-000006 (< - 15)

-29

The mean value of the temperature-coefficient, therefore,

would be

0000796 + 0000014 {t - 15).

Lord Rayleigh has given the following values for the two
different cells he investigated: —

+ 0-000827 + 0000018 {t - 16),
+ 0000740 + 0000016 {t - 15),
the mean being

+ 0000783 + 0-000017 {t - 15).

I suppose, for practical purposes, the values found by Lord
Rayleigh and by me are identical.

The most important matter is to obtain the absolute term of

FOB ELECTRICAL MEASUREMENTS 453

the E.M.F. For the purpose I used a measuring arrangement
similar to Lord Rayleigh's. The current, which produces on the
terminals of a known resistance a pressure equal to that of the
Clark cell, was obtained by the silver voltameter. It was found
that the same current deposits the more silver the more oxide of
silver is dissolved in the solution of the nitrate. I made a solu-
tion of nitrate crystals, and boiled a part of it a long time with
oxide of silver ; the deposit obtained with this basic solution was
about Tj^xstha greater than that with the original solution.
Therefore, using a certain number for the equivalent of silver,
there will be a little uncertainty of some parts in 10,000 in
measuring currents by the deposit of silver. Now, as first shown
by Professor Schuster, and also proved by me by a good deal of
experimenting, the deposit, when the voltameter is in vacuo, is
about four parts of 10,000 greater than in ordinary air. But the
absolute value of the E.M.F. is not touched by this fact, because
making the electrolysis in this manner one has certainly to take
a greater figure for the equivalent, and therefore the ratio between
the unit and the measured amount of current remains the same.
The following figures are given by taking one ohm =» 1*068 S.U.,
and by assuming that a current of one ampere strength deposits
in an hour 4*0259 grammes ; the last figure exceeds that given by
the Board of Trade only by six parts in 100,000. I found by
some thirty experiments the E.M.F. of the H cells, set up with
clean materials in the above-described manner, as 1*4332 volt at
IS^'C, and am sure that, when using the same arrangement of the
silver voltameter, this value will be right by five parts in 10,000
if the equivalent is certain to this extent. If I express the
value given by Lord Bayleigh for the cells of the original Clark's
form in this unit, it is 1*4346 volt at 15"". Lord Rayleigh finds
the E.M.F. of this H cell a few ten-thousandths of a volt greater
than that of the old form. Therefore it would, perhaps, be 1*4350
volt at 15^

Recently Mr Qlazebrook has made a new determination, and
finds the E.M.F. of the original Clark cell, in the above fixed units,
to be 1*4342 volt at 15''. He has also compared H cells set up by
me, and now brought to EIngland, and finds their E.M.F. smaller

by iTT^iHT^^^ ^^ ^ ^^^^ ^^^^ ^^^^ ^^ ^^^ original form. Therefore
the E.M.F. of the H cell is 1*4338 volt at 15"". This last value
and that found by me are in good agreement. It is to be noted

454 PRACTICAL STANDARDS

that the anodes and cathodes in my voltameters are much smaller
than those in the English ones.

It may be mentioned here that the mean E.M.F. of four H
cells set up in the same manner as before, but containing, in
accord with Professor Carhart's directions, a solution of sulphate
of zdnc, saturated at 0"", was found to be 1*442 volt at 15°, using the
same units as above.

It only remains to give some directions on the best form of
Clark cells. I suppose it will be good to distinguish such cells
which are to remain as standards in the laboratories and are used
by their maker, and such as are to be used for practical purposes.
These do not need to have the same degree of accuracy, but they
must be able to be carried about. In the Board of Trade memo-
randum the original Clark cell is adopted as the standard; but
I think Lord Rayleigh's H form gives more accuracy and is easier
to set up. In the old form not all the parts of the zinc rod are
in saturated solution, and therefore the value of the E.M.F. will
be a little uncertain. Another disadvantage is that parts of the
zinc rod may fall down in the mercury, and will so produce a
considerable variation of the E.M.F. On the other hand, the
electrodes of the H form are always in concentrated solution, and
there is no possibility of parts of the negative electrode coming
over to the positive one. I have set up about sixty H cells, and
have found no difficulty, when using carefully cleaned materials,
to keep the difference of the E.M.F. of the single cells under a
ten-thousandth of a volt.

To construct cells for practical purposes which will stand
carriage, the most simple way is to separate the two electrodes by
a porous wall. I can show here such a cell of a form constructed
by Dr Feussner, and issued hitherto by the German Reichsanstalt.
The positive electrode is an amalgamated platinum plate with
the surrounding paste in a porous vessel of clay. The zinc rod
forming the positive electrode is on the upper part protected by
a glass tube ; the lower part is blown rectangularly and covered
with crystals of sulphate of zinc. The whole glass vessel is filled
with a concentrated solution of this salt. The E.M.F. of such
cells is about y^y^ths of a volt higher than that of the H cells.
The agreement of different cells of this form is very sufficient
for practical purposes ; the difference between the E.M.F. is always
smaller than 777^ th of a volt. The only disadvantage of this

FOR ELECTRICAL MEASUREMENTS 455

form is that its E.M.F. does not follow quickly the alterations of
temperature; but I suppose one could improve that by diminishing
the size of the cell.

We endeavoured in the Reichsanstalt to make Lord Rayleigh's
H form transportable on account of its good qualities, and to do
this without introducing foreign substances as porous wall& The
investigations in this direction are not yet finished, but I am able
to exhibit a cell which was constructed for this purpose and seems
to be good. The positive electrode is formed by an amalgamated
platinum plate fixed on a wire of the same metal, which is melted
in the bottom of one of the two tubes forming the vessel. The
negative electrode is formed by the 10 percentage zinc amalgam
solidified on the bottom of the other tube, and also connected
with a platinum wire melted in the glass. The whole vessel is
filled with paste and closed by a glass stopper perforated by a
thermometer, of which the bulb is within the vessel. Such a cell
can be turned without any danger, and is suitable for transport.
The E.M.F. is about -nrW^^ ^^ ^ ^^^^ smaller than that of the H
cells; the disagreement between the e.m.f.'s of different cells
constructed in similar manner does not exceed -nr^^^^ ^^ ^ ^^^^^
But before using such cells for practical purposes, they must be
observed for a longer time till one is sure that their behaviour will
not be altered by age.

These are the principal results obtained recently in the Reich-^
sanstalt on this matter. Some of them will be already known
here ; but I hope to have given new proof that the Clark cell is a
very accurate standard for E.M.F., and a good measuring instru-
ment for practical purposes.

Appendix VI.

On the Values of certain Standard Resistance Coils*,
By R. T. Glazebrook, F.R.S.

In the Report of the Committee for 1890 it was stated that
during the early part of the year small changes had taken place
in the values of some of the old B.A. standards, which had been
subjected to a very low temperature early in that year. These
coils have been compared together at intervals since that date,
and the following tables will show that at the temperatures given

* See also Report for 1908.

456

PRACTICAL STANDARDS

there is no indication of any further change. The diflFerence
between the coils and the standard coil flat is given in bridge
wire divisions. The value of one bridge wire division is about
•00005 ohm.

The first few lines in Tables IX. and X. give the values of the
differences observed in 1890. These are followed by those of the
more recent observations which were taken at a temperature of
about le^'C.

These observations are sufficient to show that there has been
no appreciable change in the relative values of these coils.

The observations on the legal ohm standards given in Table X.
lead to the same conclusion.

Table IX. — Showing the Differences between the Platinum
Silver B.A. unite in 1890 and in 1892.

Date

Temperature

F]a%-F

Flat-G

Flat-H

May 1890

June 1890

Aug. 1890

14-4'' C.

16-5

-3-5
-3-2
-3

21-6
22-2
22-6

18-3
17-8
18

July 23, 1892...
July 24, 1892...
July 27, 1892...

16
16
16-2

-3-7
-3-5
-3-2

21-6

18-4
18-3
18-3

Table X. — Giving Differenoes in Bridge Wire Divisions
between the Legal Ohm Standards and Flat

Date

Temperature

;^ 100 -Flat

;^ 101-Flat

Aug. 1890

Nov. 1890

Jan. 1891

July 26, 1892...
July 27, 1892...

15 ^'C.

11-6

16-3

16-1

216-7
217-2
216-1
215-6
216

206-6
206-7
206-8
205-8
206-3

In November 1890 two new coils marked J and K were made
of platinum silver wire, supplied by Messrs Elliott Bros. The
wires of the coils are wound in a loose spiral inside the case, being
secured at intervals by silk threads to an ebonite central stem.

FOR ELECTRICAL MEASUREMENTS

457

They are thus in contact with air only, and there is no paraffin
inside the case. It was hoped in this way to secure freedom from
the strains set up by changes of temperature in the paraffin, which
appeared from the results given in the Report for 1890 to have
some connexion with the changes of resistance there described.
The results are shown in fig. 4. From the six observations
for each coil there recorded it is clear that there has been no change
during the time for which the observations have lasted.

Fig. 4. Showing the values of the 6. A. units Flat, J and K^ from
observations between November 1890 and July 1892.

/O //

/J /4

M9X9J^JO//a/S/4/S/>S/fJl9]»ZO

The vertical divisions are -0005 B. A. unit.

The horizontal divisions are 1"* C.

F]at=» IRA. unit at 14*8\ Temp. Coeff. -00027 (Fleming's observations,
1876).

J^ 1 B. A. unit at 14^ Temp. Coeff. -000260.

Jr» 1 B. A. unit at 14•15^ Temp. Coeff. -000261.

Two observations were made at a much lower temperature, and
these would seem to indicate a slightly less temperature-coefficient
than is shown by the observations between 10^ C. and 20° C.

The values of the two new coils in B.A. units between tem-
peratiures of 10° and 20° are given by

•/=rl + •000260(^-14).

if=l +000261 («- 1415).

We may thus conclude that during the past two years there
has been no relative change in the values of the platinum silver
unit standards of the Association.

Four of these are the original standards made in 1865-67 ;
two others were made by Messrs Elliott Bros, in 1885, and the last
two by the assistant at the Cavendish Laboratory from wire
supplied by Messrs Elliott Bros, in 1889.

458

PRACTICAL STANDARDS

Appendix VII.

On the Standard Condensers of the Association, and on certain
Resistance Coils. By R. T. Glazebrook, F.R.S.

The Report of the Committee for 1890 contains as an appendix
a very full account of the tests on the air condensers belonging
to the Association. It was there stated that while the insulation
resistance of No. I. was very high, that of No. II. was not com-
pletely satisfactory ; No. II. was therefore taken to pieces and set
up afresh. Its capacity and also that of No. I. were redeter-
mined, using the commutator method described in the previous
paper.

The following values were found : —

Table XI.

Date

Capacity iu
Miorofarads

Mean of each
Series

Dec. 23,
Dec. 29,
Dec. 23,

Dec. 23,

Dea 29

>»

1890 ..

»> ••

1890 ..

>»
II

1890, af

i»
If
11

Condi

SNSER No. I.
•021059

•021050

•021395
•021389

•021390

•021052
•021046

•021044

NSBR No. II.

•021396
•021392
•021399

•021365
021399
•021403

•021381
•021409
•021389

CONDE

temoon

>i
I)
11

Mean of the whole, "021391 microfiu^.

The different values in each series correspond to different rates
of revolution of the commutator.

The value found originally for the capacity of No. L was
'021024 microfarad; it would appesur therefore that it may be
slowly increasing; the capacity of No. II. has been changed by
being taken to pieces from '022515 to '021391.

FOR ELECTRICAL MEASUREMENTS

469

The two condensers were also compared, directly assuming the
value of No. I. to be '021050 ; that of No. II. was found to be
'021390 — practically the same value as that given by the com-
mutator.

As a further check on the values a mica condenser was com-
pared with the two in the usual way. The values found were : —

Table XII.

Date

In Terms of I.

In Terms of 11.

■

Dec. 29, 1890 i 5017

Jan. 1,1891 5013

•5013
•5012

Table XIII.

Nominal
Valae

Time of

Charging, in

seconds

Value in
Terms of I.

Value in
Terms of n.

•05

•05022
•05072
•05080

•05 +

0
2
5

10
20

•05055
•05106
•05109

•2

1
•1988 '1981
•1999 -2002
•2007 ■ -2009
•2010 ! -2013
•2012 !

•2 +

0
2
5

10
20
30

•2003
•2018

•2027
•2033
•2039
•2046

•5

0
2
5

10
20

•5032
•5058
•5078
•5081
•5092

460

PRACTICAL STANDARDS

In March 1892 the insulation of the condensers was tested
by the Secretary and Mr A. S. Bowley ; they were both found to
lose rather less than ^ of their charge per minute. A divided
condenser, Elliott No. 144, No. 3, was compared ; the results are
interesting as showing the effect on the capacity of the time of
charging, and are given in Table XIII.

Jhe first observation in each case marked as 0" was made
by connecting the galvanometer, and then momentarily making
the battery circuit for a very small fraction of a second. The
observations on the '2 condenser show that there has been no
appreciable change in the relative value of Standards I. and II.
The observations throughout are accurate to about one part in
10,000.

During the process of the work Mr Bowley compared several
of the resistance boxes of the Association together. As these are
used as standards in many experiments it will be useful to put the
results on record.

Box Elliott 1253 is a Wheatstone's bridge box of platinum
silver in legal ohms, said to be right at IT^'C. Assuming that
the two 1000-ohm coils of the bridge are equal, and the experi-
ments showed no appreciable difference, the following values were

Table XIV.

Nominal Value

Value found

EUiott 1825

10,000

10,012

20,000

20,024

dO,000

90,034

40,000

40,049

Nalder 1870

100,000 No. 1

100,042

,, 2

100,044

n 3

100,050

„ 4

100,034

„ 5

100,042

« 6

100,042

„ 7

100,052

„ 8

100,032

„ 9

100,047

» 10

100,052

FOR ELECTRICAL MEASUREMENTS 461

found for certain coils in terms of a nominal 10,000 ohms taken
from 1253.

The temperature of all the coils was about IS'G"" C.

Thus the box Elliott 1825 is right at about 4° below the box
1263, while the box Nalder 1870 is right at about 1*5° below 1263.

Appendix VIII.

On the Valves of certain Standards of Resistance and Electro-
motive Force sent from Berlin for Comparison with the
British Association Standards. By R. T. Glazebrook,
F.R.S.

Towards the end of July Dr Lindeck, of the Physikalische
Electrotechnische Beichsanstalt at Berlin, brought to Cambridge
three resistance coils of nickel-raanganese-copper alloy in order
to compare them with the British Association standards. Ex-
periments to determine the value of these coils in terms of the
resistance of mercury are in progress at Berlin ; when they are
completed a comparison of the British Association units with the
mercury standards of the Beichsanstalt will become possible.

A fourth coil, constructed for the Berlin Beichsanstalt by
Messrs Elliott Bros., was also tested. Table XY. contains the
results of the comparison.

In the experiments marked thus * a considerable length,
200 cm., of the bridge wire was used. An error of 1° C. in the
temperature of this would produce an error of "00003 in the result.
In reducing the results the temperature of the bridge has been
taken as 18° C, the value given by a thermometer laid alongside the
wire on July 28. In the other experiments a coil of 100 ohms
was put in multiple arc with the Berlin standard, and the difference
of the combination and the British Association units was found.
The length of bridge wire was very small, so that an exact know-
ledge of its temperature was not wanted, while the 100-ohm coil
was known with all the necessary accuracy.

Another set of comparisons, leading to results which do not
differ from the above by more than '00003 British Association unit,
was made by Dr Lindeck, using apparatus he had brought from
Berlin ; but a complete discussion of the whole can best be given

462

PRACTICAL STANDARDS

when the values of the Berlin coils in terms of the mercury standard
have been found.

Dr Lindeck also brought four Clark cells. Of these two, Nos.
69 and 71, were of the pattern described by Dr Kahle (Electro-
technische Zeitschrift, July 22, 1892 ; Zeitachrift fur Instrumenten-
kunde, April 1892). The mercurous sulphate is enclosed in a
porous pot. The zinc dips among the crystals of the zinc sulphate.

Table XV.

No. of CoU

Date

Temp. Value in

in ** C. B. A. units

Mean Value
of Series

Wolff, 150

July 25
,, 27
» 25
„ 26
„ 27

16-5

16-3

17-1

1-01118*

1-01116*

1-01119

1-01120

1-01118 at
16-7°

Wolff, 139

July 26

,. 27
„ 25
„ 26
„ 27

16-5
16-9
16-5
16-3
17-1

1-01110*
1-01112*
1-01 110
1-01114
1-01110

101112 at
16-7'

Wolfl^ 147

July 26

., 27
„ 26

„ 26

„ 27

16-6 1 1-01112*
16-9 1-01112*
16-5 1-01113
16-3 1-01112
17-1 ; 101115

1-01113 at
16-7"

Elliott, 250

July 26

« 27
„ 28
„ 28

17 1-01107*
17 1 1-01107*
17 : 101108*
17 1-01108

1-01107 at

its end being bent so as to be horizontal. The vertical portion of
the zinc is covered with a glass tube, so that the horizontal part
alone is effective.

The other two, Nos. 12 and 29, were H cells of the pattern
described by Lord Rayleigh.

These cells were compared with the standard at the Cavendish
Laboratory with the results shown in Table XVI., which gives
the differences between the four cells and the standard in hundred-
thousandths of a volt.

FOR ELECTRICAL MEASUREMENTS

463

Table XVI.

Date

YalaeB of Berlin CeU— Standard
in -00001 Volt

July 26

July 27

July 29

Temperature

16rC. 15-9'* C.

15-9'* C.

CeU No. 69

« 71
„ 29

» 12

1
-15 -18
-11 -15
-36 -36
-39 1 -38

-14

-36
-39

On July 29 cell No. 71 had been taken away. Dr Eahle kindly
determined the differences between No. 69 and each of the other
cells before they left Berlin. Values for these same differences
can be obtained from Table XVI.

We thus get Table XVII.

Table XVII.

Differences between Cell No. 69 and the others
sent from Berlin

Date and Place of Observation

No. 71

No. 12

No. 29

July 19, 1892)

„ 20 „ • Berlin ■

»i 21 „ J

-2
-4
-4

29
28
25

29
25
26

July 26, 1892) (
„ 27 „ [ Cambridge... \
„ 29 „ J I

-4
-3

21
19
22

24
20

Thus the relative values of the cells as found at Cambridge are
practically the same as those found at Berlin. Moreover, taking
the E.M.F. of the Cambridge standard as 1*4342 volt at IS^'C,
that of the Berlin cells, with porous pots, is 1*43405, and of the
Berlin H cells about 1*4338.

464 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

The value actually found by Dr Eahle for the E.M.F. of the
porous pot cells is 1*4336 volt, so that the agreement is satis-
factory. In all the above it has been assumed that the resistance
of 106*30 centimetres of mercury is 1 ohm, and that the amount
of silver deposited per second by a current of 1 ampere is
•001118 gramme.

The H form of cell in all cases examined at Berlin has a
slightly lower e.m.f. than those with the porous pots, the
difference being about -0003 volt.

TWENTIETH REPORT— NOTTINGHAM, 1893.

APPENDIX PAOB

I. Supplementary Report of the Electrical Standards Committee

of the Board of Trade 467

II. Experiments on the Effects of the Heating produced in the

Coils by the Currents used in Testing. By R. T. Glaze-
brook 476

III. On Standards of Low Electrical Resistance. By J. Viriamu

Jones 478

The work of testing resistance coils at the Cavendish laboratory
has been continued. The coils have all been '' ohms/' as defined
by the resolution of the Committee given in their last Report, and
since adopted by the Board of Trade Committee on electrical
standards in the following form :—

The resistance offered to an unvarying electric current by a
column of mercury at the tempeniture of melting ice 14'4521
grammes in mass, of a constant cross-sectional area, and of a
length of 106'3 centimetres, may be taken as 1 ohm. The relation
between the B. A. unit and the ohm is the following : —

1 B.A. unit = -9866 ohm.

The resolutions adopted by the Committee at Edinburgh were
communicated to the Electrical Standards Committee of the Boaid
of Trade. After consideration the Board of Trade Committee
drew up an amended report, in harmony with the Edinburgh
resolutions, for presentation to the President (see Appendix I.),

The resolutions were accepted at Edinburgh by Dr von Helm-
holtz on behalf of Germany, while in France an official committee
decided last June to adhere to the propositions of the Board of
Trade. Austria and Italy are connected by treaty with Germany
for telegraph purposes, and in consequence adopt the same units.

The Committee have learnt with pleasure from Mr W. H.
Preece, one of the English delegates to the International Congress
of Electricians at Chicago, that the Congress have accepted a series
of resolutions defining the fundamental units practically identical
with the Edinburgh resolutions.

B. K. 30

466 PRACTICAL STANDARDS

Thus these resolutions have now been accepted as a basis for
legislation throughout the British Empire, the whole of Western
Europe, and the United States of America.

The Committee are also informed that the Chicago Congress
have adopted the name " Henry " for the unit of self-induction ;
while looking with favour on this suggestion, they think it desir-
able to postpone definite action until the official report of the
Congress has been received.

In March last M. Mascart wrote to the Secretary asking the
opinion of the Committee as to a name for the standard of resist-
ance defined at Edinburgh. A circular fetter was issued inviting
members of the Committee to express their views on four names
which had been suggested, viz.: "International," "Normal,"
"Etalion," or "Ohm de 1893." After receiving replies to the
circular from twelve members of the Committee, the Secretary
wrote to Professor Mascart to the effect that the number of
members who expressed a preference for the name " International "
was greater than the number declaring in favour of any other
name, but that he thought that the Committee would accept
whichever of the first three suggestions commended itself to the
French Committee appointed to deal with the matter.

During the year Dr Muirhead has remeasured his standard
condenser. He now finds as the capacity of a condenser con-
structed twenty-three years ago to represent "1 microfarad (B.A.
unit) the value '09998 microfarad.

Tests have been made during the year on the 1-ohm and
10-ohm standards of the Association. These are still being con-
tinued. The 100-ohm and 1000-ohm standards have now been
delivered, and the tests will be shortly proceeded with. Some
experiments were made as to the amount of heating in the coils
produced by the current used for testing. These are detailed in
Appendix IL Further valuable information on this point is con-
tained in Mr Griffiths' paper on " The Value of the Mechanical
Equivalent of Heat*."

The Committee think it desirable that they should be in a
position to complete the set of resistance standards of the Associa-
tion, and recommend, therefore, that they be reappointed, with a
grant of £25, that Professor O. Carey Foster be Chairman, and
Mr R. T. Glazebrook Secretary.

♦ Phil. Tram. 1898.

for electrical measurements 467

Appendix L

Supplementary Report of the Electrical Standards
Committee of the Board of Trade.

To the Right Hon, A, J, Mundella, M,P.y President of the

Board of Trade.

Subsequently to the presentation of our former report to Sir
Michael Hicks-Beach, in July 1891, we were informed that it was
probable that the German Qovemment would shortly take steps to
establish legal standards for use in connexion with electrical supply,
and that, with a view to secure complete agreement between the
proposed standards in Qermany and England, the Director of the
Physico-Technical Imperial Institute at Berlin, Professor von
Helmhohz, with certain of his assistants, proposed to visit Eng«
land for the purpose of making exact comparisons between the
units in use in the two countries, and of attending the meeting of
the British Association which was to take place in August in
Edinburgh.

Having regard to the importance of this communication it
appeared desirable that the Board of Trade should postpone the
action recommended in our previous Report until after Professor
Helmholtz's visit.

That visit took place early in August, and there was a very
full discussion of the whole subject at the meeting of the British
Association in Edinburgh, at which several of our number were
present. The meeting was also attended by Dr Guillaume, of the
Bureau International dee Poids et Mesurea ; and Professor Carhart,
of the University of Michigan, U.S. A., who were well qualified by
their scientific attainments to represent the opinion of their
respective countries.

It appeared from the discussion that a few comparatively
slight modifications of the resolutions included in our previous
Report would tend to secure international agreement

An extract from the Report of the Electrical Standards Com-
mittee of the British Association embodying the results of this
discussion was communicated to us by the Secretary, and will be
found in the appendix to this Report.

Having carefully considered the whole question in view of this

30—2

468 PRACTICAL STANDARDS

communication, and having received the report of the suVcom-
mittee mentioned in resolution 14 of our previous Report, we now
desire, for the resolutions contained in that Report, to substitute
the following: —

Resolutions.

1. That it is desirable that the new denominations of standards
for the measurement of electricity should be made and approved
by Her Majesty in Council as Board of Trade standards.

2. That the magnitudes of these standards should be deter-
mined on the electro-magnetic system of measurement with
reference to the centimetre as unit of length, the gramme as
unit of mass, and the second as unit of time, and that by the
terms centimetre and gramme are meant the standards of those
denominations deposited with the Board of Trade.

3. That the standard of electrical resistance should be de-
nominated the ohm, and should have the value 1,000,000,000 in
terms of the centimetre and second.

4. That the i*esistance oflFered to an unvarying electric current
by a column of mercury at the temperature of melting ice 14 4521
grammes in mass of a constant cross-sectional area, and of a length
of 106*3 centimetres, may be adopted as 1 ohm.

5. That a material standard, constructed in solid metal, should
be adopted as the standard ohm, and should from time to time
be verified by comparison with a column of mercury of known
dimensions.

6. That, for the purpose of replacing the standard, if lost,
destroyed, or damaged, and for ordinary use, a limited number of
copies should be constructed, which should be periodically compared
with the standard ohm.

7. That resistances constructed in solid metal should be adopted
as Board of Trade standards for multiples and sub-multiples of the
ohm.

8. That the value of the standard of resistance constructed by
a committee of the British Association for the Advancement of
Science in the years 1863 and 1864, and known as the British
Association unit, may be taken as '9866 of the ohm.

9. That the standard of electrical current should be de-
nominated the ampere, and should have the value one-tenth (0*1)
in terms of the centimetre, gramme, and second.

FOR ELECTRICAL MEASUREMENTS 469

10. That an unvarying current which, when passed through
a solution of nitrate of silver in water, in accordance with the
specification attached to this Report, deposits silver at the rate
of O'OOlllS of a gramme per second may be taken as a current of
1 ampere.

11. That an alternating current of 1 ampere shall mean a
current such that the square root of the time-average of the square
of its strength at each instant in amperes is unity.

12. That instruments constructed on the principle of the
balance, in which, by the proper disposition of the conductors,
forces of attraction and repulsion are produced, which depend upon
the amount of current passing, and are balanced by known weights,
should be adopted as the Board of Trade standards for the measure-
ment of current, whether unvarying or alternating.

13. That the standard of electrical pressure should be de-
nominated the volt, being the pressure which, if steadily applied
to a conductor whose resistance is 1 ohm, will produce a current of
1 ampere.

14. That the electrical pressure at a temperature of 15° Centi-
grade between the poles or electrodes of the voltaic cell known as
Clark's cell, prepared in accordance with the specification attached
to this Report, may be taken as not differing firom a pressure of
1*434 volt by more than one part in one thousand.

16. That an alternating pressure of 1 volt shall mean a
pressure such that the square root of the time-average of the
square of its value at each instant in volts is unity.

16. That instruments constructed on the principle of Lord
Kelvin's quadrant electrometer used idiostatically, and, for high
pressures, instruments on the principle of the balance, electrostatic
forces being balanced against a known weight, should be adopted
as Board of Trade standards for the measurement of pressure,
whether unvarying or alternating.

(Signed) Courtenay Boyle. Kelvin.

P. Cardew. W. H. Preece.

Rayleigh. G. Carey Foster.

R. T. Glazebrook. J. Hopkinson.
W. E. Ayrton.

T. W. P. Blomefield, Secretary,

N<mmher 29, 1892.

470 practical standards

Specification referred to in Resolution 10.

In the following specification the term silver voltameter means
the arrangement of apparatus by means of which an electric current
is passed through a solution of nitrate of silver in water. The
silver voltameter measures the total electrical quantity which has
passed during the time of the experiment, and by noting this time
the time-average of the current, or if the current has been kept
constant the current itself, can be deduced.

In employing the silver voltameter to measure currents of
about 1 ampere the following arrangements should be adopted.
The cathode on which the silver is to be deposited should take the
form of a platinum bowl not less than 10 centimetres in diameter,
and from 4 to 5 centimetres in depth.

The anode should be a plate of pure silver some 30 square
centimetres in area and 2 or 3 millimetres in thickness.

This is supported horizontally in the liquid near the top of the
solution by a platinum wire passed through holes in the plate at
opposite corners. To prevent the disintegrated silver which is
formed on the anode from falling on to the cathode the anode
should be wrapped round with pure filter paper, secured at the
hack with sealing-wax.

The liquid should consist of a neutral solution of pure silver
nitrate, containing about fifteen parts by weight of the nitrate to
eighty-five parts of water.

The resistance of the voltameter changes somewhat as the
current passes. To prevent these changes having too great an
eflfect on the current some resistance besides that of the volta-
meter should be inserted in the circuit. The total metallic resist-
ance of the circuit should not be less than 10 ohms.

Method of making a Measurement,

The platinum bowl is washed with nitric acid and distilled
water, dried by heat, and then left to cool in a desiccator. Wben
thoroughly dry it is weighed carefully.

It is nearly filled with the solution, and connected to the rest
of the circuit by being placed on a clean copper support to which
a binding screw is attached. This copper support must be
insulated.

FOR ELECTRICAL MEASUREMENTS 471

The anode is then immersed in the solution, so as to be well
covered by it and supported in that position ; the connexions to
the rest of the circuit are made.

Contact is made at the key, noting the time of contact. The
current is allowed to pass for not less than half an hour, and the
time at which contact is broken is observed. Care must be taken
that the clock used is keeping correct time during this interval.

The solution is now removed from the bowl and the deposit is
washed with distilled water and left to soak for at least six hours.
It is then rinsed successively with distilled water and absolute
alcohol and dried in a hot-air bath at a temperature of about
IGC C. After cooling in a desiccator it is weighed again. The
gain in weight gives the silver deposited.

To find the current in amperes this weight, expressed in
grammes, must be divided by the number of seconds during which
the current has been passed and by 001118.

The result will be the time-average of the current, if during
the interval the current has varied.

In determining by this method the constant of an instrument
the current should be kept as nearly constant as possible, and the
readings of the instrument taken at frequent observed intervals of
time. These observations give a curve from which the reading
corresponding to the mean current (time-average of the current)
can be found. The current, as calculated by the voltameter, corre-
sponds to this reading.

Specification referred to in Resolution 14.

Definition of the CelL

The cell consists of zinc and mercury in a saturated solution of
zinc sulphate and mercurous sulphate in water, prepared with
mercurous sulphate in excess, and is conveniently contained in a
cylindrical glass vessel.

Preparation of the Materials.

1. The Mercury. — To secure purity it should be first treated
with acid in the usual manner and subsequently distilled in vacuo.

2. The Zinc. — Take a portion of a rod of pure redistilled zinc,
solder to one end a piece of copper wire, clean the whole with glass

472 PRACTICAL STANDARDS

paper, carefully removing any loose pieces of the zinc. Just before
making up the cell dip the zinc into dilute sulphuric acid, wash
with distilled water, and dry with a clean cloth or filter paper.

3. The Zinc SulphaJte Solution. — Prepare a saturated solution
of pure (" pure recrystallised ") zinc sulphate by mixing in a flask
distilled water with nearly twice its weight of crystals of pure zinc
sulphate, and adding zinc oxide in the proportion of about 2 per
cent, by weight of the zinc sulphate crystals to neutralise any free
acid*. The crystals should be dissolved with the aid of gentle
heat, but the temperature to which the solution is raised should
not exceed 30° C. Mercurous sulphate treated as described in 4
should be added in the proportion of about 12 per cent, by weight
of the zinc sulphate crystals, and the solution filtered, while still
warm, into a stock bottle. Crystals should form as it cools.

4. The Mercurous Sulphate. — Take mercurous sulphate, pur-
chased as pure, and wash it thoroughly with cold distilled water
by agitation in a bottle ; drain off the water and repeat the process
at least twice*. After the last washing drain off as much of the
water as possible.

Mix the washed mercurous sulphate with the zinc sulphate
solution, adding sufficient crystals of zinc sulphate from the stock
bottle to ensure saturation, aijd a small quantity of pure mercury.
Shake these up well together to form a paste of the consistence of
cream. Heat the paste, but not above a temperature of 30° C.
Keep the paste for an hour at this temperature, agitating it from
time to time, then allow it to cool ; continue to shake it occasionally
while it is cooling. Crystals of zinc sulphate should then be dis-
tinctly visible, and should be distributed throughout the mass ; if
this is not the case add more ciystals from the stock bottle, and
repeat the whole process.

This method ensures the formation of a saturated solution of
zinc and mercurous sulphates in water.

Contact is made with the mercury by means of a platinum wire
about No. 22 gauge. This is protected from contact with the other
materials of the cell by being sealed into a glass tube. The ends
of the wire project from the ends of the tube ; one end forms the
terminal, the other end and a portion of the glass tube dip into the
mercury.

* See Notes.

FOR ELECTRICAL MEASUREMENTS 473

To set up the Cell.

The cell may conveniently be set up in a small test tube of
about 2 centimetres diameter and 6 or 7 centimetres deep. Place
the mercury in the bottom of this tube, filling it to a depth of,
say, 1*5 centimetre. Cut a cork about '6 centimetre thick to fit
the tube ; at one side of the cork bore a hole through which the
zinc rod can pass tightly ; at the other side bore another hole for
the glass tube which covers the platinum wire ; at the edge of the
cork cut a nick through which the air can pass when the cork is
pushed into the tube. Wash the cork thoroughly with warm
water, and leave it to soak in water for some hours before use.
Pass the zinc rod about 1 centimetre through the cork.

Clean the glass tube and platinum wire carefully, then heat
the exposed end of the platinum red-hot, and insert it in the
mercury in the test tube, taking care that the whole of the exposed
platinum is covered.

Shake up the paste and introduce it without contact with the
upper part of the walls of the test tube, filling the tube above the
mercury to a depth of rather more than 2 centimetres.

Melt some marine glue until it is fluid enough to pour by its
own weight, and pour it into the test tube above the cork, using
sufficient to cover completely the zinc and soldering. The glass
tube should project above the top of the marine glue.

The cell thus set up may be mounted in any desirable manner.
It is convenient to arrange the mounting so that the cell may be
immersed in a water-bath up to the level of, say, the upper surface
of the cork. Its temperature can then be determined more
accurately than is possible when the cell is in air.

In using the cell sudden variations of temperature should as far
as possible be avoided.

474 PRACTICAL STANDARDS

Notes,

The Zinc Stdphate Solution. — The object to be attained is the
preparation of a neutral solution of pure zinc sulphate saturated
with ZnS04, 7H,0.

At temperatures above 30° C. the zinc sulphate may crystallise
out in another form; to avoid this 30'' C. should be the upper
limit of temperature. At this temperature water will dissolve
about 1'9 times its weight of the crystals. If any of the crystals
put in remain undissolved they will be removed by the filtration.

The amount of zinc oxide required depends on the acidity of
the solution, but 2 per cent, will, in all cases which will arise in
practice with reasonably good zinc sulphate, be ample. Another
rule would be to add the zinc oxide gradually until the solution
became slightly milky. The solution when put into the cell should
not contain any free zinc oxide ; if it does then, when mixed with
the mercurous sulphate, zinc sulphate and mercurous oxide are
formed ; the latter may be deposited on the zinc, and affect the
electromotive force of the cell. The diflSculty is avoided by adding
as described about 12 per cent, of mercurous sulphate before filtra-
tion : this is more than sufficient to combine with the whole of the
zinc oxide originally put in, if it all remains free ; the mercurous
oxide formed together with any undissolved mercurous sulphate is
removed by the filtration.

The Mercurous Sulphate. — The treatment of the mercurous
sulphate has for its object the removal of any mercuric sulphate
which is often present as an impurity.

Mercuric sulphate decomposes in the presence of water into an
acid and a basic sulphate. The latter is a yellow substance —
turpeth mineral — practically insoluble in water: its presence at
any rate in moderate quantities has no effect on the cell. If,
however, it is formed the acid sulphate is formed also. This is
soluble in water and the acid produced affects the electromotive
force. The object of the washings is to dissolve and remove this
acid sulphate, and for this purpose the three washings described in
the specification will in nearly all cases suffice. If, however, a
great deal of the turpeth mineral is formed it shows that there is
a great deal of the acid sulphate present, and it will then be wiser
to obtain a fresh sample of mercurous sulphate rather than to tiy
by repeated washings to get rid of all the acid.

FOR ELECTRICAL MEASUREMENTS 475

The firee mercury helps in the process of removing the acid, for
the acid mercuric sulphate attacks it, forming mercurous sulphate
and acid which is washed away.

The cell may be sealed in a more permanent manner by coating
the marine glue, when it is set, with a solution of sodium silicate
and leaving it to harden.

Appendix.

August 12, 1892.

Dear Sir, — ^I am desired by the Electrical Standards Com-
mittee of the British Association to communicate to the Electrical
Standards Committee of the Board of Trade the enclosed extract
from their report made to the Association on August 9, 1892.

I remain, yours faithfully,

(Signed) R. T. Glazebrooe,

Secretary, Electrical Standards Committee
of the British Association.

To Sir Thomas Blomefield,

Secretary, Electrical Standards Committee
of the Board of Trade.

Extract from the Report of the Electrical Standards
Committee of the British Association, Atigust 9, 1892.

The following resolutions were agreed to : —

1. That the resistance of a specified column of mercury be
adopted as the practical unit of resistance.

2. That 14'4521 grammes of mercury in the form of a column
of uniform cross-section 106*3 centimetres in length at 0° C. be the
specified column.

3. That standards in mercury or solid metal having the same
resistance as this column be made and deposited as standards of
resistance for industrial purposes.

4. That such standards be periodically compared with each
other, and also that their values be redetermined at intervals in
terms of a freshly set-up column of mercury.

476 PRAcrncAL standards

It was Airther agreed that these resolutions be communicated
to the Electrical Standards Committee of the Board of Trade.

With regard to the units of current and electromotive force it
was agreed that the number *001118 should be adopted as the
number of grammes of silver deposited per second from a neutral
solution of nitrate of silver by a current of 1 ampere, and the value
1*434 as the electromotive force in volts of a Clark cell at 15"* C.

Dr von Helmholtz expressed his full concurrence in these
decisions, which are, as he informed the Committee, in accord
with the recommendations which have already been laid by the
Curatorium of the Reichsanstalt, as well as by himself before
the German Government.

Appendix IL

Experiments on tiie Effects of the Heating produced in the Coils
by the Currents used in Testing. By R. T. Glazebbooe, F,R.S.

Various circumstances(notably the experiments of Mr Griffiths *)
have made it appear probable that the heating effect in the coils
produced by the current used in making the resistance test might
be sufficient to affect the results of the tests. Some experiments
were made to examine the point directly.

The resistance of a coil of 100 ohms (nominal value) was
measured in the usual way, i.e. by making a Wheatstone's bridge
of four coils whose nominal values were 1, 10, 10, and 100 ohms.
If the coils had been accurate there would have been a balance ; as
it was, one of the 10-ohm coils needed to be shunted, and the
adjustment was made by determining the value of the shunt when
no current passed through the galvanometer.

As the current in the battery circuit was increased by varying
the number of cells this shunt decreased in value, showing that the
effect of the heating was to produce an apparent diminution of
the resistance of the 100-ohm coil. This, of course, is as would
be anticipated; for -^ of the current goes through the 1-ohm and
one of the 10-ohm coils; the remaining -j^ goes through the
10-ohm and the 100-ohm. The rise of temperature will clearly be
greatest in the first 10-ohm coil, and to counterbalance the increase

♦ Phil. Tram. 1893.

FOB ELECTRICAL MEASUREMENTS

477

in resistance produced thereby it becomes necessary to reduce the
shunt.

The following readings were obtained : —

Carrent in Amperes

Shunt in 1000 OhmR

Correoting Factor

•06
•09
•12
•14 •
•15

365
326
♦30
305
29-5

l-'00028
•00031
•00033
•00033
•00034

The value of the 100-ohm is given by taking the product
of the values of the two 10-ohm coils at the temperature of the
observations, dividing by the value of the 1-ohm and multiplying
by a factor representing the effect of the shunt

During the above observations the temperatures remained
steady, but the factor changed from 1 - -00028 to 1 - -00034.
Thus the resistance of the 100-ohm coil apparently changed by
-034-028, or -006 ohm.

The apparatus was not sensitive with a smaller current; the
effect, however, will vary as the square of the current ; and, since
trebling the current produces so small a change, we may infer that
the total effect is itself small.

Another coil gave the following results : —

Current in Amperes

Shunt in 1000 Ohms

Correoting Factor

•05
•09
•12
•14
•15

48
45
43
41
40

1 --000208
•000222
•000233
•000244
•000260

indicating a change in the measured resistance of '0042 ohm on
100 ohms.

It is clear, therefore, that the effect of heating is small, though
appreciable when currents approaching '15 ampere are used.

* Only one observation at this carrent was made ; the others are the mean of
several.

478 PRACTICAL STANDARDS

Appendix IIL

On Standards of Low Electrical Resistance. By J. Viriamu
Jones, Principal and Professor of Physics in the Universily
College, Cardiff.

The preparation of standards of low electrical resistance of from
•001 to '0001 ohm seems to be a matter of some importance at the
present time. These standards are already in request among
engineers, and it becomes of interest to consider how they may
best be measured to a percentage accuracy comparable with that
with which the standard ohm is known.

Such standards of low resistance may be derived by potentio-
meter methods from the standard ohm by a series of downward
steps. But this is from one point of view roundabout. The
method of measuring the ohm that seems in all its details most
accurate is that of Lorenz. In this method the ohm itself is derived
from the measurement of a small resistance. It is simply going
up and down again to prepare from the ohm so derived the required
small resistance standards, and it is more direct and more accurate
to measure the latter directly in absolute measure.

" In Lorenz s method a metallic disc is made to rotate in the
mean plane of a coaxial standard coil. Wires touching the centre
and circumference of the disc are led to the ends of the resistance
to be measured, and the same current is passed through this
resistance and the standard coil. The connexions being rightly
made, we may by varying either the rate of rotation of the disc or
the resistance measured so arrange matters as to have no change
of current in the circuit of the disc and wires joining it to the ends
of the resistance, when the direction of the current through the
resistance and the standard coil is changed. When this arrange-
ment is effected there is a balance between the electromotive force,
due to the motion of the disc in the magnetic field of the current
in the standard coil, and the difference of potential at the ends of
the resistance, due to the current traversing it. If this adjust-
ment be made we will say that the apparatus is in an equilibrium
position*."

* Vide PhiL Trans. 1891, A. p. 2, **0d ihe Determination of the Spedfio
Resistance of Meroniy in Absolute Measure."

FOR ELECTRICAL MEASUREMENTS 479

If JIf ss coefficient of mutual induction of standard coil and
circumference of disc,

n = rate of rotation of disc (number of revolutions per
second),

J2 = resistance,

7 = current through standard coil and resistance,
then in an equilibrium position

Jf«7 = J27,
or J2 = Mn.

I do not think that electricians have as yet realised the accuracy
and ease with which absolute measurements of resistance may be
made by this method. The absolute measurement involves
measuring first the coefficient of mutual induction of the standard
coil and the circumference of the rotating disc, and secondly the
rate of rotation of the disc.

Now it lies well within the resources of modem mechanical
engineering to make a standard coil and disc of dimensions known
to an accuracy considerably greater than 1 in 10,000, the coil being
constructed of a single layer of wire wound in a screw thread cut
on a cylinder of large diameter ; and the measurement of the rate
of rotation to equal accuracy is a simple matter. There is difficulty
in maintaining a rate of rotation constant to this figure for four or
five minutes, but with the closest attention to the lubrication of
all the bearings this also might be accomplished. Such constancy
is well worth striving for, as the ease with which measurements of
resistance can be made by the method largely depends upon it.

I do not propose on this occasion to enter into the details of
the method I have adopted in making the measurements, the
results of which I have now to bring before the Section. But it
will perhaps be of interest if I say a few words about the time-
measurement.

In measuring a resistance we have to find the rate of rotation
corresponding to an equilibrium position. It is easiest in practice to
determine this by interpolation from two determined rates of rotation
(near together, and respectively slower and faster than the required
rate) and the galvanometer deflections corresponding to them, so
that each determination of resistance involves two determinations
of galvanometer deflection and the rates of rotation corresponding
to them.

480 PRACTICAL STANDARDS

In order that the galvanometer deflection may be obtained
with sufficient accuracy from a limited number of reversals (in
my observations the number has been almost uniformly thirty-
three, taking about four minutes in each case) the brush at the
circumference of the disc needs to be perforated and to be supplied
with a constant stream of mercury. Such a brush in its best
condition almost entirely eliminates the continual jerking of the
galvanometer needle consequent on thermo-electric changes at
the point of contact of brush and disc. A multiplication of such
brushes at three or four points of the circumference would do this
even more completely.

During the four or five minutes' run the rate of rotation is
referred by a stroboscopic method to a suitable tuning-fork pro-
vided with riders and maintained in vibration electrically. The
observer at the fork can shunt more or less current through the
electromotor driving the disc, and in this way maintains the rate
of rotation as constant as he can. But though the electrically
maintained fork is useful for purposes of control it cannot be relied
on to give us the rate of rotation. Its vibration period is not
within my experience constant to the degree of accuracy required.
If stopped and set going again it may start with a period different
by several parts in 10,000. No previous determination of the
period of the fork can therefore be relied on to give us the rate of
rotation, though once started the fork goes sufficiently uniformly
to give us a means of control.

Accordingly it is necessary to measure the rate of rotation
during each run while the galvanometer observations are being
made. The rotating disc is, by means of an eccentric attached to
its axle, made to record its revolutions on the tape of a Bain's
electro-chemical telegraph instrument side by side with the record
of the standard clock. We have, then, a time record exactly corre-
sponding to the period of observation of the galvanometer deflec-
tions. During the run the observer at the galvanometer calls out
the galvanometer readings, while the observer at the tuning-fork
controls the speed, and the Bain's instrument records it.

I have made in this way a number of measurements during the
months of July and August of a standard resistance of approxi-
mately '0005 ohm, prepared last year by my assistant, Mr Harrison,
and a student in my laboratory, Mr Parker, with the following
results : —

FOR ELECTRICAL MEASUREMENTS 481

July 17, morning ^ •00050016

„ i7, afternoon ... ... ... ... *00060016

„ 19, morning *00050015

Aug. 2, afternoon *00060020

„ 3, morning 00050021

,, 4, „ ■•* ••• .«* ... ^.^.lUOUv/lO

„ 4, afternoon -00050013

„ 5, morning -OOOSOOIO

,, ffj ,, ••. ••• ... ... \n,^Jtj\^J^ L

„ 9, afternoon "00050018

Mean 00050017

The maximum divergence from the mean is '00000004, or about
one part in 12,000. Mr Crompton has recently been issuing
standards of low resistance made of manganin sheets, and he was
kind enough, at my suggestion, to send me one for measurement
towards the end of July. .It was prepared in his laboratory as
a derivative from the Cambridge ohm by means of his poten-
tiometer. Its value so given was "00050175 at 23° C. Its
temperature coefficient appears, from the measurements made in
Mr Crompton's laboratory, to be so small that we need hardly
consider it for our present purpose. My measurements of this
standard were as follows: —

July 29, morning -00050219

Aug. 1, „ 00050225

„ 1, afternoon 00050219

„ 2, morning 00050226

Mean 00050222

which differs from Mr Crompton's value by something less than

one part in 1000. Mr Crompton's resistance ia a rectangular sheet

of manganin, and the potential terminals are two screws inserted

at a suitable distance apart in the median line. The screws are

not soldered. I thought it would be of interest to unscrew them,

screw them up again, and remeasure the resistance. The results

were : —

August 10, morning 00050328

„ 10, afternoon 00050322

„ 10| „ ... ... ... ... *UUll50o2S7

Mean 00050326

indicating a variation of about one part in 500. I unscrewed

them again, and after screwing them up made a new measure-
ment with the following results : —

B. A. 31

482 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS
August 11, morning "OOOSOdDS

,1 ^^ )f ••• ••• ••• ••• vWvOU4lW

Mean -OOOSCMOl

which, compared with the first value '00050222, shows a variation
of, approximately, one part in 280.

We may therefore conclude that if an accuracy of ^th per cent-
is required of a standard so constructed its potential terminals
ought not to be meddled with after its resistance has been
determined.

In making these measurements my direct object has been to
obtain an accurate and ready method of measuring standards of
low resistance. But I think something more than this comes out
of them. It would be possible in the light of our present ex-
perience to construct a Lorenz apparatus considerably more
accurate and easier to use than that in my laboratory at Cardifil
Such an apparatus placed, let us suppose, in the National Labo-
ratory, of which we have heard a good deal at recent meetings of
the British Association, might with advantage be kept in constant
use, not only for the calibration of low resistances, but also as
embodying in concrete form a proper ultimate standard of electrical
resistance. We have not in our electrical standard legislation
given full credit to the mechanical engineer for what he can do for
us ; and I think that a coefficient of mutual induction arranged, as
in the Lorenz method, so as to be easily combined with a time
would afford a more satisfactory standard of resistance than any
wire coil or coils, and one easier to use for purposes of ultimate
reference than any mercury column.

TWENTY-FIRST KEPORT— OXFORD, 1894.

APPENDIX PAOI

I. Report of the American Delegates at the Chicago Conference

to the Secretary of State at Washington . . . 485

II. Experiments on the Value of the Ohm. By J. Viriamu Jones 469

III. Comparison of the Standards employed by Professor Jones

with the Standards of the Association. By R. T. Qlazs-
BROOK 497

IV. Comparison of Some of the Standards of the Board of Trade

with those of the Association. By J. Rennib . 499

V. Values of Certain Coils belonging to the Indian Qovemment.

By E. O. Walker 601

VI. On the Speci6c Resistance of Copper and of Silver. By

Rev. T. C. FiTZPATRiCK 602

VII. Final Report of the Electrical Standards Committee of the

Board of Trade, and Order in Council regarding Standards

for Electrical Measurements 609

The Committee regret that the insulation of some of the coils
referred to in their last Report, which had been selected for the
new standards of resistance, as defined by the resolutions adopted
at Edinburgh, has proved defective. Traces of acid have been
discovered in the paraffin with which the coils were filled. The
two 1-ohm standards of the Association, as well as two of the
1-ohm standards of the Board of Trade, were found in January
last to have so low an insulation resistance between the coil and
the case as to be useless.

Thus the labour spent in the testing of these coils has been
wasted ; much of it will need to be done again. The insulation of
some of the other standard ohm coils is not satisfSeictory. The
single ohm standards have therefore been remade, and the others
are being refilled with carefully selected paraffin. The original
B. A. units have not, so fieir as comparisons between them can show,
changed their values during the year, and one set of new ohm
standards also has shown no sign of change. *

31—2

484 PRACTICAL STANDARDS

The Committee print, as an appendix to the Report, the Report
of the proceedings at the International Congress at Chicago,
presented to the Secretary of State at Washington by the
American delegates to the conference.

During the year Professor J. V. Jones has determined, by the
aid of his Lorenz apparatus, the absolute resistance of certain
wire coils of about O'l ohm. These have been compared with
the standards of the Association by the Secretary. An account
of these experiments is contained in Appendices II. and III.
The resistance standards of the Association have been compared
with those of the Board of Trade by Mr Rennie and the Secretary.
Details of this comparison will be found in Appendix IV., while
in Appendix V. is given, by Mr E. O. Walker, an account of a
comparison between five coils belonging to the Indian Government,
which have been for twenty-four years in India, and Dr Muirhead's
standards. Mr Fitzpatrick has continued his work on the specific
resistance of copper, and has drawn up a table (see Appendix VI.)
reducing to the same units experimental results recently obtained
by various observers. Appendix VII. contains the Final Report
of the Electrical Standards Committee of the Board of Trade
and the Order in Council relating to Standards for Electrical
Measurements.

In consequence of the difficulty met with in the insulation of
some of the coils, it was thought well to defer the purchase of other
coils for which the grant of £25 was obtained last year. The
Committee are of opinion that it is desirable to complete their
set of standards by obtaining fi*om Germany certified copies of
the standards of the Reichsanstalt. They recommend, therefore,
that they be reappointed, with the addition of the name of
Mr Rennie, and with a grant of £25; that Professor G. Carey
Foster be Chairman and Mr K T. Glazebrook Secretary.

FOR ELECTRICAL MEASUREMENTS 485

Appendix I.

Report of the Action of the International Electrical Congress
held in Chicago^ August 1893, in the Matter of Units of
Electrical Measure,

Washington, D.C.
November 6, 1893.

The Hon. W. Q. Gresham, Secretary of State, Washington, D.C,

Sir, — ^The undersigned, having been designated by you on
May 12, 1893, as delegates to represent the United States in the
International Electrical Congress to be held in August at Chicago,
beg to submit herewith a brief report showing the definitive action
of said Congress in the matter of defining and naming units of
electrical measure. The consideration of this important subject
was left to what was known as the " Chamber of Delegates " of
the Congress, consisting only of those who had been officially
commissioned by their respective Governments to act as members
of said Chamber. After conference and correspondence with the
leading electricians of Europe, it had been agreed that the maximum
number of such delegates to be allowed to one nation should be
five, and this number was allotted to the United States, Great
Britain, Germany, and France. Other nations were allowed three
or two, and in some instances one.

Delegates present and taking part in the discussions and action
of the Chamber were as follows : —

Representing the United States.

Professor H. A. Rowland, Johns Hopkins University, Balti-
more, Md.

Dr T. C. Mendenhall, Superintendent of U.S. Coast and
Geodetic Survey, and of Standard Weights and Measures,
Washington, D.C.

Professor H. S. Carhart, University of Michigan, Ann Arbor,
Mich.

Professor Elihu Thomson, Lynn, Mass.

Dr E. L. Nichols, Cornell University, Ithaca, N.Y.

486 PRACnCAL STANDARDS

Representing Oredt Britain.

W. H. Preece, F.RS., Engineer-in-Chief and Electrician, Poet
Office, England; President of the Institution of Electrical
Engineers, LondoiL

W. E. Ayrton, City and Guilds of London Central Institution,
Elxhibition Road, LondoiL

Professor Silvanus P. Thompson, D.Sc, F.R.S., Principal of the
City and Guilds Technical College, Finsbury, London.

Alex. Siemens, 12 Queen Anne's Gate, Westminster,
London, S.W.

Representing France,

E. Mascart, Membre de I'lnstitut, 176 Rue de TUniversit^,
Paris.

T. VioUe, Professeur au Conservatoire des Arts et Metiers,
89 Boulevard St Michel, Paris.

De la Touanne, Telegraph Engineer of the French Government,
13 Rue Soufflot, Paris.

Edouard Hospitalier, Professeur k TEcole de Physique et de
Chimie industrielle de la ville de Paris; Vice-Pr&ident de la
Soci^t6 intemationale des Electriciens, 6 Rue de Clichy, Paris.

Dr S. Leduc, 5 Quai Fosse, Nantes.

Representing Italy,

Comm. Galileo Ferraris, Professor of Technical Physics and
Electro-technics in the R. Museo Industriale, Turin, Via Venti
Settembre 46.

Representing Germany,

H.E. Hermann von Helmholtz, Prasident der physikalisch-
technischen Reichsanstalt, Professor a. d. Universitat, Berlin,
Charlottenburg bei Berlin.

Dr Emil Budde, Berlin N.W. Elopstockstrasse 53.

A. Schrader, Regierungsrath, Mitglied des kaiserl. Patentamts,
Berlin.

Dr Ernst Voit, Professor an der technischen Hochschule,
Miinchen, Schwanthalerstrasse 73-3.

Dr Otto Lummer, Mitglied der physikalisch-technischen
Reichsanstalt, Charlottenburg, Berlin.

FOR ELECTRICAL MEASX7REMENTS 487

Representing Mexico.
Augustin W. Chavez, city of Mexico.

Representing Austria,
Dr Johann Sahulka, Technische Hochschule, Wien.

Representing Switzerland,

A. Palaz, professeur, Lausanne.

Ren6 Thury, ing^nieur, Florissant, Genfeve,

Representing Sweden,
M. Wennman, Byr&chef i Rougle Telegra&tyrelsen, Stockholm.

Representing British North America^

Ormond Higman, Electrician, Standards Branch, Inland
Revenue Department, Ottawa,

His Excellency Dr H. von Helmholtz was made Honoraiy
President of the Congress; Dr Elisha Gi*ay, of Chicago, was
Chairman of the General Congress ; and Professor H. A. Rowland,
of Baltimore, was President of the Chamber of Delegates.

Meetings of the Chamber continued during six days, at the
end of which its members unanimously agreed in the adoption of
the following resolution : —

Resolved, That the several Governments represented by the
delegates of this International Congress of Electricians be, and
they are hereby, recommended to formally adopt as legal units
of electrical measure the following : As a unit of resistance, the
international ohm, which is based upon the ohm equal to 10* units
of resistance of the c.o.s. system of electro-magnetic units, and is
represented by the resistance offered to an unvaiying electric
current by a column of mercury at the temperature of melting ice
14*4521 grammes in mass, of a constant cross-sectional area and of
the length of 106*3 cm.

As a unit of current, the international ampere, which is one-
tenth of the unit of current of the CQJS. system of electro-magnetic
units, and which is represented sufficiently well for practical use

488 PRACTICAL STANDARDS

by the unvarj^g cuirent which, when passed through a solution
of nitrate of silver in water, and, in accordance with accompanying
specifications*, deposits silver at the rate of 0001118 of a gramme
per second.

As a unit of electromotive force, -the international volt, which
is the electromotive force that, steadily applied to a conductor
whose resistance is one international ohm, will produce a current
of one international ampere, and which is represented sufficiently
well for practical use by |f§$ of the electromotive force between
the poles or electrodes of the voltaic cell known as Clark's cell, at
a temperature of 15^ C, and prepared in the manner described in
the accompanj^g specification f.

As a unit of quantity, the intemationai coulomb, which is the
quantity of electricity transferred by a current of one international
ampere in one second.

As a unit of capacity, the international farad, which is the
capacity of a condenser charged to a potential of one international
volt by one international coulomb of electricity.

* In the following epeoifioation the tenn silver voltameter means the arrange-
ment of apparatus by means of which an electric carrent is passed through a
solution of nitrate of silver in water. The silver voltameter measures the total
electrical quantity which has passed during the time of the experiment, and by
noting this time the time average of the current, or, if the current has been kept
oonstant, the current itself, can be deduced.

In employing the silver voltameter to measure currents of about one ampere
the following arrangements should be adopted: —

The cathode on which the silver is to be deposited should take the form of a
platinum bowl not lees than 10 centimetres in diameter and from 4 to 5 oentimetres
in depth.

The anode should be a plate of pure silver some 30 sq. cm. in area and
2 or S mm. in thickness.

This is supported horizontally in the liquid near the top of the solution by a
platinum wire passed through holes in the plate at opposite comers. To prevent
the disintegrated silver which is formed on the anode from falling on to the
cathode, the anode should be wrapped round with pure filter paper, secured at
the back with sealing-wax.

The liquid should consist of a neutral solution of pure silver nitrate, containing
about 15 parts by weight of the nitrate to 85 parts of water.

The resistance of the voltameter changes somewhat as the current passes.
To prevent these changes having too great an effect on the current, some
resistance besides that of the voltameter should be inserted in the circuit. The
total metallic resistance of the circuit should not be less than 10 ohms.

t A committee, consisting of Messrs Helmholtz, Ayrton, and Garhart, was
appointed to prepare specifications for the Clark's celL Their report has not
yet been received.

FOR ELECTRICAL MEASUREMENTS 489

As a unit of work, the joule, which is equal to 10^ units of
work in the C.G.S. system, and which is represented sufficiently
well for practical use by the energy expended in one second by an
international ampere in an international ohm.

As a unit of power, the watt, which is equal to lO'^ units of
power in the c.G.s. system, and which is represented sufficiently
well for practical use by the work done at the rate of one joule per
second.

As the unit of induction, the henry, which is the induction in a
circuit when the electromotive force induced in this circuit is one
international volt, while the inducing current varies at the rate of
one ampfere per second.

The Chamber also voted that it was not wise to adopt or
recommend a standard of light at the present time.

A more complete report of the operations of the Chamber will
shortly be forwarded. This brief re»U7tii of its definite action in
reference to the matter of units is now submitted to facilitate the
prompt dissemination among representatives of foreign Govern-
ments of the important results of a congress of whose success and
fruitfiilness the United States may justly be proud.

H. A. Rowland. Elihu Thomson.

T. C. Mendenhall. E. L. Nichols.
H. S. Carhart.

Appendix II.

On a Determimttion of the Internatiofuil Ohm in Absolute Measure.
By Professor J. V. Jones, F.R.S., Principal of the University
College of South Wales and Monmouthshire, Cardiff.

The apparatus for the absolute measurement of electrical
resistance in my laboratory at Cardiff was completed in 1890, and
I first used it for the determination of the specific resistance of
mercury in absolute measure*. This determination was made by
direct measurement on a mercury column contained in a trough

• Phil. Tram,, 1891, A.

490

PRACTICAL STANDARDS

of paraffin wax. The results of five complete sets of observations
were as follows : —

94103
94074
94093
94045
94021

The mean of these is 94067 ; and the extreme variation from
the mean is 46, or about four parts in 10,000.

I suspected that much of the variation was due to the paraffin
trough, the temperature of which varied slightly (about half a
degree) during the observations, and was not accurately measurable
owing to the low conductivity of the material. With variation of
temperature there was variation of breadth, and the breadth of the
trough entered as a primary factor iato the calculation of the specific
resistance.

When I proceeded to use the apparatus for the measurement
of low-resistance standards of solid metal this was conclusively
shown to be the case. I brought a set of measurements made on
such a standard under the attention of the Section last year at
Nottingham, in which the extreme variation fix)m the mean was
only about one part in 12,000.

This may be taken to be the normal performance of the
apparatus ; and seeing that it is an instrument of such precision,
it seemed to me of interest to determine by the use of solid
metal standards the relation between its indications and the
results obtained by other observers for the value of the ohm.

With this end in view I obtained four coils fix)m Messrs Nalder
Brothers — two platinum-silver ten-ohm coils and two manganin
tenth-ohm coils. Mr Glazebrook has measured them in terms of
the international ohm ; and I am much indebted to him for the
pains he has been kind enough to take in making the determi-
nation. The following table gives their resistances and temperature
coefficients: —

Coil Number

Resistance in International
Ohm (Glazebrook)

Temperature Goeffioients
(Nalder)

3873

3874
4274
4275

9-9919 at 14-8" C.
9-9926 at 14-9" C.

•100050 at 15 2" C.

•100053 at 15-2" C.

•000300
•000276
■000013
•000013

FOR ELECTRICAL MEASUREMENTS

491

These coils were arranged in manner similar to that adopted
by Lord Rayleigh in his determination of the ohm by the method
of Lorenz (see fig. 1).

If there is no current through the galvanometer, there is
equality between the E.M.F. due to the rotation of the disc in
the field of the standard coil and the E.M.F. due to the current

Bi, R2, lO-ohm ooils,
R3, R4, *l-ohm ooils.

B, Battery.

6, GaWanometer.

D, Rotating disc.
KKy Standard ooil.

through R^; and we have, if J2i, /2,, i2,, i24 are the values of the
four resistance coils in international ohms, and if a is the value of
the international ohm in absolute measure,

22, + iia + jRj + JB4

where M = the coefficient of mutual induction of the standard coil
and the circumference of the disc, and n » the rate of rotation of
the disc.

The resistance coils are of B. A. pattern. They were immersed
in water, and the temperatures of thermometers within the coil
firames were read before and after each observation. A wooden
box surrounded the four cans containing the coils.

The method of making the observations was the same as that
described in the paper I read before the Section last year {vide
Electrical Standards Committee Report, 1893).

492 PRACTICAL STANDARDS

The results are as follows, the figure in each case giving the
value of the international ohm in true ohm&

JuLy 7. — Standard coil carefully adjusted. Three-minute
tapes.

•999703
•999761
'999807

Mean ... -999757

July 9. — No readjustment of standard coiL One-minute

tapes.

•999757

•999711

•999683

-999782

Mean ... -999733

July 10, morning. — Standard coil readjusted. One-minute

tapes.

•999734

•999818

999726

Mean ... -999759

Jvly 10, afternoon. — No readjustment of standard coil. Three-
minute tapes.

•999708
•999742
-999764

Mean ... -999738

July 11, afbemoon. — Standard coil readjusted. Three-minute

tapes.

•999693

-999692

-999679

Mean ... -999688

July 12, morning. — No readjustment of standard coil. Resist-
ance coils reversed. Three-minute tapes.

-999713
-999711
-999692

Mean ... 999705

FOR ELECTRICAL MEASUREMENTS 493

July 12, afternoon. — Standard coil readjusted. Resistance

coils removed from the mercury cups and replaced. Three-minute

tapes.

•999774

•999787
-999759

Mean ... -999773

July 13. — Standard coil readjusted. Resistance coils removed
from mercury cups and replaced. Three-minute tapes.

-999847
-999809
■999782
•999842 (morning of the 14th)

Mean ... -999820

July 14, morning. — Standard coil readjusted. Resistance coils
removed and replaced. Three-minute tapes.

•999695
•999692
-999717

Mean ... -999701

July 14, afternoon. — Standard coil readjusted. Resistance coils
removed and replaced. Three-minute tapes.

•999853
•999866
•999875

Mean ... 999865

It is clear that in the above series the chief variations are due

to changes consequent on readjustment of the standard coil,

and the removal and replacement of the resistance coils in their

mercury cups. Counting as independent only those of the

observations before which there was readjustment of the standard

coil or removal of the resistance coils from, the mercury cups, the

general mean is

•99976.

The maximum variation from the mean is '000106, or about
one part in 10,000.

494 PRACTICAL STANDARDS

Assuming that the international ohm is the resistance of a
column of mercury at 0^ of 1 sq. mm. sectional area, and 106*30 cm.
long, we have as a result of the above measurement that the true
ohm is the resistance of a column of mercury of the same sectional
area and 106*326 cm. long.

The fig^ure I arrived at in 1890, working direct on mercury,
was 106*307, with a probable error of + "Oil. The new result is
therefore a little larger than I was prepared for. The accuracy of
the result depends primarily on—

(i) The accuracy with which the resistance coils are known in
terms of the international ohm.

(ii) The accuracy with which their temperatures are known
at the times of observation.

(iii) The accuracy with which the coefficient of mutual
induction of the coil and disc has been determined.

Upon the first point I can say little. Mr Glazebrook knows
better than anyone to what figure the values of the resistances
may be relied on.

The eflFect of error in estimation of the temperatures of the
coils can be but slight. The observations have been made in two
ways, viz., with one-minute tapes, the current being put on only
during the time of observation, and with three-minute tapes, the
current being kept on continuously, whether observations were
being made or not. During the last few days of the observations
the current was kept passing through the coils night and day.
I have calculated the effect that would be produced on the result
obtained with one-minute tapes if all the heat generated by the
current were to remain in the coils — an extreme case, obviously
less favourable than the actual conditions. It is something less
than two parts in 100,000. The smallness of the effect is due to
the fact that if 7 is the main current, a current equal to |^ 7
passes through the tenth-ohm manganin coil with its small
temperature coefficient, and only ^ 7 through the platinumnsilver
coils ; while the effect of underestimating the temperature of the
manganin tenth-ohm coil is to produce an error in the result
opposite in sign to that produced by underestimating the tem-
perature of the platinum-silver coils.

There cannot, then, in the case of the one-minute tape
observations be an appreciable error due to underestimation of
the temperature. But the first four sets of observations show

FOR ELECTRICAL MEASUREMENTS

495

that the results of the one-minute tape observations and the
three-minute tape observations are practically the same. Hence
it follows that to the degree of accuracy aimed at our results are
unaffected by error due to imderestimation of the temperature.

It remains to consider the accuracy with which the coefficient
of mutual induction of the coil and disc is known.

To calculate this coefficient we must know the radius of the
disc and the mean radius of the coil. The circumference of the
disc is a sufficiently true circle, the disc having been ground true
in place. The measurement of its diameter presented no difficulty.
It was determined on my Whitworth measuring machine to the
ten-thousandth of an inch.

The mean radius of the coil cannot be determined with the
same accursu^y ; but I believe that it is known to the thousandth
of an inch. The coil consists of a single layer of silk-covered wire
wound in a screw thread cut on a brass frame. It was measured
along eighteen diameters in the Whitworth machine with the
following results: —

Diameter

Measarement

Diameter

Measoremeut

0»— 180"

21 0838

90"— 270"

21-1038

10' lOO**

21-0929

100" 280"

21-1056

20"— 200'

21-0951

110"— 290"

211041

30"— 210'

21-0933

120" 300"

211014

40"— 220"

21-0960

130"— 310"

21-0979

50" 230"

21-0998

140" 320"

21-0945

60" 240"

211017

150"— 330"

21-0924

70"— 260"

21-1026

160" -340"

21-0900

80"— 260"

211044

170-— 350"

21-0910

Max

211056

Mean

21-09757

Min. ...

21-0898

<=17"

•0158

These measures clearly show that the coil is elliptical in section,
the difference between the major and minor axes being about *008
inch, or about one part in 1,300.

In considering the possible effect of this ellipticity on the result,
it must be borne in mind that the formula R » Mn implies that the
coil is circular. The true formula is

R s 27m I aJSda,

J a»

where o^ and Oi are the distances from the centre of the disc at

496 PRACTICAL STANDARDS

which the internal and external brushes are applied, and H is the
magnetic force at a distance a firom the centre when unit current
is passing through the coil.

This is an unpleasant integral for an elliptical coil, and it has
not yet yielded to persuasion. It is, however, satisfactory to note
that as in my apparatus the brush radius makes but a small angle
with the minor axis (about 15**), I am, in so far as the ellipticity
of the coil affects matters at all, underestimating the integral, and
hence underestimating the international ohm. Any correction for
ellipticity hereafter calculated will make the value of the later-
national ohm deduced from my observations nearer to and not
further from the true ohm.

It is further to be noticed that the formula JJ = i/n applies
only if there is exact coincidence of the axes of the disc and coiL
It has been customary to consider the adjustment for centre as of
secondary importance in Lorenz's method. It would be so if the
formula R = Mn were applicable when the centres of coil and disc
do not coincide, for a slight displacement only affects the coefficient
of mutual induction to a secondary degree. But we are not con-
cerned with the coefficient of mutual induction in this case. We
are concerned with another integral, viz.,

27r I aHda ;

and the adjustment for centre is ia truth of primary importance.
Special attention should therefore be paid to this in designing
apparatus for the absolute measurement of resistance by this
method.

One other poiat remaias to be noticed in this connexion, viz.,
the possible effect of the difference of the temperature of the coil
and disc when measured and when in use. On calculating the
correction to be applied for this cause I find it negligible.

Again, I would say, as I said last year, that the chief value of
these observations consists in the proof they afford of the precision
with which the absolute measurement of resistance may be made
by this method. A well-constructed apparatus of the kind in a
national laboratory — say the Laboratory of the Board of Trade —
will, I believe, prove to be the best ultimate standard of electrical
resistance.

FOB ELECTRICAL MEASUREMENTS 497

Appendix ni.

Comparison of the Standard Coils used by Professor Jones with
the Standards of the Association, By R T. Qlazebrook.

The tenth-ohm standards of manganin wire whose value in
absolute measure was determined by Professor Jones by means of
the experiments described in Appendix II. were compared with
the standards of the Association in the following manner. A
Wheatstone's bridge was formed in which the arms were the
tenth-ohm to be tested, two single-ohm coils and a ten-ohm coil ;
if the coils had these values exactly, there would of course always
have been a balance ; since, however, the coils were not accurately
correct there was usually a small current through the galvano-
meter ; the balance, however, could be obtained by placing a large
resistance as a shunt either to one of the one-ohm coils or to the
ten-ohm coil : this resistance, which varied from 10,000 to 20,000
ohms, was taken from a good box of coils. The resistance of the
ten-ohm and of the two one-ohm coils being known, that of the
tenth-ohm coil could readily be found.

The four coils dipped into four mercury cups cut in an ebonite
block ; the bottoms of these cups were copper pieces some 3 to
4 mm. thick.

Binding screws screwed into these copper pieces and rising
above the mercury served to connect the bridge to the galvano-
meter and the battery.

The mercury cups were somewhat large — about 2*5 cm. in
diameter — and it was found on January 16 that distinct differences
could be observed by moving the tenth-ohm coils slightly so as to
bring their terminals either close to or as far as possible from the
feet of the one-ohm coils which dipped into the same cups. After
this date two sets of measurements were made for each coil at
each observation : in the one the terminals of the coils in any
cup were put as close together as possible, in the other the
terminals of the tenth-ohm coils were placed at some distance from
those of the other coil in the same cup.

Both sets of values are given in the table as a means of showing
the delicacy of the observations and the error arising from this
cause. The tenth-ohm coils were weighted so as to press firmly
on to the copper bottoms. No variation was produced by shifting
the ten-ohm coil in its cup.

B, A. 32

49«

PRACTICAL STA^OARDd

One or two Leclanch^ delld were used in the various experi-
ments: the coils were in water-baths and the temperatures read
by a standardised Kew thermometer.

The standard coils usied were —

Elliott 264= 1 + -000312 («-15-45).
Nalder 3715 = 1 + -000260 (t - 14-95).
Elliott 289 = 10-1- -002600 (t - 15-40).

The results of the experiments are given in the following tables.

In the results of the experiments made after January 15 the
two values given correspond to the two positions of the coil in the
mercury cup. They are included to show the magnitude of the
error, which may be due to the resistance of the copper bottoms of
the cup.

Tables giving Values of Nalder No, 4274 '^ No. 389 in
terms of the Ohm Standards of the Association,

Date

Temperature

Value of Eesistance

Dec. 29, 1893...
Jan. 13, 1894...

>» 15, >» •••

» 1"> » •••
„ 1/, )) •••

»» *»> » •••

Feb. 20, „ ...
March 17, „ ...

Mean

14-4" C.

152

14^8

15-5

16-4

16-9

141

15-2

•100052

•100051

•100056
•100051 100056
•100049 -100058
•100057 -100066
•100036 -100041
•100045 -100046

-100050 -100054

Valvjes of N

'alder No. 421

r5 ^ No, 390.

Date

Temperature

Value of Besifitanoe

Dec. 29, 1893...

Jan. 13, 1894...

„ 10, „ ...

»i 1"» » •••

M l'> »5 •••

n *'> »> •••

Feb. 20, „ ...
March 17, „ ...

Mean

14-6'' C.

15-8

164

16-6

14-1

13-8

16-2

•100059

-100053

•100058
-100056 100061
-100051 -100059
•100058 -100066
•100043 ^100047
•100061 -100061

•100053 -100057

FOR SUCTIOC^At MfiASURElfENTS 49$

Thus, the values oTtTie "coils at'l5*2°C. are respectively for

^389 -100050 ohm,

and for

^390 -100053 ohm,

while in each case the resistance introduced by placing the^ contact
pieces of the tenth-ohm coils at some distance from those of the
other coils is -000004 ohm.

Appendix IV,

Comparison of certain Ohm-Standards of the Board of Trade.

By J. Rennie.

In the accompanying table (p. 500) are given the results of
comparisons which were made on May 29 and 30, 1894, at the
Cavendish Laboratory, between the three unit coils: —

Elliott's No. 261,

Elliott's No. 263,

Nalder's No. 3876.

belonging to the Electrical Department of the Board of Trade, and
the B. A. standards, Flat, F, 0, and H.

The bridge was of the Carey Foster pattern, constructed for
the Department by Nalder Bros, and Co., and the slide wire
used was the one marked B, having a value of '0000509 ohm
per division.

A 100-ohm coil, Elliott's No. 291, was placed in parallel with
the Board of Trade coil for each comparison, this being effected by
a large meicury-in-paraffin bath.

Temperatures were measured by a mercury-in-glass thermo-
meter, which had been standardised at Eew.

32—2

500

PRACTICAL STANDARDS

B.A. CoU

Flat

Flat
F
O
H

Flat ...

F ...

O ...

a ...

Temp, of
B.A. CoU

Temp, of B.
of T. Coil

12-60
12-60
12-70
12-40

Observed Valae

Chart*
Value

No. 261 Coil

No. 3876 Coil

12-45 -998808

12-36 -998727

12-23 -998697

1230 ! -998768

-99879
-99876
•99874
•99876

Difference,
Chart-
observed

12-54'C.

12-53

12-90

12-58° C.

12-45

12-97

•999156
•999064
-999217

-999142
-999106
-999262

-•000014
+ -000042
+ •000045

No. 263 Coil

12-60
12-60
12-80
12-85

12-53
12-48
12*98
13-08

•999136
-999074
-999217
-999299

-999140
•999124
-999271
•999304

+ ■000004
+ •000050
+•000054
+ -000005

-■000018
+ •000033
+ •000043
-•000008

* The chart referred to, for No. 261 and No. 263 coils, is one supplied for these
coils by Mr Glazebrook, and is dated March 1892. The chart referred to in the
case of No. 3876 was constructed from comparisons made by N alder Brothers
between it and their ** master coil/' No. 8717. The coils Nos. 263 and 261 were
compared on May 29, 1894, before beginning the above-mentioned series of com-
parisons. They were found exactly equal, when the temperatures were — No. 263,
12-65° C. ; No. 261, 12-62° C. The chart values at these temperatures are— No. 263,
0^999175 ; No. 261, 0-999156 ; showing a difference of 19 x YT* ohms. The corre-
sponding differences deduced from the above table are — from Flat, 18 x 10~-* ohms ;
from i^, 8 X 10~^ ohms ; from G, 9 x 10~* ohms. The comparison No. 261 — H is
omitted, as the difference obtained was obviously much too large, and must have
been caused by some undetected interference. It is evident from the eleven results
given in the table that the difference between the coils Nos. 263 and 261 as deduced
from comparison with K must be something like 10 x 10~' ohms. [Note added
October 5, 1894.]

FOB ELECTRICAL MEASUREMENTS

601

Appendix V.

Table showing valvsa of five standard coils B,A. Units belonging
to the Indian Government as compared with Dr Muirhead^s
standard at his Laboratory, By E. O. Walker, C.I.E.,
M.I.E.E., Late Superintendent in the Government Telegraph
Department in India.

Standard used, No. 78, marked right at 15*7° C, taken as
correct. This standard, tested April 27, 1893, against a No. 68

Glazebrook, gave a ratio ^^^^ of 1-00015 at le*" C, and 1*00018

at 15*4° C.

Temperature of water, 20*2° C.

Number

Marked right at

Difference

Correct at

106
108
110
111
114

151" C.
15-3" 0.
16-3' C.
15-5" C.
151" C.

+ 023 per cent
+ 126 „
-•028 „
+ -055 „
+ •004

14-0" C.
11-7" C.
16-6" C.
13-9" C.
15-6" C.

Apparatus used one metre bridge of platinum-iridium wire
with a supplementary coil at each end of 20,012 millimetres.
Suspended coil galvanometer, resistance 15 ohms (Muirhead and
Co,'s). Trough, 45 x 7 J x 5 inches ; depth of water, 2 J inches ;
quantity of water, 6{ gallons ; battery used, 1 Hellesen's Dry Cell,
No. 3; E.M.F., 1*4 volt

The interest attaching to these tests lies especially in the
fact that the standard coils have been exposed to the climate of
Calcutta for twenty-four years They were made, I understand,
by Dr A. Muirhead when in Dr Matthiessen's laboratory, under
the supervision of the latter.

In reducing the observations from 20*2** C. to the temperatures
given, it has been assumed that all the coils have the same
temperature coefficient.

^2 PBACnCAL STANDARDS

Appendix VI.

On the Specific Resistance of Copper and of Silver.
By Rev. T. C. Fitzpatrick.

As lately several observers have published the results of
measurements made on the specific resistance of copper, it may
be worth while to collect these results together in tabular form.

The resistances of metals may be expressed in teims of equal
weight or of equal volume ; that is, as the resistance of a wire of
the given material such that one metre of it weighs one gramme
or as the resistance between opposite fetces of a cube of the material
each face of which is one square centimetre. I have pointed out
that Matthiessen* considered the first as the most satisfactory
mode of expressing resistances, and for these results alone did he
make all the actual experiments ; the results for specific resistances
were calculated from these with the help of specific gravity values
obtained in many cases from tables, and not determined directly
for the wires used.

Only in cases where considerable masses of the material are
used can the specific gravity, and fix)m this the cross-section of
the wires, be accurately determined. There is therefore an evident
advantage in expressing results in terms of weight, as then the
determination of the cross-section of the wires becomes unnecessary,
and there is no reason why an accuracy of one in two or three
thousand should not be attained.

Again, it is found that difierent samples of copper have
difierent densities, according to the method by which they have
been prepared; in a tablet which I published on a previous
occasion the variation is fi-om 88C to 8*95. Mr SwanJ gives a
value as high as 8-9587.

From samples of copper of the same quality I have had wires
drawn which differed in density; it was always found that the
denser the copper the less is its resistance, and the difference
affects much more the results expressed as specific resistances than
when expressed as the resistance for a gramme per metre.

This is another reason for expressing the results in these terms,

* B.A. Report, 1890, p. 410. t B.A. Report, 1890, p. 404.
X Nature, vol. l. p. 165.

FOR ELECTRICAL MEASUREMENTS

SOS

at l^ast as well as specific resistances, and for actual practical
purposes it is a question of weight rather than volume.

In the following tables the results are given for a temperature
of 18° C. in C.G.S. units.

Table A. Hard-drawn Copper Wires,

Be&istanoe of wire
Buoh that the metre
weighs one gramme

Specific resistance

per
onbio centimetre

Observer

1550 xlO»

1550
1527

1743
1720
1726
1708

Matthiessen *
Swan and Rhodin f
Fitzpatrick %

»» (cjopper sup-
plied by Messrs Bolton)

Table B. Annealed Wires.

Resistance of wire Specific resistance
snch that the metre per

weighs one gramme ' cubic centimetre

Observer

1616 xlO»

1488
1488

1704
1681
1680
1665

Matthiessen *
Fleming and Dewar §
Swan and Rhodin f
Fitzpatrick { (copper sup-
plied by Messrs Bolton)
(Wire sent by Mr Swan)

From Table A it will be seen that Messrs Swan and Rhodin
obtained a value rather lower than that which I got for copper,
prepared by myself, and which, expressed as the resistance of a
wire one metre long weighing one gramme, is identical with the
value that Matthiessen obtained ; but the resistance of all these
specimens is distinctly greater than that of the copper kindly sent
me by Messrs Bolton — which seems to bear out the statement,
which I have previously made, that it is impossible to prepare
wires on the small scale which are of the same quality, i,e, probably
due to density, as the best specimens specially prepared by large
manufacturers.

* B. A. Report, 1864.

t B.A. Report, 1890, p. 404.

t Nature^ yoI. l. p. 165.

g PkiL Mag. voL xzzvL p. 287.

504

PBACTICAL STANDARDS

In Table B are given the results of measurements on three
specimens of copper prepared by Mr Swan: one was given to
Profs. Dewar and Fleming ; a second was examined by Mr Swan
himself, and a third specimen he kindly sent to me ; the quality
of the copper in the three cases may therefore be expected to be
the same.

The results of Profe. Dewar and Fleming and Messrs Swan
and Rhodin are expressed only as specific resistances, whilst
the result of my measurement is only given for a wire one
metre long weighing one gramme. The weight of the copper
wire, as measured, was only three grammes, and that does not
allow the accurate determination of the specific gravity of the
sample. The value I obtain for its resistance ia identical
with that for the sample of annealed copper wire sent me by
Messrs Bolton.

If it be considered to have the same specific gravity as that
sample (8*94) its specific resistance in C.G.S. units is 1665 ; a value
distinctly smaller than that obtained by Messrs Swan and Rhodin,
whose result is practically identical with that of Profs. Dewar and
Fleming.

Not only may wires drawn from the same specimen of copper
have different densities and different resistances, but the variation
of that resistance with change of temperature may be also different.

In the following table are given the temperature coefficients of
various specimens of copper : —

Observer

1561
1603*
1563
1592 •

•00387
•00428
•00408
•00417
■00405
•00406
•00364

Matthieesen t
Dewar and Fleming {
Swan and Rhodin §

Fitzpatrick

Kennedy and FeHnenden |

Benoitl

* Hard-drawn wires.

t B.A. Report, 1864. % Phil. Mag. vol. xzxyi. p. 287.

§ Nature, vol. l. p. 165. || Electricity^ vol. y. p. 165.

% Ccmptet Rendm^ vol. lxzvi. p. 845.

FOR ELECTRICAL MEASUREMENTS

505

Influence of AnneaUng. — As is well known, annealed wires
have a less resistance than hard-drawn wires, but the variation
of resistance according as the wires are annealed or hard-drawn
differs considerably for different materials. For silver it is as
much as 10 per cent., whereas for copper it is less than 3 per cent.

I have made observations from time to time on the resistance
value of specimens of hard-drawn copper wire, all pieces of the
same coil, which were sent me in 1889 by Messrs Bolton and Son.
From the results of these measurements it will be seen that a
hard-drawn wire seems to fall in resistance with lapse of time.
The coil of wire has been left hanging in the laboratory, and has
not been treated with any special care.

Date

Temperatore

Resistance of wire 1 metre
long weighing 1 gramme

July 1890

August 26, 1891...

March?, 1892

January 1894

July 1894

18" C.

18"

1528 X 10*

1526

1522

1520

1519

The fall in resistance is small, and for the period of nearly five
years does not amount to more than ^ per cent.

I have, for the sake of comparison, made a measurement of the
resistance of a specimen of annealed copper wire sent me by the
same firm, and for this the resistance value is identical with that
obtained at a previous date : —

Date

Temperature

Resistance of wire 1 metre
long weighing 1 gramme

October 1889

July 1894

18" C.
18"

1488 XlO*
1487-8

This, on the whole, is what one would expect. In the case of
wires of other material the change would probably be greater, as
the difference in resistance between annealed and hard-drawn
copper wires is less than that for wires of other materials.

5^ PRACTICAL STAN|>AI»PSr

In my previous communication a^ method*^ of annealing was
described which gave satisfactory results. The ¥dre was packed
in asbestos and fine carbon in a copper vessel and heated for
twenty-four hours. The following results amongst others were
obtained : —

Hard-drawn 18° Annealed 18° Difference

1527x10* 1489 38

1626 1488 38

Matthiessen s values are : —

1550 1516 34

Messrs Swan and Rhodin give for the values of the specific
resistance : —

Ha'-d-drawn 18° Annealed 18° Difference

1720 1680 40

I have recently been annealing copper wires by heating them
in boiling paraffin (220'') ; and after slow cooling the wires seem to
be completely annealed : —

Hard-drawn 18° Annealed 18° Difference

1526 1486 40

A wire sent me as annealed gave the result : —

Annealed 18°, 1488

This wire was then hardened, and, reannealed as above described^

gave the value : —

Annealed 18% 1489

Either of these two methods seems to give satisfactory results.
For completely annealing silver wires the temperature of the
paraffin bath is not sufficiently high, but from the results of my
measurements for silver, for which the influence of annealing
is very considerable, it can be seen that the first method is quite
satisfactory.

Silver.

Many of the older measurements for resistances and con-
ductivities are expressed in terms of the resistance of pure silver :
this was the case with Matthiessen*s earlier results.

* B.A. Report, 1890, p. 405.

FOR ELECTRICAL MEASUREMENTS

5or

' ' Sdme measurements therefore made on silver wires Are given
together with the results obtained by Matthiesaen and Profs, Dewar
and Fleming for the sake of comparison.

Several samples of silver wires were supplied by Messrs Johnson
and Matthey : one of these was stated to be absolutely pure.

The result* are expressed for wires weighing one gramme per
metre.

Hard-drawn

Annealed

Silver I

1816x10*

1814

1816

1799

. 1777

.^773

1767

1739 X 10*

1741

1721

1722

1666

V j> •

Silver II

Silver III.,. pure.. s-.

The difference between the values for the hard-drawn wires
is probably due to the feet that they had to be further drawn
down after I had received them to enable me to measure them on
my bridge.

Matthiesaen's valae

For wire 1 metre long
weighing 1 gramme

Resistance per c.o.

Hard-drawn

Annealed

1779 X 10*
1639

1694*
1561

Profs. Dewarf and Fleming give as the value for an annealed
pure silver wire 1468 c.a.s. at 0°C. with the temperature co-
efficient of 0004 ; the value at 18° C. is therefore for the specific
resistance 1574.

For most of the wires which I measured the specific gravities
were determined; there is practically no difference between the
values obtained for the annealed and hard-drawn wires, the values
varying from 10*496 to 10'511.

* Using the value 10*5 as the specific gravity of silyer.
t Phil, Mag. toI. zxzvi.

508

PRACTICAL STANDARDS

For the wires Silver IIL the values varied from 10*49 to 10*50 ;
using the mean value 10*496 I get for the specific resistances the
following values : —

Hard-drawn wire

Annealed

Specific resistance in CG.s.
units at 18** C.

1689

1587

In the case of copper* with increase of purity there is a
decrease in the difference in resistance between annealed and
hard-drawn wirea With silver the reverse is the case.

Silver I. — Difference
II. — Difference
III. — Difference

»>

77
107

The value that I obtained for the hard-drawn wire is very
nearly the same as that given by Matthiessen, but he obtained
a greater decrease in resistance on annealing. He states f that for
different pieces of the same wire there was a variation of from
6 to 10 per cent.; so that the difference between his value and
that which I have obtained for a sample of pure silver is not
greater than might be expected.

The considerable variation iq all the values given above makes
it clear that the values of the specific resistance depend, not simply
on the purity of the material, but on a number of other factors,
which will be different in the cases of different wires of the same
material, and that therefore we cannot expect to attain to any
great degree of accuracy in the determination of specific resistances
as distinguished from the accurate measurement for some particular
wire.

B. A. Report, 1890.

t Phil. Trans, 1862, p. 7.

FOB ELECTRICAL MBASUBEIIENTS 509

Appendix VIL

Final Report op the Electrical Standards Committee

OP THE Board of Trade.

To the Right Hon. James Bryce, M.P.,
President of the Board of Trade.

Since the date of our last Report the Board of Trade have laid
before us a riaumi of the action of the International Electrical
Congress held in Chicago in August 1893 to determine the units
of electrical measurement. We were also infonned by the Board
of Trade that her Majesty's Government had been invited by the
United States Ambassador in London to take steps to adopt the
recommendations of the Congress.

These recommendations, so far as they refer to the units of
electrical resistance, electrical current, and electrical pressure, are
substantially the same as those suggested for adoption in our
previous Reports.

We see no reason for further delay in the legalisation of
standards of the above-mentioned units, and we have prepared
and attach a revised Draft Order in Council*, which we advise
may be submitted for her Majesty's gracious approval.

The accompanying notesf to the specification for the Clark's
cell have been communicated by Mr Glazebrook, and will be found
of great assistance in the preparation of this form of cell.

(Signed) Courtenay Boyle. Kelvin.

Francis J. S. Hopwood. P. Cardew.

W. H. Preeoe. Rayleigh.

G. Carey Foster. R. T. Glazebrook.

J. HOPKINSON. W. E. Ayrton.

T. W. P. Blomefield, Secretary.

August 2, 1894.

* The Order in Council is printed in the form in which it has since received
her Biajesty's spprovaL
t For^the notes see p. 516.

wo • PRACTKTlL ^TAKDARDS .

Oi'der in Council regarding Standards for Electrical

. Mea>8arement9K

At the Court at Oibome Houae^ Ide of Wight^ August 23, 1894.
Present: The QueetCs Most Excellent Majesty in Council,

Whereas by 'The Weights and Measures Act 1889' it is
among other things enacted that the Board of Trade shall fix>m
time to time cause such new denominations of standards for the
measurement of electricity as appear to them to be required for
use in trade to be made and duly verified.

And whereas it has been made to appear to the Board of Trade
that new denominations of standards are required for use in trade
based upon the following units of electrical measurement, viz. —

1. The ohm, which has the value 10® in terms of the centi-
metre and the second of time, and is represented by the resistance
offered to an unvarying electric current b}^ a column of mercury
at the temperature of melting ice 14*4521 grammes in mass of a
constant cross-sectional area and of a length of 106*3 centimetres.

2. The ampfere, which has the value -^ in terms of the
centimetre, the gramme, and the second of time, and which is
represented by the unvarying electric current which when passed
through a solution of nitrate of silver in water in accordance with
the specification appended hereto, and marked A, deposits silver
at the rate of 0*001118 of a gramme per second.

3. The volt, which has the value 10® in terms of the centi-
metre, the gramme, and the second of time, being the electrical
pressure that if steadily applied to a conductor whose resistance
is one ohm will produce a current of one ampere, and which is
represented by 0*6974 (|^J) of the electrical pressure at a tem-
perature of 16"* C. between the poles of the voltaic cell knovm as
Clark's cell set up in accordance with the specification appended
hereto, and marked B.

And whereas they have caused the said new denominations of
standards to be made and duly verified.

Now, therefore, her Majesty, by virtue of the power vested in

her by the said Act, by and with the advice of her Privy Council,

is pleased to approve the several denominations of standards set

forth in the schedule hereto as new denominations of standards for

electrical measurement

C. L. Peel.

FOR ffiLSCTRICAL MfiASURXlffENTS 811

Schedule.

I. — Standard of Electrical Resistance.

A standard of electrical resistance denominated one ohm being
the resistance between the copper terminals of the instrument
marked 'Board of Trade Ohm Standard Verified 1894' to the
passage of an unvarying electrical current when the coil of
insulated wire forming part of the aforesaid instrument and con-
nected to the aforesaid terminals is in all parts at a temperature
of 154" C.

II. — Standard of Electrical Current

A standard of electrical current denominated one ampere being
the current which is passing in and through the coils of wire
forming part of the instrument marked ' Board of Trade Ampere
Standard Verified 1894* when on reversing the current in the
fixed coils the change in the forces acting upon the suspended coil
in its sighted position is exactly balanced by the force exerted by
gravity in Westminster upon the iridio-platinum weight marked A
and forming part of the said instrument.

III. — Standard of Electrical Pressure,

A standard of electrical pressure denominated one volt, being
one-hundredth part of the pressure which when applied between
the terminals forming part of the instrument marked ' Board of
Trade Volt Standard Verified 1894/ causes that rotation of the
suspended portion of the instrument which is exactly measured by
the coincidence of the sighting wire with the image of the fiducial
mark A before and after application of the pressure and with that
of the fiducial mark B during the application of the pressure, these
images being produced by the suspended mirror and observed by
means of the eyepiece.

In the use of the above standards the limits of accuracy
attainable are as follows: —

For the ohm, within one-hundredth part of one per cent.
For the ampfere, within one-tenth part of one per cent.
For the volt, within one-tenth part of one per cent.

The coils and instruments referred to in this schedule are
deposited at the Board of Trade Standardising Laboratory,
8, Richmond Terrace, Whitehall, London.

512 PRACTICAL STANDABDB

Specifications referred to in the forbqoing Order

in cjouncdu

Specification A.

In the following specification the term silver voltameter means
the arrangement of apparatus by means of which an electric
current is^passed through a solution of nitrate of silver in water.
The silver voltameter measures the total electrical quantity which
has passed during the time of the experiment, and by noting this
time the time-average of the current, or if the current has been
kept constant the current itself, can be deduced.

The anode should be a plate of pure silver some 80 square
centimetres in area and 2 or 3 millimetres in thickness.

This is supported horizontally in the liquid near the top of the
solution by a platinum wire passed through holes in the plate at
opposite comers. To prevent the disintegrated silver which is
formed on the anode from falling on to the cathode, the anode
should be wrapped round with pure filter paper, secured at the
back with sealing-wax.

The liquid should consist of a neutral solution of pure silver
nitrate, containing about 15 parts by weight of the nitrate to
85 parts of water.

The resistance of the voltameter changes somewhat as the
current passes. To prevent these changes having too great an
effect on the current, some resistance besides that of the voltameter
should be inserted in the circuit. The total metallic resistance of
the circuit should not be less than 10 ohms.

Method of making a Measurement.

The platinum bowl is washed with nitric acid and distilled
water, dried by heat, and then left to cool in a desiccator. When
thoroughly dry it is weighed carefully.

It is nearly filled with the solution, and connected to the rest

FOB ELECTRICAL MEASUREMENTS 513

of the circuit by being placed on a clean copper support to
which a binding screw is attached. This copper support must
be insulated.

The anode is then immersed in the solution so as to be well
covered by it and supported in that position ; the connexions to
the rest of the circuit are made.

The solution is now removed from the bowl and the deposit is
washed with distilled water and left to soak for at least six hours.
It is then rinsed successively with distilled water and absolute
alcohol and dried in a hot-air bath at a temperature of about
IdO"" C. After cooling in a desiccator it is weighed again. The
gain in weight gives the silver deposited.

To find the current in amperes, this weight, expressed in
grammes, must be divided by the number of seconds during which
the current has been passed, and by 0*001118.

The result will be the time-average of the current, if during
the interval the current has varied.

In determining by this method the constant of an instrument
the current should be kept as nearly constant as possible, and the
readings of the instrument observed at frequent intervals of time.
These observations give a curve from which the reading cor-
responding to the mean current (time-average of the current)
can be found. The current, as calculated by the voltameter,
corresponds to this reading.

Specification B.
On the Preparation of the Clark Cell.

Definition of the Cell.

The cell consists of zinc or an amalgam of sine with mercury
and of mercury in a neutral saturated solution of zinc sulphate and
mercurous sulphate in water, prepared with mercurous sulphate in
excess.

B. A. 33

514 PBACnCAL STANDARDS

Preparation of the Materials.

1. The Mercury. — To secure purity it should be first treated
with acid in the usual maimer, and subsequently distilled in vacuo.

2. I%e Zinc. — Take a portion of a rod of pure redistilled zinc,
solder to one end a piece of copper wire, clean the whole with
glass paper ot a -steel burnisher, cM^fiilly removing any loose pieces
of the zinc. Just before making up the cell dip the zinc into dilute
sulphuric acid, wash with distilled water, and dry vrith a clean cloth
or filter paper.

3% The ' Mercuroue Sulphate. — ^Take mercurous sulphate,
purchased as pure, mix with it a small quantity of pure mercury,
and wash the whole thoroughly with cold distilled water by
agitation in a bottle ; drain off the water, and repeat the process
at least twice. After the last washing drain oS as much of the
water as possible.

4. The Zinc Sulphate Solution. — Prepare a neutral saturated
solution of pure (" pure recrystallised ") zinc sulphate by mixing in
a flask distilled water with nearly twice its weight of crystals of
pure zinc sulphate, and adding zinc oxide in the proportion of
abotit 2 per cent, by weight of the zinc sulphate crystals to
neutralise any firee acid. The crystals should be dissolved with
the aid of -gentle heat, but the temperature to which the
solution is raised ishould not exceed 30'' C. Mercurous sulphate
treated as described in 3 should be added in the proportion
of about 12 per cent, by weight of the zinc sulphate crystals
to neutralise any free zinc oxide remaining, and the solution
filtered, while still warm, into a stock bottle. Crystals should
form as it cools.

5. The Mercuroue Sulphate and Zinc Sulphate Paste. — Mix
the washed mercurous sulphate with the zinc sulphate solution,
adding sufficient crystals of zinc sulphate from the stock bottle to
ensure saturation, and a small quantity of pure mercury. Shake
these up well together to form a paste of the consistence of cream.
Heat the paste, but not above a temperature of 30'' C. Keep the
paste for an hour at this temperature, agitating it fix>m time to
time, then allow it to cool ; continue to shake it occasionally while
it is cooling. Crystals of zinc sulphate should then be distinctly
visible, and should be distributed throughout the mass. If this is

FOR ELECTRICAL MEASUREMENTS 515

not the cose add more crystals from the stock bottle, and repeat
the whole process.

This method ensures the formation of a saturated solution of
zinc and mercurous sulphates in water.

To set up the Cell.

The cell may conveniently be set up in a small test-tube of
about 2 centimetres diameter, and 4 or 5 centimetres deep. Place
the mercury in the bottom of this tube, filling it to a depth of,
say, 0'5 centimetre. Cut a cork about 0*5 centimetre thick to fit
the tube ; at one side of the cork bore a hole through which the
zinc rod can pass tightly ; at the other side bore another hole for
the glass tube which covers the platinum wire ; at the edge of the
cork cut a nick through which the air can pass when the cork
is pushed into the tube. Wash the cork thoroughly with warm
water, and leave it to soak in water for some hours before use.
Pass the zinc rod about 1 centimetre through the cork.

Contact is made with the mercury by means of a platinum
wire about No. 22 gauge. This is protected from contact with
the other materials of the cell by being sealed into a glass tube.
The ends of the wire project frx)m the ends of the tube ; one end
forms the terminal, the other end and a portion of the glass tube
dip into the mercury.

Clean the glass tube and platinum wire carefully, then heat
the exposed end of the platinum red-hot, and insert it in the
mercury in the test-tube, taking care that the whole of the
exposed platinum is covered.

Shake up the paste and introduce it without contact with the
upper part of the walls of the test-tube, filling the tube above the
mercury to jbi depth of rather more than 1 centimetre.

Melt some marine glue until it is fluid enough to pour by its
own weight, and pour it into the test-tube above the cork, using
sufficient to cover completely the zinc and soldering. The glass

33—2

516 PRACTICAL STANDARDS

tube containing the platinum wire should project some way above
the top of the marine glue.

The cell may be sealed in a more permanent manner by coating
the marine glue, when it is set, with a solution of sodium silicate,
and leaving it to harden.

The cell thus set up may be mounted in any desirable manner.
It is convenient to arrange the mounting so that the cell may
be immersed in a water-bath up to the level of, say, the upper
surface of the cork. Its temperature can then be determined
more accurately than is possible when the cell is in air.

In using the cell sudden variations of temperature should as
far as possible be avoided.

The form of the vessel containing the cell may be varied. In
the H form the zinc is replaced by an amalgam of ten parts by
weight of zinc to ninety of mercury. The other materials should
be prepared as already described. Contact is made with the
amalgam in one leg of the cell and with the mercury in the other
by means of platinum wires sealed through the glass.

Notes to the Specification on the Preparation of

THE Clark Cell.

The Mercarous StUphate. — ^The treatment of the mercurous
sulphate has for its object the removal of any mercuric sulphate
which is often present as an impurity.

Mercuric sulphate decomposes in the presence of water into an
acid and a basic sulphate. The latter is a yellow substance —
turpeth mineral — practically insoluble in water; its presence at
any rate in moderate quantities has no efifect on the cell. I£^
however, it is formed, the acid sulphate is formed also. This is
soluble in water, and the acid produced afifects the electromotive
force. The object of the washings is to dissolve and remove this
acid sulphate, and for this purpose the three washings described
in the specification will in nearly all cases suffice. If, however,
a great deal of the turpeth mineral is formed it shows that there
is a great deal of the acid sulphate present, and it will then be
wiser to obtain a fresh sample of mercurous sulphate mther than
to try by repeated washings to get rid of all the acid.

The free mercury helps in the process of removing the acid.

FOR ELECTRICAL MEASUREMENTS 517

for the acid mercuric sulphate attacks it, forming mercurous
sulphate and acid which is washed away.

Pure mercurous sulphate when quite free from acid shows
on repeated washing a faint primrose tinge, which is due to
the formation of a basic mercurous salt, and is distinct fit)m the
turpeth mineral or basic mercuric sulphate. The appearance of
this primrose tint may be taken as an indication of the fact
that all the acid has been removed and the washing may with
advantage be continued until this primrose tint appears. Should
large quantities of this basic mercurous salt be formed the sulphate
should be treated as described in the instructions for setting up
Clark's cells issued from the Physical Technical Institute of Berlin,
Zeitschrift filr Instrumentenkunde, 1893, Hefb 5.

The Zinc Stdphate Solution, — The object to be attained is the
preparation of a neutral solution of pure zinc sulphate saturated
with ZnSO* . 7H,0.

At temperatures above 30"^ C. the zinc sulphate may crystallise
out in another form; to avoid this, 30'' C. should be the upper
limit of temperature. At this temperature water will dissolve
about 1*9 times its weight of the crystals. If any of the crystals
put in remain undissolved they will be removed by the filtration.

The zinc sulphate should be free fit)m iron, and should be tested
before use with sulphocyanide of potassium to ascertain that this
condition is satisfied. If an appreciable amount of iron is present
it should be removed by the method given in the directions already
quoted, Zeitschrift filr Instrumentenkunde, 1893, Heft 5.

The amount of zinc oxide required depends on the acidity of
the solution, but 2 per cent, will, in all cases which will arise in
practice with reasonably good zinc sulphate, be ample. Another
rule would be to add the zinc oxide gradually until the solution
became slightly milky. The solution when put into the cell should
not contain any fi^e zinc oxide ; if it does, then, when mixed with
the mercurous sulphate, zinc sulphate and mercurous oxide are
formed; the latter may be deposited on the zinc and affect the
electr motive force of the cell. The difficulty is avoided by
adding as described about 12 per cent, of mercurous sulphate
before filtration : this is more than sufficient to combine with the
whole of the zinc oxide originally put in, if it all remains free.
The mercurous oxide formed, together with any undissolved
mercurous sulphate, is removed by the filtration.

518 PBACTICAL STANDABDS

The Merouroui Sulphate and Zinc Sidphate Paste. — Although,
after the last washing of the mercnrous sulphate, as much water
as jjoesible may have been drained off, sufficient water generally
remains to necessitate the addition of a very considerable quantity
of crystals of zinc sulphate from the stock bottle, in order to insure
saturation, when the washed mercurous sulphate is added to the
zinc sulphate solution as described in No, 4 of Specification B
appended to the Order in Council.

If the sides of the test-tiibe above the cork be soiled by the
introduction of the paste, the marine glue does not adhere to the
glass ; the liquid in the cell rises by capillary action between the
glue and the glass, and may damage the cell.

Fig. a.

The form of the vessel containing the cell may be varied. In
the H form devised by Lord Bayleigh and modified by Dr Kahle
the zinc is replaced by an amalgam of zinc and mercury. The
other materials should be prepared as already described. Contact
is made with the amalgam in one leg of the cell and with the
mercury in the other by means of platinum wires sealed through
the glass.

The amalgam consists of about ninety parts of pure mercury
mixed with ten parts of pure redistilled zinc. These are heated in
a porcelain crucible to about 100° C, and gently stirred until the
zinc is completely dissolved in the mercury. The amalgam is
liquid while warm, and must be poured into the cell before it
becomes solid on cooling.

FOR ELECTRICAL MEASUREMENTS 519

The vessel containing the element consists of two vertical
tubes. These, as shown in the figure, are closed below and open
above into a common neck, which can be closed by a ground
stopper of glass. The two tubes should be 2 cm. in diameter and
3 cm. in length. The neck should be at least I'o cm. in diameter
and 2 cm. long. A short length of platinum wire is sealed through
the bottom of each tube.

The end of the wire in one tube is covered by a slsiaH quantity
of pure mercury, that in the other tube by the zinc-mercury
amalgam.

Above the mercury a layer about 1 cm. thick of the mercurous
sulphate paste is placed ;• above this, and also above the amalgam,
a layer, also about 1 cm. in thickness, of zinc-sulphate ciystals, and
the vessel is filled up with the saturated zinc-sulphaie. solution.

The zinc-sulphate crystals are obtained by evaporating at a
temperature of less than S0° C. some of the zinc-sulphate solution
prepared as in 4 of the specification.

The stopper is then inserted, leaving a small air bubble above
the liquid, and sealed on the outside with shellac dissolved in
alcohol.

The ends of the platinum wires outside the cell form the two
poles, and should be connected to suitable terminals.

TWENTY SECOND REPORT— IPSWICH, 1895.

APPKMDn PAOB

On Maguetic UniU. By Dr 0. J. Lodge 521

Remarks on the above. By Professor EvbUbtt .... &96

„ ,, „ By Professor Q. Carkt Footbr and

Dr Q. Johnstone Stokey 538

The resistance coils referred to in the last Report as defective
in insulation have been refilled, and up to the present their insula*
tion has proved satisfisu^tory.

The publication of a paper handed in by Dr Muirhead, giving
further results of tests made by Mr E. O. Walker on the coils
made by Dr .Matthiessen twenty-five years ago, and since exposed
to an Indian climate, is deferred until the Cambridge coil, against
which they were tested, can be re-examined by the Secretary.

The set of standards ordered from Germany has only just
arrived. In the course of the next year a careful comparison will
be made between their values and those of the standards of the
Association.

During the year the Committee have had imder discussion a
paper on magnetic units prepared by Dr Lodge and printed as an
appendix to this Report, together with a communication received
firom Dr Everett.

Taking into account the bet that the question of magnetic
units is still under discussion by various bodies, the Committee
wish to come to no hasty decisions, but they recommend for ten-
tative adoption the following terminology : —

1. That, as a unit for magnetic field, a hundred million
"CO. 8. lines" be called a weber.

Note, — A weber added per second at a steady rate to the field
girdled by a wire circuit induces one volt in every turn of that
circuit.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 521

Hence the webers " cut " by a closed wire circuit of n turns
are equal to the quantity of electricity in coulombs thereby im-
pelled round that circuit multiplied by - th its resistance in ohms.

2. That the co.s. unit of magnetic potential or of magneto-
motive force be called a gauss.

Note. — An ampere-turn corresponds to TrrCor 1*2566) gauss.

Hence the number of gausses round any closed curve linked
on an electric circuit is equal to 1'2566 times the number of
ampfere-tums in this circuit.

3. That the termination -ance be used in general for words
expressing the properties of a definite body or piece of matter ;
e,g., resistance, conductance, inductance, permeance, reluctance,
etc.; and that the termination, -iviby or -ility or the like be used
for words expressing the specific properties of a material; e,g,,
conductivity, resistivity, inductivity, refractivity, permeability, etc.

The Committee recommend that they be reappointed; that
Professor Q. Carey Foster be Chairman and Mr R. T. Glazebrook
Secretary.

APPENDIX.
Magnetic Units.

To the British Association Committee on Electrical Standards,

Believing that the Committee is impressed with the convenience
of affixing names to some of the more important units connected
with the magnetic circuit, I beg to suggest the following con-
siderations and recommendations, which I will write out as briefly
as possible. The statements are intended to be precise in their
terms; but in several cases alternative forms of definition are
given.

(1) That the unit coefficient of self-induction, though frequently
useful, is by no means one of the most fundamental units, but
should be defined in a suitably subordinate manner, with reference
to other and more important quantities.

(2) That it would be a mistake so to define it as to discourage
the emplo}rment of the same term for as many other quantities

522 PRACTICAL RTANDARDS

of the same ''dimensions" as possible; especially for the unit
coefficient of mutual induction, and for unit ''permeance."

(3) That the essentially different quantities commonly called
H and B should be careftiUy kept distinct, although their measures
in air have been conventionally so arranged as to be numerically
equal.

Summary of known facts and definitions. — H being the
intensity of magnetic force at a point, or the slope of magnetic
potential (o»),

<k>0 ~ <k>5 = I Hds, along any length ab ;

and in a closed magnetic circuit the circuitation of H is equal to
47r times the total electric current through the area bounded
by the magnetic circuit ; or,

cycle I Hds = circuitation of if = 1 1 4nrcdS = 4BTrG,

or, at any point of space,

curl H = Vy H = iwc ;

where c is current density.

If the electric circuit consists of n turns of wire threading the
magnetic circuit, and each conveying the current Ci, then (7=n(7j.

First quantity to be named,

(4) The first thing requiring a name is this quantity magnetic
potential, sometimes cjiUed nutgneto-motive force; a quantity
spoken of and, measured, not inconveniently but with insufficient
generality, by electrical engineers as amp^re-tums. It has been
proposed (by Mr Heaviside, for instance) that it be called gaussage,
and that its c.o.s. unit be one gauss.

(6) The circuital gaussage round a closed curve is 4ir times

the total electric current through the area bounded by that curve.

In the case of a magnetic circuit wound with wire the gaussage

is 1^ times (tt^ or 1*2566 j the amp^re-tums threading that circuit.

Note. — It may be best to retain the word " gaussage " for the
whole of a closed circuit only, and to speak of the difference of
magnetic potential between two points as the fall of gausses or
the " gauss-fsJl " from a to 6.

FOR ELECTRICAL MEASUREMENTS 523

(6) The gaussage, or gauss-fall, in any portion a& of a magnetic
circuit, is measured by the change in the potential energy of a
unit pole as it moves from a to 6 by any path which involves
neither the cutting of magnetic layers nor the encircling of
currents (a long channel being imagined for its motion through
solid material if necessary). Or, more practically, it is measured
by the induction through a long narrow tube whose ends are at
a and b respectively, divided by the permeance of that tube.
(Cf. Chattock on a magnetic potentiometer, Phil, Mag. July 1887.)

In practice, however, gaussage is frequently calculable fit>m the
ampfere-tums to which it is due.

(7) Intensity of magnetic force, or H, will be naturally ex-
pressed as gauss-fall per centimetre, or the gauss-gradient. For
instance, the earth's horizontal intensity at some place is 0*18
gauss per linear centimetre or 5'4 gausses per foot.

Note. — H should not (strictly speaking) be expressed as so
many lines per square centimetre ; that mode of expression should
be reserved for induction-density B. H is the cause, and should
be thought of as the slope of magnetic potential, B is the effect.
In a medium of so-called unit permeability the two quantities are
numerically equal, but they should not be confounded ; any more
than the slope of electric potential, or electric-intensity (e), should
be thought of as identical with current-density, even in a medium
of unit conductivity.

(8) The gauss-gradient inside a long or closed magnetic
solenoid of length /, wound uniformly with n turns of wire each
convejdng the current Oj, is irrrnCijl^^imfii) where w, is the
total number of turns of wire (in all the layers) to the linear
centimetre.

This is the measure of ^ in the interior of such a solenoid,
quite irrespective of the material with which it may happen to be
filled.

(8 a) That the rotation of the plane of polarisation caused by
any transparent body is equal to the number of gausses between
the points where the ray enters and leaves the body, multiplied
by the appropriate specific constant of its material (sometimes
called Verdet's constant); in other words, that Verdet's constant
may be expressed in degrees or radians per gauss.

524 PRACTICAL STANDARDS

Second qtuintity to be named,

(9) The second quantity requiring a name is the total induction
in a magnetic circuit, also called " total flux/' " total lines/' " electro-
magnetic momentum/' and " electrotonic state/' It is the quantity
whose time-rate of variation gives the voltage induced in an
electric circuit surrounding it once. It is proposed that its prac-
tical unit be called a ** weber/' and be defined as equal to 10" ccs.
lines (or unit tubes) of induction.

(Denote the quantity for the present by N, and its density

by A)

Summary of known facts. — e being the intensity of electric
force, or the slope of electric potential, or the volt-gradient at a
point ; the circuitation of 6 = the induced E.M.F. in a closed circuit
s the rate of change of induction through it ; or,

cycle I ede^ E^^-N^- 11 BdS;

or, at any point of space,

jS«-curle = -FV«.

Through a simple closed electric circuit of constant resistance,

N^^JEdt^JRCdt^RQ.

(10) The total induction through any area may be practically
measured by suddenly surrounding it with a closed wire circuit of
n turns connecting the terminals of a ballistic galvanometer, and
measuring the quantity of electricity thereby impelled through
the galvanometer. The induction is equal to the quantity so
impelled, multiplied by 1/nth the resistance of the circuit If the
quantity is one coulomb, and the resistance one ohm, the induction
is 1/nth of a weber.

Or, otherwise, if the induction through any boundary changes
at the rate of one weber per second, the E.M.F. excited in that
boundary is 1 volt.

In the case of a spiral wire circuit through which induction is
varying at the rate of a weber per second, one volt is excited in
each turn of the ynre.

FOR ELECTRICAL MEASUREMENTS 525

(11) Another mode of measuring the total induction through
an area is to surround that area with a movable electric circuit
of n turns of wire conveying a known current, and to measure
the potential (or mechanical) energy of the circuit under those
conditions. The induction is equal to the potential energy of the
circuit divided by n times the current circulating in each turn of
wire.

Or, if the induction through a simple circuit carrying one
ampere is one weber, the potential energy of the circuit is one
joule.

Derived quantities.

(12) Induction-density, or B, may be expressed as so many
webers per unit area; say per square centimetre or per square
inch, or whatever is preferred for practical purposea

For instance, the earth s horizontal induction-density at some
place is 0*18 co.s. unit « 018 x 10"* weber per square centimetre
ss 18 microwebers per square metre.

(13) The inductivity (fi), or absolute permeability of a medium
at any point under specified circumstances, is the ratio o{ Bto H
at that point, and under those circumstances. In many substances
this ratio is far from constant. [It may be expressed in terms of
heniys or other units of permeance per unit length (see below),
instead of in CG.s. units, if convenient. For example, the in-
ductivity of air is y^th of a microhenry per centimetre, or one
millihenry per kilometre.] More explicitly it is measured by the
webers per unit area divided by the gauss-SeJl per unit length;
in other words, by the ratio of the weber-density to the gauss-
gradient

(14) The relative inductivity of a substance as compared with
that of empty space (a^/mo) Q^ay be called simply its ** permeability"
as at present, and is a mere number.

(Its electrical analogue is specific inductive capacity (K/Ko),
as contrasted with absolute electric inductivity (K) ; which latter
could be defined in practical units as the ratio of the coulombs
displaced per unit area to the volt-gradient)

526 PRACTICAL STANDARDS

Third quantity to be named.

The third quantity for whose unit a name is required is some
form of ratio between the two fundamental quantities whose units
are here named after Weber and Gauss respectively. It has been
practically decided in America that this unit shall be named after
Pro£ Henry, of Washington, and that it shall equal 10" CQ.s.
units, being equivalent to the earth-quadrant or secohm; but
the precise mode of definition has not yet been finally agreed
upon.

There are two quantities of the same physical dimensions to
which the name is applicable, viz., the coefficient of self or mutual
induction of a coil or coils of wire, and the permeance or inverse
reluctance of a magnetic circuit.

The most logical order is to define permeance first, as the ratio
of the webers of induction to the exciting gaussage, and then to
say that the inductance of a coil of n turns of wire is n* or 47m*
or *4?rn' times the permeance of the magnetic circuit which it
embraces, €uxx>rding to the units of gaussage and current which
have been decided on.

If the units of gaussage and current are both the CG.s. units,
then 47m' is the numerical factor connecting inductcmce with
permeance.

If the co.s. unit of gaussage is adopted along with the ampere-
current, then *47ni* is the factor.

But if the circulation of H due to one ampire'tum is adopted
as the practical unit of gaussage, then n' is the feustor ; and the

permeance of a cylinder, instead of being simply ^ , is ^. ■ .

The apparent simplicity of this last sjrstem has much to
recommend it for commercial use, though it will complicate the
specification not only of permeance but also of magnetic fields
and potentials; but some inconvenience due to the unfortunate
definition of the unit pole, and the only less unfortunate definition
of the practical unit of current, cannot be avoided ; and our aim
must be to place the inconvenience where least likely to be felt
in everyday work.

FOR ELECTRICAL MEASUREMENTS 527

First system.

We will begin with the more logical system, and with general
statements which apply to both.

(15) In a complete magnetic circuit the ratio of the total
induction to the corresponding gaussage under specified conditions
is called the '' permeance " of that circuit under those conditions.
It is not in general constant.

Or, the permeance of any solenoidal portion of a magnetic
circuit, if free from intrinsic magneto-motive force or magnetic
boundary layers, is the webers through it divided by the gausses
between its ends.

(16) The practical unit of permeance is that of a circuit in
which a weber is excited by a gauss. Its reciprocal is the unit
of reluctance. The practical unit so defined is 10^ CG.s. units.'

Examples. — The permeance of a cubic metre of air to parallel
induction from one face to the opposite is 1 microweber per gaus&r

Under circumstances such that the permeability of iron is
400 times that of air, the permeance of an iron ring of one deci-
metre cross-section and one metre in mean diameter is

p- = 20r = 100 CG.S. = again one microweber per gauss.

It is, perhaps, a question whether this amount of permeance
could be called "a microhenry'* without confusion.

Explanation. — The inductance (or 3elf-induction-coefficient) of
an electric circuit consisting of n turns of wire, so far as it is
constant, is defined to be equal to n times the induction produced
through it by a current of one ampere in each turn. But the

gaussage due to n ampfere-turns is -rr^ or *4im; hence the in-
ductance of a wire coil la '47rn* times the webers caused by each
gauss in the magnetic circuit surrounded by it ; %.e., is *47m' times
the permeance of that circuit considered as constant.

(17) A coil of wire threading n times a complete magnetic
circuit of unit permeance under any given circumstances is said
to have '4?m' units of inductance under those circumstances;
and in general the inductance of a coil of n turns is '4nrn* times
the permeance (as above defined) of the magnetic solenoid enclosed
by it. (The permeance may here be considered variable.)

528 PRACTICAL STANDARDS

[With the amp^re-tum as unit gaussage the *47r is prefixed
similarly to both inductance and permeance, so that only the
factor n' is needed to convert one into the other. See below.]

(18) The c.o.s. unit of inductance is equal to n times the
induction excited through a coil per c.G.s. unit of current in every
turn of wire ; whereas the practical unit of inductance is n times
the webers excited per ampere ; hence the practical unit of in-
ductance is 10" times the CG.s. unit.

The practical unit is called a "henry." (It has also been
called secohm and quadrant.)

Example, — If the above iron ring were wound closely with

1000 turns of wire, the coil would have a coefficient of self-indue*

4rtr
tion equal to ^tt or 1^ heniys whenever the permeability of the

iron was 400.

A coil of 20,000 turns of wire, wound closely on the same core,
would have an inductance of 1^ heniys if it contained air or other
non-magnetic substance.

AUemative mode of definition of inductance on first system.

In view of one of the above practical methods of measuring
induction experimentally, the inductance of coils of wire, both self
and mutual, may be defined more directly thus : —

(19) When of two simple circuits one conveys a current, the
other in general has induction caused through it; and the ratio
of the induction through either to the inducing current in the
other is called the mutual inductance of the circuits.

(20) Of two coils, with n and n' turns respectively, the total
mutual inductance is to be reckoned for every turn of wire on
each coil, and is therefore nn times the inductance of the mean
turn of one coil on the mean turn of the other.

(21) The mutual inductance of two coils is ^wnn' times the
permeance of the largest magnetic solenoid which threads both.
For if every turn of one conveys a current C, while every turn of
the other surrounds an induction N' in consequence, the per-
meance of the magnetic solenoid threading the second coil is
P TsN'/^TmC; but the total effective mutual induction, J/C,
through all the turns is n'N'; hence M ^4nmn'P.

FOB ELECTRICAL MEASUREMENTS 529

(22) When two coils each conveying one ampere are con-
stantly connected by one henry of mutual inductance, the kinetic
energy of the field due to their mutual action is one joule.

(23) If the self-induction coefficient of a coil is being con-
sidered, its total inductance may be taken as n* times the
inductance of the mean turn ; that is n times the total induction
through it divided by the inducing current. Or the weber-tums
per ampere give the self-inductance in henrys.

(23 a) The expression weber-tums, to signify the product of
the total field into the number of spires surrounding it, though
at first sight not in precise correspondence with the phrase
ampfere-tums wliere the current circulates in the spiral instead
of forming its core, is really accordant with it, because a spiral
and its core are geometrically interchangeable.

(24) The practical unit of inductance, whether self or mutual,
is called a henry ; and a coil of n turns has a henry of inductance,
on itself or on another of n' turns, when an ampere in , one
maintains 1/n' weber of induction (through itself or) through the
other.

(25) When the induction through a coil varies for any reason
at the rate of one weber per second, the E.M.F. generated in each
turn is one volt.

(26) When the inductance of a coil is one henry, on itself or
on another, a small variation of current in it at the rate of one
ampere per second induces an e.m.f. of one volt in itself or in the
other.

(27) When the inductance of a coil conveying one ampere
varies at the rate of a henry per second, the induced E.M.F. is one
volt.

(28) When the self-inductance of a coil is constantly, or on
the average, one henry, while an ampere current is generated in
it, the kinetic energy of the field due to that ampere is half a
joule.

Second system.

(29) If instead of taking a gauss as equal to a co.s. unit of

magnetic potential, we take the circulation of H caused by one

ampere-turn as the practical unit of magneto-motive force, we

10
shall have 1 ampfere-turn = j— CG.S. units of gaussage.

B. A 34

680 PRACTICAL STAHDABD6

(30) The practical unit of permeance will then be that in
which a weber of total induction is excited by each amp&re-tum ;
in other woid9» it will be 4f7r x lO' c.a.8. units of permeance.

(31) And the practical unit of inductance will be that of a

coil in which an ampere in every turn excites - th of a weber

through every turn ; that is to say, the inductance of a coil will
be n' times the permeance of the magnetic circuit surrounded
by it.

(32) The difference between inductance and permeance is
only one of reckoning. Permeance is webers per ampere-tuna.
Inductance, is weber-tums per, ampere. *

SUMXABT OF THE ADVANTAGES OF THIS SORT OF MODB OF

DEFINING Unit Inductance.

The special feature of this mode of defining the ** henry " is
that it makes inductance depend on the simple ratio N/C, or
weber-tums per ampere, instead of on something more compli-

cated.

It might possibly be defined as the ratio dN/dC, that is, as
proportional to the tangent of the slope of the B : H curve ; and
such a definition would emphasise its variability; but certain
practical advantages would be lacking, because it would be
detached from any connexion with the permeance of the circuit
The N/C ratio on the other hand instantly connects itself with
permeance, and represents the slope of the secant drawn fix>m
the origin to any point of the B : H curve. It exhibits the
variability sufficiently ; making the inductance reach a maximum
at the shoulder of the curve, and then slowly decrease as saturation

sets in.

It is sometimes said — but the mode of expression is, to say
the least, very inconvenient — that there are three different
principles on which to define L, all leading to a different result :
viz., numbe^ring them inversely, but giving them in their usual

order : —

(3) Energy W^^LC\

. ' • '

(2) E.M.F. E ^LdC/dt,

(1) Total induction... JV =Za.

FOR KLEGTRICAL MEASUREMENTS 581

But the real facts to be expressed are not here exhibited.

The real fiicts are

(1) N^LC,

(2) E = dNjdt,

(3) dW^GdN.

The essential thing to name is therefore N\ and if lO" c.G.s.
lines or unit tubes be called a " weber," or a " weber-tum," then
a volt is a weber or a weber-tum per second, and a joule is a
weber-amp^re-tum. Nothing can be handier than that; and a
henry can be defined as a weber-tum per ampere.

Instead of saying as above that there are three ways of defining
Ly the simplest thing is to say that two of the three equations as
first given above are incorrect, except for the special and in practice
comparatively rare case when L is constant. Written out correctly
they stand as follows : —

(1) N^LC,

(8) Tr-iXC* + ir(7*dX.

It is then obvious that (2) and (3) are too complicated to base
a definition upon, and that the first alone. gives a feasible system.

The fact that L is decidedly not in general constant deprives
the henry of any such importance as the ohm possesses ; moreover,
it refers explicitly to rather a special thing, viz., a coil of wire,
and that under specified conditions, if it contain iron ; hence it
would be rather absurd to name this alone of all magnetic units.
In the above communication, in addition to a certain mode of
defining the henry, it is urged that unit total induction be named
too ; for this is the quantity which is of real engineering import-
ance— this is the quantity to attain which field-magnets are built,
and in the midst of which armatures are spun.

It is also urged that it would be convenient if unit magnetic
potential could likewise be named, since electrical engineers have
shown that they have need of some such unit for the exciting
cause of induction, by their practical employment of the phrase
** ampere- turns." The introduction of a gauss unit, in some form
not too obviously limited to the case of a wire-wound coil, would
assist teaching and would clarify magnetic ideas

34—2

532 PRACTICAL STANDARDS

The present writer does not presume to decide between the

two alternative systems of defining " the gauss " as given above :

viz., the c.G.S. unit on the one hand, and the amp^re-tum on the

other.

Oliver J. Lodge.

Liverpool, December 9, 1894.

Postscript — Another subject for discussion is whether L had
better not be defined as dN/dC; with permeability as fA = dB/dH
to correspond. This would make the three equations on p. 531
stand thus:^-

(n N=JLdG,

(2) E^LC,

(3) w=cir.

A letter just received from Mr Heaviside indicates that he would
probably favour this course, and there is evidently much to be
said for it. I need hardly add that he contemns my temporising
method of dealing with the 47r nuisance.

It need hardly be said that in the last resort it rests with
practical men to employ or decline any suggested system of units.
Those who daily deal with the quantities under consideration are
the best judges of the utility or otherwise of a suggested unit,
provided always that they take the trouble to give it a fair trial,
and see how it works in practice. It may be hoped that the
above or similar suggestions will meet with criticism at the hands
of such men, and in order to make a beginning of criticism I
asked the Departmental Lecturer on £lectrotechnics at University
College, Liverpool (Mr F. G. Baily), to consider them with special
reference to

(1) The large size of the weber and henry units ;

(2) The handiest definition of the gauss ; and

(3) The least troublesome mode of bringing in the 4nr.

His reply, which is annexed, covers these points, and also
incidentally refers to the quantity called / or intensity of
magnetisation.

Now, as must often have been pointed out, the equation
B = H + 4nrl is a barbarous one, involving as it does quantities
of different dimensions in one equation. Its true meaning is of

FOR ELECTRICAL MEASUREMENTS '533

course B^ ^H'^(fjk'-fjLo)H; which, although algebraically only
a roundabout method of writing B^fiH, is yet convenient, as
exhibiting separately the part of the induction due to the ether
and the part due to a material medium.

The customary convention of further denoting (ji^fh)lfh hy
the symbol ^k, and then christening kH as the magnetisation /,
is likewise convenient. With this definition ic is a pure number,
and / is a gauss-gradient or field-intensity. Another, but on the
whole less satisfactory, definition, viz., the omission of /io from the
denominator, would make x of the same dimension as /i, and /
an induction-density.

The pull between two parallel magnetised surfaces of area A is
^ABH-i-^, that is to say, NH/Svy and is therefore measured in
webers multiplied by the gauss-gradient, or in joules per centi-
metre. But to maintain an induction-density B in air requires
a gauss-gradient £//!«, hence we might write the pull across an
air-gap as i\r* -«- STTfi^A, If the induction-density across ah air-gap
is expressed in microwebers per square centimetre the tension
there comes out in units of which 2,500 would make an atmo-
sphere; or, roughly, in pounds per square foot.

As for the strength of a magnetic pole — a quantity which,
though fundamental in one sense, is seldom really dealt with —
it will naturally be expressed in ergs per gauss, or in joules per
gauss if it is very strong.

Mr Baily's chief practical suggestions are first that a special

unit of permeance, other than the henry, is desirable ; and next

that the 47r/10 had best be thrown on to the /x, so as to keep the

gauss equal to one ampfere-tum.. The fact that the inductivity

of air will then cease to appear in its artificial garb of unity may

even be regarded as a positive advantage, because its existence

will then be less likely to be ignored. But I much fear that the

amp^re-tum as unit of gaussage, so near the CG.s. unit in size

and yet not equal to it, will be awkward and may lead to

mistaken

O. J. L.

6S4 PRACTICAL STANDARDS

UNiriBsmr Ck>LLBoiB^ Livbbpool,
January 15, 1896.

Dear Professor Lodge,

In reference to the sizes of the magnetic units proposed by
you, I find that the weber lO' co.s. would only be used in frac-
tions. The largest dynamo thai I know of has a magnetic flux,
or, as you propose to call it, an induction, of '6 weber. From this
the value will go down to about Ol in small motors. These
figures are, however, by no means inconvenient.

Transformers will be rather smaller. In these the weber-tum
is a convenient size and an interesting quantity, as it is given by
\ of the mean volts per cycle, or, more accurately expressed, mean
volte per unit frequency -r 4. Its numerical value will lie between,
say, \ and 60, according to the volts and the firequency; but it
gives no indication of the size of the transformer.

The henry, 10* ca.s., is also large. The inductance of choking
coils would in general be jGractions of a heniy. The inductance
of the winding of a transformer has no very important meaning,
but it has a convenient size. Measuring it as mjean volte per unit
frequency 'T' four tim^ee the open circuit current in ampiree, the
inductance of the primary coil on a 2 H.P. closed magnetic circuit
1,000-volt transformer would be about 40 henrys.

The inductance of pairs of cables would run fix>m 100 to 1,000
microhenrys per kilometre, but the value would vary with the
arrangement.

The induction per unit area is good ; having a value in practical
work fi^m 1,000 to 20,000 co.s. units, it is given by 10 to 200
microwebers per sq. cm.

Gaussage would be about 60 in small transformers, up to 40,000
in large dynamos. The latter could be conveniently reckoned in
kilogausses. To make the gauss a 1 ampfere-tum appears to have
great advantages in practice, and connects it directly with its
usual source.

The idea of permeance is very useful, and the identification
of its dimensions with those of inductance is neat But I think
it is liable to cause confusion, for the permeance of the core of a
coil will be a different number of henrys from the inductance of
its wire. Moreover, the argument as to identical dimensions
might equally be applied to the case of amperes and gausses.

FOR ELEOTRICAL MEASUREMENTS 535

I would therefore have a new unit strictly cofmected with the
henry, so that inductance «* n* x permeance in a coil>of:n turns*

As to the units of permeance : with the above meanings of
gauss and weber the permeance of a circuit would be '4irful/10/,
as yoii point out, instead of ftAjl, But I wish to suggest a dhange
in. the method of reckoning, namely » still to retain, thier value of

cm.'

., webers , ,.,.^ wiebeis per sq.
the permeance as , and permeability as ■. — ^-t-~-

gausses' ^ *^ * gausses {)er cm. ■

therein giving up the convention of unit permeability, of space,
and giving it the value 1*2666 x 10~* units of permeance for a unit
cube. In this way both the troublesome 10^* and . 47r/I0 are
dealt with in an easily intelligible way. To avoid the high power
of 10 it may be measured in micro-units of permeance, so that
permeability of space and air » '012566 micro-unit of permeance
for a unit cube, and permeability of soft iroias=up tO;25 micro-
units for a unit cube. Thus we have permeance — il/x//, where /t
is to be obtained from tables of its value, which can. easily be

weber-turns

altered to this method. Inductance then becomes — ^-^-^ ;

amperes

. webers

or ti" =n" permeance.

gausses '^

[Of course your phrase "weber-turns per ampferes" means tb^
same as the above webers -f- amperes, and does not necessarily
mean the weber-turns caused by one ampere.] - ; ..

It may be objected that the c.G.s. units of stren^h of field,
unit magnetic pole and intensity of magnetisation do hot bear
any simple relation to these practical unit& This is chiefly
important in the use of the magneto-metric measurement of iron,
and in the measurement of the mechanical form of attraction
between two magnetic surfeu^es in contact. But the expressions
are not in reality much complicated; e.g., present O.G.S; unit of
intensity of magnetisation is given by 4nrI^(jArrf^iyM, wher^
/!« a permeability of space »1, and H^CG.s. unit of jtnagnetic
force. This becomes 4m'I^(jA ^ fig)H' W, where J5f' is th^
gauss-gradient in the magnetic substance, and fig » '012566 micror
unit of permeance for a cubic centimetre.

As the single magnetic pole is unchanged, the force on it will
be - strength of pole x gauss-gradient x 1*2566 ; but as this is not
a calculation of frequent occurrence, except in magnetic surveys,
the complication will not be serious. Other mag;netic relationshipii

536 PRACTICAL STANDARDS

are almost entirely of academic interest only, and would be carried
out in C.G.S. units. Also the transition would present no diffi-
culties to people with a little scientific knowledge.

I am of opinion also that as the legal volt has no direct con-
nexion with induction and velocity of motion, it is not necessary
to define the practical units as they are defined absolutely. That
is, ohm and ampere are the starting points, volt is obtained fix)m
them, weber &om volt, gauss from ampere, permeance from weber
and gauss, henry from weber and ampere or fix>m permeance, and
so on. This is much more easily explained to practical and un«
scientific men than the absolute derivations are, and it is the

order in which they learn them.

Francis G. Baily.

Remarks on the Above (especially on pages 530 — 532).

According to the proposal of the Chicago Chamber of Delegates,
the quantity which we call '' inductance," and which is to be ex-
pressed in '' henrys," is defined by the equation

E= L -j-,ioT self-induction,
at

or E — M-j~,{oT mutual induction,

both being comprehended in one definition, the inductance L or

M being calculated by dividing E in volts by -, in amperes per

N
second. This implies that L or M is not to be defined as ^ ,

but as ^.

N dN

I think names are desirable both for -^ and (or-rp » I would

suggest that the former be called " the total inductance," and the
latter "the difierential inductance." The distinction would be
somewhat analogous to the distinction between the " mean specific
heat 6t>m 0^ to f " and the '' true specific heat at f** Both total
and differential inductance should be expressed in " henrys," for
they are quantities of the same kind, and when there is no iron,
etc., in the field they are equal.

I think that the above mode of definition, involving as it
does no magnitude except current and time, is more readily

FOR ELECTRICAL MEASUREMENTS 537

comprehended than Dr Lodge's proposed definition, in which the
magnitudes involved are current, flux of induction, and the
number of convolutions of the coil through which the flux passes.
In the definition proposed by the Chicago delegates the con-
sideratijjns of the number of convolutions does not enter.

For a circuit or two circuits not having iron, etc., in the field
we may define inductance (in henrys) as the E.M.F. (in volts) due
to variation of current at unit rate (one ampere per second).
When the field is modified by the presence of magnetic material
the above will be the definition of "differential inductance."

The " total inductance " for any specified strength of current
will be the mean value of differential inductance for equal incre-
ments of current fix)m zero up to the specified strength.

I would suggest similar nomenclature in the case of per-

meability: -j^ should be called differential permeability, and

jf total permeability.

In some respects "mean" or "average" would be a more
correct designation than "total"; but these words would be
liable to be misunderstood as referring to an average taken over
the different parts of the body or circuit. . "Total" is to be
understood as standing for "calculated on totals."

As regards the magnitude of the unit of inductance. While
I agree with Mr Heaviside and Dr Lodge that the unit pole
ought to have been so defined that the mutual force between
two poles is equal to their product divided by the surface of a
sphere whose radius is their distance, a definition which would have
made the line-integral of H due to a current C equal to C itself
instead of to 47rC, I deprecate a mixing up of the two systems.
So long as we employ our present unit of intensity of magnetic
field, which results from our present definition of the unit pole,
we cannot consistently reckon the line-integral as equal to the
amp^re-tums. It must be reckoned as 4i7r times the ampere-turns,
and the flux N must be reckoned as 4m'fi times the amp^re-tums.
The practical inconvenience of retaining the factor 47r cannot be
considerable, for it is as easy to tabulate the values of 4nrp, as
the values of fi.

Next as regards "permeance." I do not think it can con-
veniently be reckoned in heniys. I would rather reckon it in

538 PRACTICAL 8TAMDA&DS POB ELBCTRIOAL MEASUREMENTS

"webers per amp^re-tum," which would be written "web. per
amptu " ; and there can be no possible doubt as to the meaning
intended when once we have fixed the magnitude of the '' weber."
There seems to be no difference of opinion as to what this mag*

nitude should be. It is fixed by the relation -ff = ->- , iS being

in volts, If in webers, and t in seconds. This is in accordance
with Dr Lodge's proposal ; but Dr Lodge has not explicitly recom-
mended any name for the j^ysical quantity which is measured in
webers. Shall we call it '' weberage "* ? It greatly needs a name ;
for '^ induction '' may mean B instead of the surface integral of B,
besides having many other meanings.

When permeance varies according to the strength of current,

iV ' .

I would distinguish between " total permeance " -^ and " differ-

ential permeance - -^ .

As regards " gaussage *' and " gauss fall." I think the names
will be convenient in the senses proposed by Dr Lodge, but I
cannot agree with his selection of a unit of measurement. The
present definition of the unit pole (on which the present unit
current is based) requires us to equate the line-integral in question
to AtwnC,

To be consistent we must reckon gaussage as equal to 4i7r times

the number of amptus. Dr Lodge's proposal is to reckon C, not

in amperes but in C.Q.S. units, thus introducing, as it appears to

me, an awkward breach of continuity.

J. D. Everett.

Professor Carey Foster has written objecting to the term
" gauss-gradient," instead of " magnetic gradient " ; he prefers the
latter, just as he would prefer "temperature-gradient" to "degree-
gradient."

Dr Johnstone Stoney has also written, urging strongly that
not the c.o.s. unit of magnetic potential, but one-tenth of this
quantity, should receive a name, in order to make it harmonise
with the ampere series; and further recommending that the
names " weber " and " gauss," as above suggested, should be inter-
changed.

TWENTY-THIRD REPORT— LIVERPOOL, 1896.

-.1
JkPPEMDIZ PAaH

I. Exiraeit from LsUisn reoeived^ dealing toith the Qtte$tum of the

Unit of Beat , 544

II. The Capacity for Heat of Water from 10° to 20" C. referred to

its Capacity at W C. <u Unity . . . . 554

III. Becalculation of the Total Heat of WcUer from the Experiment^

of Regnauh and Roidand. By W. N. Shaw . . 55$

The comparison between the set of standards ordered from
Germany — referred to in the last Report — ^is not yet completed.
The work will be continued during the current year.

At the Ipswich Meeting of the Association the question of
a standard thermal unit was referred to the Electrical Standards
Committee, and has been under their consideration during the
year.

After the Ipswich Meeting Mr E. H. Griffiths sent the follow-
ing letter to a number of physicists in various foreign countries,
together with a copy of the paper* he had communicated to the
Association : —

Herewith I forward you a copy of a recent oommunication to the Philo-
eophical Magazine, in which I have endeavoured to call attention to the
unsatisfActory nature of our present system of thermal measurements.

At the Ipswiuh Meeting of the British Association the oonsideratiun of the
question of a standard thermal unit was referred to the Electrical Standards
Committee.

As a member of that Committee I now approach you with a request that
you will communicate to me any suggestions which you may regard as cal«
culated to assist our deliberations on the subject.

I am anxious to lay before the Committee the opinions of the leading
authorities of all countries ; I trust, therefore, that you will favour me with
some expression of your views, particularly as to the nature and magnitude
of the thermal unit (if any) that you would reoommend for adoption.

Unless you state that I am to regard your reply as ** for Committee only ''
or ** private,'' I shall conclude that you have no objection to its publication.

The importance of arriving (if possible) at some general agreement r^ard-
ing the thermal unit will, I hope, be accepted as a sufficient excuse for thus
troubling you.

* Pkil Mag,, November) 1S95.

540

PRACTICAL STANDARDS

Copies of the circular letter,
Unit, were sent to the following

Professor Abbe, Washington, U.S.A.
Professor Ames, Baltimore.
Professor Bartoli, Pavia.
Professor Bams, Providence, R. I.
Professor Benott, Sevres.
Professor Berthdot, Paris.
Professor Boltzmann, Vienna.
Professor Cbillencbir, Montre^
Dr Chappuis, Bureau International,

S&vres.
Dr Curie, Paris.
Professor Dieterici, Hanover,
Professor Dom, Halle.
Professor Du Bois, U.S.A.
Professor Willard Gibbs, Yale, U.S.A.
Dr Quillaume, Bureau International,

Sevres. .
Professor Hall, Harvard, U.S. A.
Professor Himstedt, Freiburg.
Professor Hittorf, Muuster.
Professor Joubert, Paris.
Professor Kayser, Bonn.
Professor Kohlrausch, Berlin.
Professor de Kowalski, Freiburg,

Switzerland.
Dr S. P. Langley, Washington, U.S. A.
Professor Landolt, Berlin.
Professor Le Chatelier, School of

Mines, Paris.

and of the paper* on the Thermal

Professor Lippmann, Paris.
Professor Victor Meyer, Heidelberg.
Professor Nemst, GOttingen.
Professor Nichols, Ithaca, U.S.A.
Professor Olszewski, Cracow.
Professor Ostwald, Leipzig.
Professor Overbeck, Tubingen.
Profe^r Paschen, Hanover.
Professor Planck, Berlin.
Professor Pellat, Paris.
Professor Pemet, Zurich.
Professor Potier, £oole Polytech-

nique, Paris.
Professor Quincke^ Heidelberg.
Professor Remsen, Baltimore, U.S.A.
Professor Rowland, Baltimore, U.S.A.
Professor Runge, Hanover.
Professor Schuller, Budapest.
Professor Stohmann, Leipzig.
Professor J. Thomsen, Copenhagen.
Professor Van 't Hoff, Amsterdam.
Professor Vaschy, £oole Polytech-

nique, Paris.
Professor E. Warburg, Berlin.
Professor Wartha, Budapest.
Professor Weber, Zttrich.
Professor £. Wiedemann, Erlangen.
Professor G. Wiedemann, Leipzig.
Professor WuUner, Aachen.

Replies were received from the following, and the Committee
desire to thank those who so courteoosly responded to Mr Griffiths'
inquiry for their very valuable assistance.

Professor Kichols, Ithaca, U.S. A.
Professor Olszewski (and Colleagues),

Cracow.
Professor Ostwald, Leipzig.
Professor Paschen, Hanover.
Professor Planck, Berlin.
Professor Quincke, Heidelberg.
Professor Remsen, Baltimore, U.S.A.
Professor Rowland, Baltimore, U.S. A.
Professor Runge, Hanover.
Professor Stohmann, Leipzig.
Professor WuUner, Aachen.

Professor Ames, Baltimore.
Professor Boltzmann, Vienna.
Professor Callendar, Montreal.
Dr Chappuis, Bureau International,

Sevres.
Professor Dieterici, Hanover.
Professor Dom, Halle.
Dr Quillaume, Bureau International,

Sevres.
Professor Le Chatelier, School of

Mines, Paris.
Professor Victor Meyer, Heidelberg.
Professor Nemst, Gdttingen.

* Phil. Mag., November, 1895.

FOR ELECTRICAL MEASUREMENTS 541

Extracts from such replies as contain definite suggestions
bearing on the question of the^ unit of heat are printed in
Appendix I.; the letters have been translated, and those which
merely give general approval to some such scheme as that out-
lined have not been included. No replies were received adverse
to the suggestion that an endeavour should be made to secure
common agreement in the matter.

The concluding propositions of Mr Qriihths' paper were subr
stantially as follows: —

(I) To adopt as the theoretical unit of heat a multiple
(42 X 10") of the erg.

(II) To adopt as the practical unit of heat, the heat required
to raise 1 gramme of water l"" C. of the nitrogen thermometer at
some temperature C" C. as given by that thermometer.

(III) To adopt provisionally some formula expressing the
specific heat of water in terms of the temperature over a range
of, say, 10^ C.

If the number, 42 x 10" ergs, be adopted for the theoretical
unit, then, according to the experiments of Rowland, the theoretical
and the practical unit agree, provided that the temperature f C.
be 10° C.

Mr Griffiths, in the paper already referred to, has made a
comparison of the results obtained by Joule, Rowland, Schuster,
Miculescu, and himself, for the amount of energy required to raise
1 gramme of water 1"" C. at various temperatures. The results
differ according as the readings of Joule's mercury thermometer
are reduced to the scale of Rowland's air thermometer, or to the
scale of the nitrogen thermometer, as has been done by Schuster.

In the first case the mean values are —

At 10° C. (41-971 ± 023) x 10" ;

and at 15° C. (41891 ± 023) x 10" ;
and in the second —

At 10° C. (41-958 ± -029) x 10« ;

and at 15° C. (41875 ± '029) x 10«.

Tables of the values of the specific heat of water between
10° C. and 20° C. have been calculated by Mr Griffiths, and are
given in Appendix II.

The Committee have made an analysis of those replies which
contain definite suggestions.

542 PRACTICAL STANDARDS

Most of the writers wish to see some multiple of the eig
adopted as the theoretical unit, but there are differences of opinion
as to the multiple to be chosen.

Thus, Professors Dom and Wtillner, Dr Chappuis, and Pro-
fessor Ames would prefer 42 x 10" ergs. Professor Ostwald,
Professor Olszewski and his colleagues, and Professor Callendar
suggest 10' ergs. Professor Planck and M. Le Chatelier suggest
10" ergs, or in the case of the latter, as an alternative, 5 x 10^.

Professors Rowland and Nichols consider the ice unit as
theoretically best ; the latter, however, would be willing to adopt
42 X 10" ergs as the theoretical unit, while Professor Rowland
writes: "From a practical standpoint, however, the unit depending
on the specific heat of water is certainly the most convenient. It
has been the one mostly used, and its value is well known in
terms of energy."

There is £drly general agreement in the view that as a practical
unit the heat required to raise 1 gramme of water 1° C. at some
fixed temperature must be taken, but views differ as to the tem-
perature which it is most convenient to choose.

Mr Qriffiths suggested the nitrogen thermometer as the standard
of temperature. The French physicists agree in the opinion that
the hydrogen thermometer should be adopted, and reasons are
given for this in the letters of M. Quillaume and M. Chappuis.
The Committee concur in this view.

The Committee are of opinion that Mr Griffiths' paper, and
the replies received by him, show clearly that it is desirable to
come to an agreement as to the definition of the unit of heat.

They understand that a Committee of the French Physical
Society have the question at present under consideration, and they
hope it may be possible for the Electrical Standards Committee
of the British Association to co-operate with this Committee and
with representatives of other foreign countries in the matter.

The Standards Committee have provisionally approved the
following propositions, with the view of opening international
discussion of the question. They propose to send the propositions
to representative bodies throughout the world, with a letter
stating that they have been provisionally approved, inviting
further discussion, and asking those bodies to take the steps
which seem to them most desirable in order to secure international
agreement on the matter.

FOR KLKCTBICAL BfSASUREMENTS 543

Proposition I. — For many purposes heat is most conveniently
measured in units of energy, and the theoretical CQ.s. unit of
heat is 1 eig. The name Joule has been given by the Electrical
Standards Committee to 10' ergs.

For many practical purposes heat will continue to be measured
in terms of the heat required to raise a measured mass of water
through a definite range of temperature.

If the mass of water be 1 gramme, and the range of tem-
perature 1° C. of the hydrogen thermometer from 9'5** C. to
10*5*^ C. of the scale of that thermometer, then, according to the
best of the existing determinations, the amount of heat required
is 4*2 Joules.

It will, therefore, be convenient to fix upon this number of
Joules as a secondary unit of heat.

This secondary thermal unit may be called a " Calorie."

For the present a second proposition is

Proposition II. — The amount of heat required to raise the
temperature of 1 gramme of water 1"^ C. of the scale of the
hydrogen thermometer, at a mean temperature which may be
taken as 10° C. of that thermometer, is 4*2 Joules.

If further research should show that the statement in II. is
not exact, the definition could be adjusted by a small alteration
in the mean temperature at which the rise of V takes place.
The definition in I. and the number (4*2) of Joules in a Calorie
would remain unaltered.

In Appendix II. a table is given showing the capacity for
heat of water between 10"* C. and 20"" C, and in Appendix III.
the value of the total heat of water has been calculated by
Mr Shaw from the experiments of Regnault and Rowland.

Professor J. V. Jones has, during the year, calculated the
correction to be applied to the value of the international ohm in
absolute measure given by him at the Oxford meeting (1894), in
consequence of the ellipticity of the standard coil used in his
experiments. The required correction is '00684 per cent., and
the corrected value of the international ohm is '99983 x 10"
absolute units.

In conclusion the Committee recommend that they be re-
appointed, with a grant of £6; that Professor G. Carey Foster
be Chairman, and Mr R. T. Olazebrook Secretary.

544 PRACTICAL STANDARDS

APPENDIX I.

Extracts from Letters received, dealing with the
Question of the Unit of Heat.

1. — From Dr C. Dieterid, Professor of Physics, Hanover.

[This reply baa, since it was sent to Mr Griffiths, been printed in full
in Wiedemann's Annalen for February, 1896. It is therefore not thought
neoesaary to print it again here.]

2,— From Dr Dom, Professor of Physics, Halle,

December 27, 1895.

[Translation.]

...I quite agree with you that it is very necessary there should
be an improvement in the department of calorimetry, and that
the first step must be the determination of sharply defined units.
I agree with you in the opinion that the new unit ought not to
differ in a marked degree from the present, for it would otherwise
cause great inconvenience to both physicists and chemists, and
there would be no hope of introducing the new unit technically.

I have really no objection to offer to the thermal unit being
42 X W ergs (or i-ather 41 89 x 10« ergs).

3. — From Dr TT. Ostwald, Professor of Chemistry, Leipzig,

February 12, 1896.

[Translation.]

I entirely agree with your proposal to take some multiple of
the erg as unit of heat. Such a step seems to me so undoubtedly
necessary that, in my opinion, the question is when and not if
such a change should be carried out. I therefore regard your
proposition as a welcome opportunity for going into the neglected
question, and 1 may say that I am determined to recalculate, iu
the forthcoming third edition of my text-book, the whole of the
thermo-chemical data in such a manner as to do my utmost to
diminish the difficulties consequent on the transition. I have
already (in 1891) expressed my opinion very clearly, and I now
send you the memoir referring to it*.

* See Studien tur Energetik, p. 677.

FOR ELECTRICAL MEASUREMENTS 645

I differ from your proposals, however, as regards the magnitude
of the unit to be adopted. I believe that only an erg multiplied
by some integral power of 10 should be chosen. I formerly pro-
posed a Mega-erg, but have now altered my opinion.

As a practical multiple of the erg, we already possess one in
electricity, viz., the Joule = 10* ergs ; and it appears to me to
have the great advantage that the practical unit of energy in
constant use in the two great departments of electrical and thermal
measurements would be identical; therefore I do not think that
any other choice could be so advantageous.

4. — From Dr F. Paschen, Tit Professor of Physics, Hanover,

November 24, 1895.

...We must have an absolute unit simply related to other
absolute units, and that would be your " Rowland " ; but we must
also know how to realise this unit. For this purpose the specific
heat of water must be fixed for each temperature.

I think, as the different observations on the variability of the
specific heat of water differ so greatly your statement III. (p. 541)
is a very preliminary one.... I think it would be best to propose
that a new determination of the changes in the specific heat of
water should be undertaken by some institute that has the neces-
sary apparatus and money.

5. — From Dr M, Planck, Professor of Physics, Berlin,

November 25, 1895.

[Translation.]

If I may venture on giving my opinion on the propositions
made by you, I must emphasise, before all things, that I agree
with you as to the necessity of having a well-defined universal
unit of heat, and I should be very glad if your well-considered
plans led to a definite result. As a theorist I would make even
more radical demands as to the unit to be defined. The ideal
universal unit of heat appears to me to be still more closely
related to the definition of the electrical units; consequently
I would define : —

I. One " Rowland " (or " Meyer," or " Kelvin ") £U9 that quan-
tity of heat which is equivalent to 10" ergs.

II. According to the best measurements hitherto obtained
1 " Rowland " is that quantity of heat which raises 1 gramme of

B. A. 35

546 PRACTICAL STANDARDS

water at IS"* C. through 2*39'* C. It would be possible to modify
this number in the light of subsequent experiments. We should
thus avoid the arbitrary character involved in the choice of such
numbers as 41*89 x 10" or 42 x 10*. At the same time I quite
acknowledge that the establishment of this unit will cause a
considerable revolution in present thermal calculations which
will be difficult to carry out, and it will therefore probably meet
with energetic opposition from practical physicists and from
technical men. Still, as I have already remarked, I should con-
sider it a great step in advance if even the value of the equivalent
of heat were established.

6. — Friym Dr WnXlner, Professor of Physics, Aachen,

February 23, 1896.

[Translation.]

I, also, have finally decided on determining the unit of heat
by the work done, inasmuch as I have endeavoured to determine
the work which is equivalent to the mean calorie measured by
the ice calorimeter.

I hope I made it evident that I am quite aware of the un-
certainty of this method of calibration. I thus arrived at the
value 4175*8 x 10*, or, in whole numbers, 4176 x 10*, which,
according to Rowland, corresponds to the heat required to raise
the unit weight of water through l^'C. at 22''C. of the air
thermometer.

I am, however, quite willing, if an agreement can be arrived
at, to discard the always uncertain relation to the mean unit of
heat, and to accept your proposed unit 42 x 10". The temperature
IS"*, at which the specific heat of water is then unity, is more
convenient The consequence of such an agreement will be that
all thermal measurements in which absolute values are aimed at
will be made with the water calorimeter, in which case it appears
easier to experiment with temperatures about 15' ; also we are in
better agreement as to the behaviour of water between 10"" C.
and 20"* C, although, even then, there is not complete certainty.
I should, for example, prefer to make the reductions at 15° entirely
according to the observations of Rowland, as he has directly
measured the equivalent of heat at these temperatures. Finally,
as regards the designation of the new unit, I do not approve of

FOR ELBCTRICAL MEASUBBMENTS 547

giving it the name of a physicist; also the name ''therm" is
suitable for English physicists, but not for others.

Why should we not simply preserve the name " thermal unit " ?
Or, if a distinctive name is used, then, approximating to the long-
used " calorie," call the new unit a " calor." The definition would
then be, "A calor is the heat value of 41*89 x 10" ergs," and, until
further notice, the calor will be equal to the amount of heat
which will raise the unit mass of water at IS"" through 1^ C.

No especial name has been given to the length of the mercury
column which is equivalent to 1 ohm. In no case would I advocate
the adoption of a second definition for the practical unit (besides
" Rowland," " calor," or simply " thermal unit "), as that would
lead to confusion.

7. — From Dr Boltzmann, Professor of Theoretical Physics^ Vienna,

November 26, 1895.

The unit ought to be as simple as possible and capable of
accurate determination, as all other qualities are of less import-
ance. It would be simplest to choose the heat which raises the
temperature from 10^ to 11^ C

In general I am in accord with all you say in your paper.
The most important thing is that the same conception should be
adopted everywhere, and for this reason I will fully accept the
decision of the majority of the Committee.

8. — From Dr K. Olszewski, Professor of Chemistry, Cracow,

December 14, 1895.

I have taken the advice of my colleagues in the Cracow
University, Professors Witkowski and Natanson, and I beg to
submit to your attention, as well as to that of the British Associa-
tion Electrical Standards Committee, the following suggestions,
being the conclusions arrived at conjointly by the above-named
gentlemen and mysel£

1. It would be advisable, on theoretical grounds, to select a
Joule, or lO' ergs, as the fundamental theoretical or ideal unit of
heat-energy. Hence the following proposal is brought forward : —

"That the theoretical or thermO'dynamical, or, say, a 0.8.
standard thermal unit, be defined as the heat equivalent of
a Joule or of IW ergs, and termed a thermal Joule"

35—2

548 PRACTICAL STANDARDS

2. That, as a practical thermal unit, the quantity of heat
required to raise 1 gramme of pure water through V of the thermo-
dynamical scale at 15^ of that scale be temporarily adopted.

3. That, in view of the exceptional importance of the question,
steps be taken, by international co-operation or otherwise, leading
to the determination of the numerical value of the ratio between
the theoretical unit and the practical unit, defined by IS"", as above
stated, by some at least of the leading physical and metrological
laboratories and institutions of the world, with the highest degree
of accuracy nowadays attainable ; and to the extension (if possible)
of such determinations over as great a range of temperature as
practicable. Added to the highly valuable work already done,
such an investigation cannot fail to settle the question of the
specific heat of water ; and if this be done, the subject of thermal
units will have lost nearly all of its present difiiculty.

9. — From Dr Chappuis, Bureau International dee Poids et
Mesures, Shres^ February 2, 1896.

[Translation.]

...Your arguments have led me to accept the propositions given
by you on p. 541.

If, however, I may be allowed to express a wish, it is that the
values may be reduced to the normal scale of temperature, i.6., to
that of the hydrogen thermometer, and not to the air or nitrogen.

It is true that the difference between these scales is veiy small,
but still it is perfectly measurable Some experiments of the Bureau
International des Poids et Mesures (not yet published) have led me
to the conclusion that the thermometric scale of hydrogen is inde-
pendent of the initial pressure between 0*5 and 2 atmospheres, and
that the hydrogen thermometer at constant pressure gives sensibly
the same values as that thermometer at constant volume. It is
not so with the nitrogen or the air thermometer.

The difference between the nitrogen and hydrogen scales is
indicated both in the original memoir (Trav. et M^m. du Bureau
International, Vol. Vl.) in the pamphlet on thermometry of pre-
cision by M. Guillaume, as well as in Landolt and Bomstein's
physical tables, 2nd edition, p. 93. Also a great number of
physicists have adopted the decision of the International Com-
mittee of Weights and Measures to take, as the normal scale of
temperature, that of the hydrogen thermometer at constant volume.

FOB ELECTRICAL MEASUREHENTS 549

10. — From Professor Le Chatelier, School of Mines, Paris.

[Translation.]

. . .1 should like the theimal unit to be a number of ergs chosen
arbitrarily ; either !()• ergs, or, in order to approach more nearly
to the present unit, 5 x 10' ergs. Then, as practical unit, I
should like two : (1) A unit, of precision analogous to the ohm,
which should be the quantity of heat yielded by a given mass of
mercury in passing from one state to another, the states being
defined by volume or electrical conductivity. (2) The present
unit should be the specific heat of water at IS*'.

The use of water is indispensable for current researches, but
it appears to me very doubtful for researches of precision.

It is supposed that the condition of water and, consequently,
its internal energy are completely determined when the pressure
and temperature are ascertained. Now, nothing is less probable.
Since Ramsay's researches, we know decisively that water is
formed of a mixture of molecules at various degrees of associa-
tion ; it is a sjrstem in equilibrium. The state of equilibrium of
analogous systems is in theory entirely defined when. the pressure
and temperature are known. But in practice the state of equi-
librium is only attained with an extreme slowness, and sometimes
it is never reached. The lower the temperature, the more serious
are those delays in reaching the state of equilibrium. It is
therefore possible that the specific heat of water varies with the
temperature, and that it differs according to whether the initial
temperature of the experiment has been reached when ascending
or descending.

11. — From Dr Guillaume, Bureau International des Poids
et Mesures, Sivres, November 19, 1895.

[Translation.]

I believe that if the French Committee adopt your proposal
as to the fixing of the new unit, they will declare themselves still
more decidedly in favour of the name which you have given them,
as it has already been proposed here to name " therm " the equi-
valent of heat of the erg or of one of its decimal multiples.

I do not think, in return, that we could agree with you as to
the scale of the nitrogen thermometer. There appears to be no

550 PRACTICAL STANDARDS

doubt that the hydrogen thermometer gives a scale extremely
like the thermo-dynamic, and that it is, at all events, the most
analogous we can have. Sooner or later it will be necessary to
adopt the thermo-dynamic scale, and it is well to now approach
to it as nearly as possible.

Besides, this scale is one of a certain small number of units
on which a legal authority has been conferred. It is now included
in the decisions arrived at by the International Committee of
Weights and Measures, which a certain number of States have
introduced into their legislation.

In itself the thing is actually of little importance; but it
becomes more important in proportion as experiments become
more exact, and it is best to have as little as possible to change
in the end.

12. — Froni Professor J. S. Ames, Johns Hopkins University , n.8.A.,

December 10, 1895.

...I must say your proposal appeals to me in every way. The
10° unit seems to me to be preferable to the 15** one.

13. — From Professor H. i. Gallendar, Professor of Physics^
McOill University, MontrecU, December 5, 1895.

I entirely agree that it would be a very great improvement
to adopt an absolute unit in place of the present various and
uncertain units based upon the peculiar properties of water. I
think, however, that it would be better to connect it more simply
and directly with the system of electrical units, and to use only
names which are already familiar to all engineers, than to attempt
to retain a close approximation to the value of any of the old
specific heat units, which are essentially arbitrary.

The following are the names of the series of thermal units
which I should be inclined to suggest as being already familiar
in practice: —

1. The thermal watt-second, or "Joule," defined as being
equivalent to 10' C.G.s. units of work. A rider might be added
to the effect that, according to the best determinations, this unit

is approximately equal to j;^ of the gramme degree centigrade
at 10" C.

FOR ELECTRICAL MEASUREMENTS 551

2. The thermal watt-hour, which would be equivalent to
3600 Joules, and would therefore be of a similar magnitude to
the kilogramme degree centigrade, which is so largely used in the
thermo-d}mamics of the steam-engine. The watt-hour, in fact,
would be exactly f ths of the kilogramme degree centigrade at
some temperature in the neighbourhood of lO"* C.

3. The thermal kilowatt-hour, or simply kilowatt-hour, which,
as the Board of Trade unit of electrical energy, is already so
£Etmiliar and useful for the commercial measurement of large
quantities of energy.

In connexion with the latter unit it may be remarked that
it would be a great advantage if engineers could be induced to
adopt the kilowatt as their unit of mechanical power in place of
the horse power. The latter unit differs from the " cheval-vapeur,"
and being based upon the foot-pound has different values in
different latitudes. For the order of accuracy generally attainable
in steam-engine work, it would, as a rule, be sufficient to take the
horse power as being }ths of the kilowatt power.

For steam-engine work undoubtedly one of the most important
units at present in use is the British thermal unit, or pound
degree Fahrenheit. It happens that the watt-hour is very nearly
equal to 3*400 B.T.U. The reduction of the latter to watt-hours
may be very readily effected by multiplying by 0*3 and then
reducing the result by 2 per cent.

It would seem, on the whole, not improbable that the simple
adoption of all the femiliar units of electrical energy, with the
prefix '* thermal," if necessary, as our absolute units of heat, would
result in a more general agreement and a greater simplification of
expression than any attempt to re-define one of the older units
in terms of the absolute S3rstem. The latter course might readily
lead to confusion, and would necessitate the retention of the
constant factor J'=s4*2xl0' in our equations whenever they
involved electrical or mechanical measurements.

To put the question in a brief and concrete form for the
consideration of the Committee, I think that the views above
expressed might be embodied in some such resolutions as the
following : —

1. That the thermal equivalents of the practical units of
electrical energy above mentioned may be taken as convenient
absolute units of heat.

552 PRACTICAL STANDARDS

2. That when used to denote quantities of heat these
units may be distinguished, if necessary, by prefixing the word
"thermal."

3. That the "thermal watt-second/' which is intended to
represent 10' C.GJ3. units of energy, be also called a " Joule."

4. That the heat developed by an electromotive force equal
to that of a standard Clark cell at 15° C, when acting through a
resistance equal to one standard ohm, may be taken as 1*4340
Joules per second.

5. That (pending the results of further investigations) the
quantity of heat required to raise the temperature of one gramme
of water through one degree of the centigrade air thermometer
in the neighbourhood of 10"" C. may be taken as 4*200 Joules.

6. That the thermal watt-hour, which is equal to 3*600 Joules,
may be taken as equal to f ths of the kilogramme degree centigrade
at 10"" C, or as equal to 3*4 times the pound degree Fahrenheit
at 50** F.

7. That for the reduction of observations to the standard
temperature of 10"* C. or 50° F., the temperature coefficient of the
diminution of the specific heat of water may be taken as "00036
per V C, or "00020 per V F., over the range 10° to 20°.

With regard to the last resolution I do not see that anything
would be gained in the present state of our knowledge by adopting
a more complicated or discontinuous formula of reduction, until
we are prepared to extend it to higher ranges of temperature.

The name "Joule," as that of the father of the mechanical
measurement of heat, would not, I think, be open to objection.
At the same time I feel that the choice of a special name for the
absolute unit of heat is one comparatively of secondary import-
ance. The really essential points to impress upon the world of
science in general, and upon engineers in particular, are, that
the specific heat of water is for from constant, and that 772 foot-
pounds are not very accurately equivalent to the B.T.U. Also
that in measuring quantities of heat by the rise in temperature
of a mass of water it is most important to have an accurately
verified thermometer, and to state the limits of temperature
between which the observations were taken. It would certainly
be a great advantage for the reduction and comparison of obser-
vations to use always the same standard formulae, such as those
which you suggest ; but it would still be. necessary in accurate

FOB ELECTRICAL MEASUREMENTS 558

work to state the limits of temperature for subsequent identifica-
tion, should these formulae prove on more exact investigation to
be not sufficiently approximate.

14. — From Professor E. L, Nichols, Professor of Physics, Cornell
University, Ithaca, U.S.A., January 12, 1896.

The suggestion of defining the heat units by means of the
melting of ice strikes me so favourably that, in spite of the
difficulties which have hitherto been found in determining the
precise heat of fusion, I am considering the question of the re-
determination by new methods with a view of finding whether
one can obtain a sufficient degree of accuracy to warrant the
adoption of the heat of fusion of water as the basis for thermal
measurement.

15. — From Professor Rowland, Professor of Physics, Johns
Hopkins University, U.S.A., December 15, 1895.

As to the standard for heat measurement, it is to be considered
fix>m both a theoretical as well as a practical standpoint.

The ideal theoretical unit would be that quantity of heat
necessary to melt one gramme of ice. This is independent of
any system of thermometry, and presents to our minds the idea
of quantity of heat independent of temperature.

Thus the system of 'thermometry would have no connexion
whatever with the heat unit, and the first law of thermodynamics
would stand, as it should, entirely independent of the second.

The idea of a quantity of heat at a high temperature being
very different fix>m the same quantity at a low-temperature would
then be easy and simple. Likewise we could treat thermo-
dynamics without any reference to temperature until we came
to the second law, which would then introduce temperature and
the way of measuring it.

From a practical standpoint, however, the unit depending on
the specific heat of water is at present certainly the most con-
venient. It has been the one mostly used, and its value is well
known in terms of energy. Furthermore, the establishment of
institutions where it is said thermometers can be compared with
a standard renders the unit very available in practice. In other
words, this unit is a better practical one at present. I am very
sorry this is so, because it is a very poor theoretical one indeed.

554

PRACTICAL STANDARDS

But as we can write our text-books as we please, I suppose
that it is best to accept the most practical unit. This I conceive
to be the heat required to raise a gramme of water 1* C. on the
hydrogen thermometer at 20^*0.

I take 20® because in ordinary thermometry the room is usually
about this temperature, and no reduction will be necessary. How-
ever, 15"* would not be inconvenient, or 10^ to 20^

As I write these words I have a feeling that I may be wrong.
Why should we continue to teach in our text-books that heat has
anything to do with temperature ? It is decidedly wrong, and if
I ever write a text-book I shall probably use the ice unit. But
if I ever write a scientific paper of an experimental nature I shall
probably use the other unit.

APPENDIX II.
The Capacity for Heat of Water from lO"* to 20*" C.

REFERRED TO ITS CAPACITY AT 10*" C. AS UnITY.

—

Rowland

Griffiths

Bartoli and
Stracciati

Mean

10°

lOOOO

I'OOOO

1-0000

11°

-9995

'9997

•9997

•9996

12°

•9990

*9994

•9994

-9993

13°

•9985

'9991

•9991

•9989

14°

•9980

•9989

•9988

•9986

15°

■9974

•9986

•9985

•9982

16°

•9969

•9983

•9981

•9978

17°

•9964

•9981

•9979

•9975

18°

"9959

•9978

•9972

19°

•9954

•9975

•9977

•9969

20°

•9950

•9973

■9977

•9967

(Numbers given in italics are obtained by extrapolation.)

Note. — If we assume the validity of the numbers in the last
column, then any quantity of heat (Qt) expressed in terms of the
capacity for heat of water at f C. may be expressed with sufficient
accuracy in terms of the thermal unit at lO"" C. (Qio) by means of
the following formula : —

(2i« = Qe {1- 00033 («- 10)},

where t lies between 10*" and 20° C.

Then Q^ x 4'2 gives the equivalent in Joules,

FOR ELECTRICAL MEASUREMENTS 555

APPENDIX III.

Recalculation of the Total Heat of Water from the
Experiments of Regnault and Rowland. Bt W. N. Shaw.

Tables of Thermal Data eapressed in terms of Joules.

The thermal data depending upon a thermal unit, which are,
as a rale, included in tables of physical constants, comprise the
following : —

The variation of the specific heat of water with variation of
temperature.

Specific heats of various substances, solid, liquid, or gaseous.

Latent heats of fusion.

Latent heats of evaporation.

Heat of chemical action.

Thermal conductivities of various substances.

The tables are mainly compiled by grouping the results ob-
tained by a number of observers. Such results are only, strictly
speaking, comparable where the scales of temperature, and the
thermal units adopted for the reduction of the observations, are
identical. With difierent observers this is only the case if very
rough approximation be allowed; but the experimental data
communicated in the description of observations sometimes afford
the possibility of putting the results upon a better footing for
comparison than that upon which the author's own reductions
leave them. It is clear that the auxiliary data which must be
used in order to render the results strictly comparable, are in
effect precisely those which are necessary to express the author's
data in absolute measure, except that for the mere purposes of
comparison one datum — the djmamical equivalent at one specified
temperature — is not actually required. At the same time the
comparison of data is in no way vitiated by the use of some
number (for the present a conventional one), in order to convert
a result bom some definite gramme-degree-unit to Joules.

An examination of the tables of thermal data with a view
to expressing the results in Joules furnishes, therefore, a very
effective test of the comparability of the results obtained by
different observers for the same thermal constants, and, moreover,
the difficulties to be met with in making the reduction to Joules

556

PRACTICAL STANDARDS

give the best indication of the points which must be settled before
the results of thermal measurement can be regarded as final. To
cany out such an examination completely, using numbers for
reduction that can only be regarded as provisional, would be an
unnecessary labour; but a few selected instances may help to
exhibit some of the uncertainties which might reasonably be
expected to disappear if observers once recognised the desirability
of expressing all thermal measurements in Joules, or in some
recognised equivalent.

Table I. — Total heat of water at variotia temperatures of the
absolute scale (hydrogen thermometer) between W and 36"",
expressed in Joules {Rowland's experiments).

Total Heat in

Joules between

Joules betweeu

0°and r°

(f and r°

S*'

21'044*

21'

88-144

25-254

22'

92-321

29-462

23'

96-496

33-668

24'

100-671

37-871

25'

104-844

10'

42-072

26'

109-017

46-271

27'

113188

12'

60-468

28'

117-369 1

13'

54-663

29'

121-530

14'

58-856

30'

125-700

15'

63-046

31'

129-871 '

16'

67-234 1

32'

134-042

17'

71-420

33'

138-214

18'

76-604

34'

142 386 !

19'

79-786

35'

146-558

20'

83-966

36'

150-731

As an example, I have computed the total heat of water at
various temperatures as determined experimentally. I have used
Rowland's numbers for lower temperatures, and have recomputed
Regnault's experiments, accepting Table I. (computed fix>m Row-
land) as correct.

I think it might be possible to find data enough to recompute
some others, e.g„ the latent heat of steam at 100'', the specific
heat of air at constemt pressure, which, by the way, is almost
exactly a Joule. The labour is, however, very considerable, and

* The total heat between 0° and 4J° is obtained by extrapolatton from
Rowland's numbers.

FOR ELECTRICAL MEASUREMENTS

657

it might be abbreviated (for the Committee) if those who are
or have recently been engaged in thermal measurements would
supply the Committee with the results of their own observations
reduced to Joules and thermometric units.

The numbers are reduced from the table in the Mimoires de
VInstitut, tome xxL p. 743, by assuming the mean specific heat
of water for the calorimetric range of each experiment to be the
specific heat of water as given in Rowland's table for the meaii
calorimetric temperature of the experiment, and adding to the
heat thus computed, as that given out by one gramme of water
in cooling from T to the final calorimetric temperature, the
further amount which, upon an estimation based on Rowland's
data, would be given out on cooling to 0°.

+ «.

+ 40
+ 20

■

—

a
II

T"^

« m

■o

0
-2-0

190

Absoissfte. — Air TemperatoreB (T),

Ordihates.— Differenoes (in Joules) between Total heat from 0° to r° and 4-2 x T.

Some doubt has been thrown on the accuracy of the data
quoted by Regnault in the table referred to. I have adopted
Mr Macfarlane Gray's conclusion that the computations of the
mean specific heat are correct, though the data are erroneously
printed in Regnault's paper.

The results of the individual experiments are shown in the
following table. In order to obtain a mean result a curve of
differences (see figure) between total heat at temperature T and
4'2 X T has been plotted, and the means of observations, collected
into seven groups, have been taken and also plotted. These are
indicated in the diagram by circular dots, the individual results
being shown by crosses.

568

PRACTICAL STANDARDS

Table II. — Regnaulfa Observations for the total heat of water
between 0** C. and various temperatures (T) of the ''Air-
thermometer ** above the boiling-point of water.

(Reduced from Regnault's and Rowland's resulta Expressed in Joules.)

Total Heat
ftom

4*2 X T

Differ-

Total Heat
from

4-2 xT

Differ-

QPioT"

enoe

QPioT"

enoe

107W

1 451-83

452*34

- -51

153-68°

646-44

645*46

+ -98

107-90°

453-60

45318

+ -42

154-80°

651*97

650*16

+ 1-81

107-7r

453-36

452-72

+ -64*

155-61°

654*27

653-56

+ 71

109-38*'

460-69

459-40

+ 1-29

156*82°

660-89

658*64

-2-25

109-25"

461-44

458-85

+2-59

158-82°

668-38

667-04

+ 1-34

109-25'

460-94

458-85

+ 2W

159*19°

669*64

668-60

+ 1'04*

109-25°

460-84

458-85

+ 1-99

160-34°

675*74

673*43

+2*31

110-80°

46576

465*36

+ *40

160*61°

677*61

674*56

+3*06»

111-61°

467-60

468-34

- 74

113-86°

478-56

478-21 + -35

116-60°

491-35

48972

+ 1-63*

172-66°

728*47

725-17

+3-30

116-91°

492-46

491-02

+ 1*44*

172-75°

730*82

725-55

+5-27

118-54°

498-76

497*87

+ -99*

17271°

730-68

725*38

+5*30

120-39°

504-86

505-64

- -78

172*66°

730-59

725*17

+5*42

120-84°

507-36

507-53

- -17

121-86°

512-72

511*81

+ -91

III

•

128-91°

54230

541-42

+ -88

179-23°

759-70

75277

+6-98*

130-40°

548-07

547-69

+ -38

183*56°

776*57

770*95

+5*62*

186-00°

787*34

781-20

+614*

IV.

186-65°

791*62

783-93

+7-69*

137-16°

577-27

576-07

+ 1-20

186-89°

790*86

784*95

+5-91*

137-27°

577-96

576-53

+ 1-43

187-75°

795*35

788-55

+6-80*

138-27°

581-58

580-73

+ -85

190-36°

805*80

799-51

+6«

A curve based upon these means as accurate would show a
minimum ordinate above 100® C. Without any definite experi-
mental reason I have considered this as outside the range of
probability, and have drawn a curve corresponding to a gradual
increase of specific heat between the limits of the experiments,

* It will be remembered that Begnault gives for these two values 1*0050 and
1*0133 respeotivelj.

FOR ELECTRICAL MEASUREMENTS

559

viz., 107^ and 190^, which (airly connects the means. Continuing
that curve beyond 107° to 100° and reading off from it the value
of the total heat — 4*2 x T at intervals of 10° we get the following
result : —

Table III. — Total heat of water between 0° and 7° {air-
thermometer) according to Regnault and Rowland,

Total Heat be-

tweenO°andT°

iD Joules

1
Excess over
4-2xr

100'

4d0-e8

■68

no*

462-73

-73

120''

504-80

•80

130'

546-88

•88

140''

589-04

1-04

150'

631-40

1-40

ISO'

674-02

2-02

170'

717-62

3-52

180'

761-60

6^60

Whence we obtain —

Mean specific heat of water between 0° and 100° is 4*2068
Joules = 1*0016* thermometric units.

Mean specific heat of water between 0° and 180° is 4*2312
Joules s 1*0075* thermometric units.

* It will be lemembered that Regnault gives for these two Tallies 1-OOSO and
1*0188 tespeotirely.

TWENTY-FOURTH REPORT— TORONTO, 1897.

APPENDIX PAGE

I. Note on the Constant-volume Gas-thermometer, By G. Cabby

FoBTBR, F.R.S 564

II. (hi a Determination of the Ohm made in Testing the Lorenz
Apparatus of the McOiU University^ Montreal. By
Professor W. E. Ayrtox, F.R.S., aiid Professor J.
VlRIAMU JONBS, F.R.S 567

At the Liverpool Meeting the Committee agreed that the
"calorie," defined as the heat equivalent of 4*2 x 10' ergs, should be
adopted as the unit for the measurement of quantities of heat, but
the question as to the exact part of the absolute thermodynamic
scale of temperature at which this quantity of heat could be taken
as equal to one water-gramme-degree was for the time being left
open.

This resolution has made it incumbent on the Committee to
consider carefully —

1. The relation between the results of measurements of
intervals of temperature by accepted methods and the absolute
scale;

2. The specific heat of water in terms of the erg and its
variation with temperature.

With regard to the first point there appears to be no reason to
doubt that the scale of a constant-volume hydrogen-thermometer
is very nearly identical with the absolute scale*. The Committee
have therefore decided to recognise the standard hydrogen-ther-
mometer of the Bureau International des Poids et Mesures as
representing, nearly enough for present purposes, the absolute
scale. This convention has at least the advantage of giving a
definite meaning to statements of the numerical value of intervals
of temperature within any range for which comparison with the

* See Appendix No. I to this Report.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 561

hydrogen-thermometer is practicable. If fiiture investigation
should show that it is inaccurate to any appreciable extent, corre-
eponding corrections can be applied when necessary.

Experience of the use of the platinum resistance-thermometer
in various hands encourages the hope that it will afford a con-
venient and trustworthy working method of referring the indications
of mercury- or other thermometers to those of the standard hydro-
gen-thermometer. The Committee have consequently much
satisfaction in learning that Dr J. A. Harker, formerly of Owens
College, is at this moment canying out at Sevres, on behalf of the
Committee of the Kew Observatory, and with the concurrence of
the Director of the Laboratories of the Bureau International, a direct
comparison of platinum thermometers belonging to the Eew Obser-
vatory with the standard hydrogen-thermometer of the Bureau.

As to the djmamical value of the specific heat of water — in
other words the mechanical equivalent of heat — it was pointed out
by Professor Schuster and Mr Gannon in 1894* that the results of
the best determinations by direct mechanical methods agree
among themselves much more closely than they do with those
that are founded upon electrical measurements of the energy
expended, although these in turn are in good agreement among
themselves. Additional significance is given to this remark by
the comparison of those determinations which, by extending over
an appreciable range of temperature, indicate the rate of variation
of the specific heat of water. Of such determinations there is one
of each kind, that of Professor Rowland by the mechanical method,
and that of Mr E. H. Grifiiths by the electrical method The
results of the former of these have recently undergone an elaborate
revision at the hands of one of Professor Rowlands pupils,
Mr W. S. Dayt, who has compared the three principal ther-
mometers employed in the experiments ynth the Sfevres hydrogen-
standard by means of three Tonnelot thermometers which had
been compared at the Bureau with the hydrogen-standard.
Messrs C. W. Waidner and F. MalloryJ have also compared two of
Rowland's thermometers with a platinum thermometer made by
Mr Griffiths. The result of this discussion is to leave Rowland's

* Phil, Tram. vol. clxxxyi. p. 462 ; Proc, Roy, Soc, vol. Lvn. p. 81.
t Joknt Hopkim Univenity Cireulan, pp. 44, 45 (Jane 1S97) ; also PhiL Mag.
zuv. 169-173.

t nnd, pp. 42, 48 (Jane 1897) ; Phil. Map, xuv. 165-169.

B. A. 36

562

PRACTICAL STANDARDS

original value unchanged at 15^, and to raise it by four parts in
4000 at 25°, making the rate of variation of the specific heat of
water almost exactly the same as that given by Griffiths's experi-
ments throughout the same range.

The following table gives the numerical values : —

Values of the Specific Heat of Water at 15° C.
1. By mechanical friction : —

Author

Date

Resalt

Joule

Miculescu

Rowland

1878
1892
1879

4172x10* ergs
4181 „
4189

Reynolds and Moorby...

1897

>< 1 QQ w 1 A4 S mean specific heat
^^®^^^^|fromOUolOO»C.

2. By electrical methods : —

Author

Date

Result

Griffiths

Schuster and Qannon...

1893
1894

4199-7 X 10« Gtg^
4197

VaricUion of the Specific Heat of Water,

Temperature

Specific Heat

Rowland

Griffiths

6
10
16
20
25
30
35

4204x10*
4197 „
4189 „
4183 „
4177 „

4173 „

4174 „

4199-7 X 10
4193-2 „
4187-4 „

Joule's (1878) result is given by Schuster and Gannon (Froc Roy. Soe,
Lvii. p. 31) as 775 foot-pounds at Greenwich per degree Fahr. As Professor
Schuster has examined the thermometers employed by Joule, this value is

FOR ELECTRICAL MEASUREMENTS 563

adopted as the most trustworthy statement of the result of Joule's experi-
ments: it is reduced to ergs and the Centigrade scale.

Miculeecu {Ann, Chim. Phys, [6], xxvn. 237) states his result as
426*84 kilogramme-metres per kilogramme-degree of the normal hydrogen-
thermometer between 10° and IS"*. Taking ^=:d80-96, this is equivalent to
4187x10 ergs per gramme-d^p^ee. The mean temperature 11***5 has been
adopted and reduction to W has been made by means of the rate of
variation given by Rowland's experiments.

Reynolds's and Moorby's experiments {Proc, Roy, Soc, LXi.) refer to the
whole range from 0"* to 100°. Their result is stated, in foot-pounds at
Manchester and degrees Fahr., as 776*94. To reduce to ergs and the
Centigrade scale this number has been multiplied by 1*8 x 90-48x981*34.

Schuster and Gannon {Proc. Boy. Soc. Lvn. 25-31).

Rowland's and Qriffiths's results are quoted from Day {PhU. Mag.y
August 1897, p. 171), whose statement is adopted by Griffiths in Nature
for July 15, 1897.

The agreement between the separate determinations by the
mechanical and by the electrical methods respectively, and the
regularity of the differences between Rowland's values and those
of Griffiths, is such as to raise a strong presumption that, in the
experiments by both methods, errors of observation have been
reduced to a very small amount. At the same time the difference
between the two sets of results points to some constant source of
energy affecting one or both. The mechanical method is, in principle,
so direct and simple that it is difficult to suppose its results affected
by a constant error. On the other hand, the electrical method
being less direct and more complicated, there is here more room
for uncertainty in the data.

The electrical determinations depend upon the well-known
relation between thermal and electrical energy, which is expressible
in the three forms —

Schuster and Qannon's experiments are based upon the second
form of the equation, those of Griffiths on the third. In both of
them electromotive force was measured by comparison with a
Latimer Clark's cell. Schuster and Gannon measured, in addition,
the strength of their current by means of a silver- voltameter, and
Griffiths measured a resistance in terms of the ohm.

The accepted value of the electromotive force of the Clark's
cell depends in its turn on the electrochemical equivalent of silver
as determined by Lord Rayleigh and Professor F. Eohlrausch, and

36—2

564 PRACTICAL STANDARDS

consequently it appears that the electrical determinations of the
mechanical equivalent involve a double reference to the electro-
chemical equivalent of silver, so that any inaccuracy in the adopted
value of this quantity would involve a duplicate error in the value
of the mechanical equivalent deduced therefrom.

In this connexion it may be mentioned that, in a recent letter
to Naturey vol. LVi. p. 292, Lord Rayleigh has stated that he does
not consider that a possible error of one part in 1000 is excluded
from his determination of the electrochemical equivalent of silver.
If it be assumed that his value is one part in 1000 too small, this
would almost exactly account for the difference between the
electrical determinations into which this quantity enters as a
factor and the direct mechanical determinations.

It thus appears to be a matter of urgent importance that a
redetermination of the electrochemical equivalent of silver should
be made, and that the general question of the absolute measure-
ment of electric currents should be investigated. In order to
enable them to carry out this investigation, the Committee have
decided to ask for reappointment and to apply for a grant of
£100 towards the expense of the necessary apparatus and experi-
ments.

Appendix I.

Note on the Constarvt-volume Oas-thermoineter.
By G. Carey Foster, F.R.S.

The absolute thermodynamic scale of temperature introduced

by Lord Kelvin is connected with the properties of real fluids by

the equation*

dT dv

Y~^ <l>

where dv is the infinitesimal increment which unit mass of a fluid
occupying the volume v undergoes when it is heated, under
constant pressure, from the absolute temperature T to the
infinitesimally higher absolute temperature T+dT, and Sw is the

* Compare equation (16) of Lord Kelvin's article '* Heat '* in the Encyelopadia
Bfitannicaf vol. zi. p. 571 ; Mathematical and Phyiieal Paperi, vol. m.

FOR ELECTRICAL MEASUREMENTS 566

amount of work required to restore the original temperature of
unit mass of the fluid when it has undergone a fall of pressure, Sp,
by passing through a porous plug, as in Joule and Thomson's
experiments, without loss or gain of heat.

It follows that, if there is any fluid which does not undergo a
change of temperature when forced through a porous plug, an
infinitesimal change of temperature is to the total temperature on
the absolute scale as the resulting change of volume of this fluid
is to the total volume. Such a fluid would be called a perfect gas.

The following discussion of the bearing of the results of the
porous-plug experiments on the indications of a constant-volume
gas-thermometer is taken from a copy which the writer made
in January 1894 of a fuller discussion of these experiments
communicated to him by his friend and former pupil, Mr John
Rose-Innes. Mr Rose-Innes will shortly read a paper on this
question before the Physical Society of London. In the meantime
the writer has his permission to make the present use of his
hitherto unpublished results.

It will be remembered that Joule and Lord Kelvin found that
all the gases they experimented on were, with the exception of
hydrogen, slightly cooled by being forced through the plug. With
hydrogen the effect was smaller than with other gases and was a
rise of temperature. At a given temperature the cooling effect
was, up to five or six atmospheres, proportional to the difference of
pressure on opposite sides of the plug. For a given change of
pressure the effect decreased with rise of temperature, and Joule
and Lord Kelvin concluded that it was approximately proportional
to the inverse square of temperature reckoned fix>m — 273°C.
With hydrogen the variation with temperature was too small for
them to consider it as clearly established ; if anything the effect
became greater as the temperature rose.

Mr Boee-Innes's discussion of these results is founded upon
his remark that an empirical formula with two constants, a and fi,
namely

where 0 is the cooling effect and t temperature on the ordinary
centigrade scale, represents the experimental values rather more
accurately than the inverse-square fonnula. The values of a and
fi calculated by him for air, carbonic acid gas, and hydrogen, the

566 PRACTICAL STANDARDS

change of pressure being represented by 100 inches of mercury,
are as follows : —

Air ••• •••

• ••

• • •

a
441-5

/9
-0-697

Carbonic Acid

• ■•

•••

261-6

-4-98

Hydrogen

• «•

•••

64-1

- -331

To apply equation (1) to the discussion of the gas-thermometer,
we may begin (like Joule and Kelvin) by expressing the work 8w,
required to restore the gas to its initial condition, in terms of the
observed cooling eflTect, and may write

Sw = JC0^Jc{^ + ^y

where J is the mechanical equivalent of heat and C the specific
heat of the gas under constant pressure. If we remember that J
may be written J— W/viff, where W is the work that must be
spent to raise the temperature of a mass m of water by the amount
^, we see that the thermometric scale on which 0 and O' are
expressed is of no consequence, provided it is the same for both.

Putting n for the change of pressure producing a cooling effect
0, we may write equation (1) thus, taking reciprocals of both sides :

4;— f&-^) (^)

or, dividing throughout by jP* and integrating between limits T
and infinity —

\t)» t nV22"^ry ^^^

With regard to the first tenn on the right, it may be remarked
that all gaees appear to approximate more and more nearly as

temperature rises to agreement with the equation ^ « ^ (a con-
stant). Applying this to (3), we get

■R V _JC ( a

T)'

RT JC pf a „\

Neglecting, provisionally, the Joule-Kelvin effect, we have, as

a first approximation,

FOR ELECTRICAL MEASUREMENTS 567

and we may take thie value as accurate enough for use in the
small term containing p on the right-hand side of (4).
We thus get, as a second approximation to the value of

P"7[^-S<*«+^^)] <«)

Now let V remain constant, and let p^, T^ and pi, Ti represent
pressure and temperature at the melting-point of ice and at the
boiling-point of water respectively ; we then get

By subtraction

or, finally, if we assume 100 as the numerical value of the interval

whence we may conclude that, to the degree of approximation
attained in this calculation, the scale of the constant volume gas-
thermometer is identical with the absolute thermodynamic scale.

Appendix II.

On a Determination of the Ohm made in Testing the Lorenz
Apparatus of the McGHll University, Montreal, by Professor
W. E. Ayrton, F.RS., and Professor J. Viriamu Jones, F.RS.

This apparatus, made by Messrs Nalder Brothers, is in general
arrangement and dimensions similar to the Cardiff apparatus
described in the Philosophical Transactions of the Royal Society,
1891, A, pp. 1-42, and in the Electrician, June 1896, vol xxxv.
pp. 231 and 253.

568

PRACTICAL STANDARDS

The field coil, in pursuaDce of a suggestion contained in the
Royal Society paper, consists of a single layer of wire wound in a
helical groove of semicircular section, cut in the cylindrical sur&ce
of a massive marble ring of about 21 inches outside diameter,
15 inches inside diameter, and 6 inches thick. This helical groove
has 201 complete turns with a pitch of 0025 inch. Bare wire, of
mean thickness 0*02136 inch, was first used, and the outside
diameter of the coil so wound was measured in the Whitworth
machine with the following results : —

Diameter

1
Near front faoe

Near middle

Near back faoe

0" - 180"

21-04772

1
21-04765

21-04765

10' -190"

21-04795

21-04765

21-04952

20' -200'

21-04768

21-04755

21-04905

30' -210'

21-04805

21-04745

21-04818

40' -220'

21-04785

21-04755

21-04825

60' -230'

21-04808

21-04730

21-04812

60' - 240'

21-04752

21-04755

21-04805

70' - 260'

21-04755

21-04822

80' - 260'

21-04785

21-04795

21-04895

90' -270'

21-04812

21-04780

21-04942

100' -280'

21-04805

21-04815

21-04925

110' -290'

21-04808

21-04825

21-04898

120' -300'

21-04786

21-04840

21-04906

130' - 310'

21-04828

21-04835

21-04916

140' -320'

21-04828

21-04815

21-04908

150' -330'

21-04805

21-04806

21-04932

160' -340'

21-04872

21-04795

21-04858

170' -350'

21-04778

21-04785

21-04812

Mean 21-04797

21-04784

21-04872

General mean =21*04818 inches.

The temperature, which was taken at each observation, varied
between 19*^*9 C. and 2V C, and had a mean value of 20°-4 C.
Correcting for the difference between the temperature at which
the bars of the Whitworth machine have their specified value and
this mean temperature, we have for the mean outside diameter of
the coil, when wound with bare wire 0'02136 inch thick,

2104932 inch at 20"-4 C.

From the above measurements it is clear that the wire lay on
a very true circular cylinder. With bare wire, however, of the
thickness used it was found impossible to obtain sufficient insulation

FOR ELECTRICAL MEASUREMENTS 569

between pairs of convolutions. Hence, after much time had been
spent in endeavouring to insulate the successive turns by forcing
paraffin wax in between them, etc., the coil was unwound and re-
wound with double silk covered wire which had been first dried,
then drawn through paraffin wax, and lastly baked before the
winding was commenced. To wind so large and heavy a ring was
not an easy matter, and it was not until the winding had been
performed three times that the layer looked sufficiently uniform
and quite free from abrasion of the silk.

The mean thickness of the double silk covered wire used in the
last winding was 0*01914 inch, so that the outside diameter of the
wound coil, calculated from the value given above for the coil
wound with bare wire, was

2104488 inches at 20^4 C.

The coil was then brushed over with melted paraffin wax,
bound round with silk ribbon that had been soaked in a solution
of shellac, and finally loosely covered up with a wide silk ribbon
that had been passed through paraffin wax.

During the time that the ring was unwound the linear
coefficient of expansion of the marble was measured by Messrs
Spiers, Twyman, and Waters, three of the students of the City
and Guilds Central Technical College. The experiment was
attended with difficulty, for it was far firom easy to bring so large
a mass of a badly conducting substance to the same temperature,
but ultimately the result 0*000004 per V C. was obtained.

At the conclusion of the resistance observations recorded

further on, the silk ribbons and the protecting layer of paraffin

wax were carefully removed until the silk covering of the wire

appeared, and the diameter of the coil was measured along two

directions at right angles to one another. The maximum difference

between four measurements was only five hundred-thousandths of

an inch, and after the introduction of the proper temperature

corrections, the mean value of the outside diameter of the coil was

found to be

21-04687 inches at 20^*4 C.

This result is about one part in ten thousand larger than the
calculated value given above, and the difference is probably due to
the silk covering of the wire having swollen slightly when the
wound coil was brushed over with melted paraffin wax. In the

570 PRACTICAL STANDARDS

calculation, therefore, of the coefficient of mutual induction we
have considered it more accurate to use t]ie value obtained by
direct experiment Subtracting from that value — 21'04687 — the
thickness of the double silk-covered wire — 0*01914 — ^we have for
the mean diameter of the coil ^rom aada to aads of the wire

2102773 inches at 20"-4 G

Shortly before the last set of resistance measurements was
carried out, the edge of the phosphor bronze disc was ground in
position so as to be made quite true with the axis of rotation, and
immediately after the completion of the investigation the diameter
of the disc was measured and found to be 13*01435 inches at
19'''5 C. Messrs Spiers, Twyman, and Waters had previously
determined its linear coefficient of expansion to be 0*0000125 per
1"" C, so that its diameter was

1301451 inches at 20°-4 C.

During 1896 Mr W. G. Rhodes, when he was an Assistant at
the Central Technical College, carried out the long calculation of
the coefficient of mutual induction between the coil, as wound with
bare wire, and the disc by using the method given in the paper in
the Philosophical Transactions above referred to, and with the
following values: —

Diameter of coil or 2 A = 21 02673 inches.

Diameter of disc or 2a = 13*01997 inches.

Axial length of helix or 2a; = 5*025 inches.

Number of convolutions or w = 201.

He found Jf = 18056*36 inches

= 45862*33 centimetres.

The calculation was checked by Mr Mather and independently
by one of the authors.

Now it can be shown that for the above values of A, a, x,
and n

^= 1-246 ^ + 2-346^ + 0-0997^.
M A a a:

and so the value of M for the particular values of 2 A and 2a given
above, viz. 2102772 and 13-01451 can be calculated. When this is
done we find

Jlf= 18037-51 inches

= 45814*45 centimetres,

FOR ELECTRICAL MEASUREMENTS 571

and this was the value of M which we employed in our final
determination, after allowance had been made for the effect of the
central brush, as will be described further on.

The accuracy of the preceding calculations was tested in the
following way. Values of 2il and 2a, differing slightly from those
employed by Mr Rhodes, were selected, and by means of the

formula for -^ the proportional change in M was determined by

Mr Twyman. Then the value of M for these changed values of
2A and 2a was calculated by the authors from a new formula
involving an elliptic integral of the third kind*.

The centre brush consists of a tube, 0*135 inch outside diameter,
which projects into an axial hole in the disc of 0*144 inch diameter.
Contact with the edge of the disc is made by three small tan-
gential phosphor bronze tubes lightly pressed on it, at points
separated by angular distances of 120''. Through all four tubes a
small stream of mercury is kept flowing, as this is found to greatly
diminish the disturbances caused by variations in the thermo-
electric effects; and the employment of three brushes at the
circumference, as suggested by Rowland, eliminates small errors
due to imperfect centering of the coil and disc.

To prevent the mercury which drops out of the central tube-
brush touching the disc at a larger radius than that of the hole in
its centre an ebonite boss is cemented to the disc, and this causes
the mercury to drop away quite clear of the metal of the disc.

If we take as the effective outside diameter of the central tube
0'139 inch, that is the mean of 0'135 and 0*144 inch, calculation
shows that the coefHcient of mutual induction is reduced by
4*50 centimetres, so that finally we have

M= 46809*96 centimetres.

As the allowance for the central brush only diminishes M by
one part in ten thousand it is clear that, for that degree of accuracy,
an error of a few per cent, in estimating the diameter of the central
brush is of no consequence.

The method of making the observations was the same as that
described in the papers on the Cardiff apparatus read before
Section A of the British Association at Nottingham and Oxford

* An aoconnt of this new formola as weU as of that for -^ wiU shortly be

pablished by Professor Viriamn Jones.

672

PRACTICAL STANDARDS

{vide Report of the Committee on Electrical Standards, Appendices
1893 and 1894). The use of an extremely sensitive Ayrton-
Mather galvanometer of the d'Arsonval type materially facilitated
the readings being taken. Two such narrow coil galvanometers
were specially constructed by Mr Mather himself for use with the
Lorenz apparatus, and the data of the second instrument are con-
tained in the following table : —

Resistance of suspended coil ...
„ „ coil and suspension

Periodic time of complete swing

Scale distance actually used ...

1*9 ohms.

7-6 seconds.

{1412 millimetres.
1340 scale divisions.
Deflection in divisions at actual scale distance ri37 per micro-ampere.

used (23-8 „ micro-volt

Deflection in divisions at scale distance equal j204 „ micro-ampere.

to 2000 scale divisions (35-8 „ micro- volt

The resistance coils used were those previously employed in the
Cardiff determination of the ohm {vide Report of the Committee
on Electrical Standards, Appendices II. and III., 1894). They
have been tested once by Mr Glazebrook, and twice by the kindness
of Major Cardew in the Board of Trade Electric Standardising
Laboratory, with the following results : —

Coil

Mr Olazebrook,

Jan.— Maroh 1894

B
Board of Trade,
November 1896

1
C

Board of Trade,

August 1897

No. 3873
„ 3874
„ 4274
„ 4275

9-9919 at H'S" C.
9-9926 at 14-9" C.

•100050 at 15-2'* C.

'100053 at 15-2'' C.

9-992994 at 14-86*0.
9-993213 at 14-91° C.

•1000595 at 14-77" C.

•1000722 at 1514° C.

10-00712 at 19-3° C.
10-00775 at 19-3° C.

-100078 at 19-4° C.

•100081 at 19-4° C.

The coils Nos. 3873 and 3874 were stated by the makera,
Messrs Nalder, to be wound with platinum silver wire, and the
two others, Nos. 4274 and 4275, with manganin.

In the following table are given the temperature coefficients as
supplied originally by the makers, and as calculated from the tests
A and (7, and B and C

These figures show that a redetermination of the temperature
coefficients, which we are now carrying out, is necessary.

Fortunately the last set of determinations of the resistance of
these four coils was carried out at Westminster, within a fortnight

FOR ELECTRICAL MEASUREMENTS

573

of the completion of our absolute measurements, and we are much
indebted to Major Cardew for his kind promptness in the matter.
The temperatures of these 1897 Board of Trade measurements

Temperature Coefficients of Resistance per 1° C.

Con

As sapplied by
Biessrs Nalder

From teflts
A andC

From tests
BandC

No. 3873

., 3874
„ 4274
„ 4275

0-000276
0-000300
0-0000127
00000127

0-000360
0-00a}44
0-0000667
0-0000667

o-ooa3i8

0-000331

0-0000399

00000207

were so nearly those of the coils during our final absolute deter-
minations, which were from IS^^'S to 19'"4 C, as to render the effect
of possible errors in the temperature coefficients negligible to the
degree of accuracy aimed at by us. We have, therefore, used the
August 1897 Board of Trade values for these coils as transmitting
the Board of Trade ohm to the laboratory in Exhibition Road.

The standard thermometers used in the investigation were sent
to Kew and their errors were determined at the time by the
kindness of Dr Chree ; also, thanks to Sir J. Norman Lockyer, the
clock in the Mechanical Department of the Central Technical
College, which transmitted seconds to the fast running Bain
Chronograph, was frequently timed by reference to the current sent
hourly to his room from the General Post Office, and at 10 a.m.
from Greenwich.

The results of successive measurements of the absolute re-
sistances became very concordant after, little by little, various
possible causes of small errors had been eliminated. Nine sets
taken on July 30, 1897, gave the following results for the value of
the Board of Trade ohm in true ohms, without allowance for the
error in the clock rate : —

1000286

1-000277

1-000256

1-000306

1-000286

1-000284

1-000361

1-000307

1000295

Mean 1

-000294

674

PBACTICAL STANDARDS

or, since the clock was found to lose, during the daytime, at the rate
of three seconds per twenty-four hours, it follows that according to
this investigation

1 Board of Trade ohm = 100026 true ohms.
It is important to consider in which direction this result will
be affSpcted by sources of error that cannot be removed by careful
adjustment, centering, etc. They may be classified as follows: —

Soarce of Error

1. Over-estimation of the diameter of the coil

arising, for example, from the stress on
the copper wire having caused it to com-
press the under side of its silk covering.

2. Under-estimation of the diameter of the

phosphor bronze disc from a neglect of
the tips of the circumferential brush tubes
being; possibly pushed away from the disc
by the stream of mercury issuing, eta

3. Presence of iron pipes, girders, etc. in the

neighbourhood of the apparatus.

4. Traces of iron in the phosphor bronze disc.

5. Defective insulation between the support of

the central brush and the supports of the
circumferential brushes.

6. Defective insulation between the convolutions

on the coil.

7. Traces of iron in the marble ring.

8. Defective insulation of parts of the circuit

from one another.

9. Permanent magnetic field at the apparatus.

Effect Produoed

Result would be too
small.

Result would be too
smalL

Result would be veiy
slightly too smalL

Result would be too
large.

Result would be too
laige.

Result would be too
large.

Effect would depend
upon the position
of the leaks.

No effect, for the
current through
the field ooil was
periodically re-
versed.

As regards 4 and 7, special induction balances were constructed
and used by Mr Mather to test the permeability of both the marble
ring and the phosphor bronze disc; but, although a deviation fit>m
unity of one part in fifteen thousand could have been detected in
the permeability of either, no such deviation was observed.

As regards 5 and 8, careful tests were made every day of the
insulation resistance of the apparatus, and it was alwajrs found to
be greater than one thousand megohms.

6. The insulation between the adjacent convolutions of wire
could not be measured when they were silk covered and buried in
paraffin wax, since a small leak between a pair of turns would not

FOR ELECTRICAL MEASUREMENTS 575

change the apparent resistance of the copper coil by as much as
the variation in temperature of a fraction of one degree. We had,
therefore, to content ourselves with the precautions, previously
described, which were taken to secure high insulation in the
winding of the coil.

When the ring was wound with bare wire it was possible to
roughly compare the insulation resistance between pairs of convo-
lutions by sending a constant current through the coil and
measuring, very accurately, the p.d. between every adjacent pair
of the 201 turns. This we did several times, but it was a long
and laborious task.

When constructing a new Lorenz apparatus it will be well to
consider whether two separate helices should not be cut in the
cylindrical surface of the marble ring in which two independent
bare wires would be wound, a turn of the one being everywhere
(except at the extreme ends) between two turns of the other.
The insulation resistance, therefore, between the two windings
would measure the insulation between the adjacent turns, while in
the ordinary use of the apparatus the two windings would be joined
in series so as to constitute a single coil. In this way it may be
possible to be more sure of the absence of 6 than by using
paraffined double silk covered wire, and at the same time, to
entirely remove 1.

The direction of our experimental result, which shows that the
Board of Trade ohm is between two and three parts in ten thousand
larger than the true ohm could not, however, arise from 1. Nor
could it arise from either 2 or 3, still many experiments were made
to detect any evidence of the effective diameter of the disc being
larger than its true diameter, as measured in the Whitworth
machine. But no change in the pressure of the circumferential
brush-tubes, nor alteration in the shape of their ends, etc., indicated
that, with the brushes as we employed them, the effective diameter
of the disc differed from its true diameter.

Our thanks are due to the three students whose names are
given above for much assistance in carrying out the long series
of observations; to Mr Harrison for bringing to bear, from time to
time, the experience that he had previously gained in the use of
the Lorenz apparatus ; and we are especially indebted to Mr Mather
for the suggestive aid which he rendered us throughout the whole
of the present investigation.

TWENTY-FIFTH KEPOKT— BRISTOL, 1898.

APPENDIX PAGE

I. Cotnparisan of the Stafidard Coils used by Frofeseors J, Vtriamu
Jones and W, E, Ayrton in their determination of the absolute
resistance of Mercury with the Standards of the Association,
By R. T. Qlazebrook, F.R.S. 577

II. On the Determination of the Temperature Coefficients of two

lO^hm Standard Resistance Coils {Nos. 3873 and 3874)
used in the 1897 determination of the ohm. By M. Solomon 581

III. An Amphre Bcdatice. By Professor W. E. Atrton and

Professor J. V. Jones 589

In consequence of his appointment as Treasurer of the Associa-
tion, Professor Carey Foster has resigned the position of Chairman
of the Committee, which he has held for many years. The
Committee in asking for reappointment recommend that Lord
Rayleigh be the Chairman.

The standards of the Association have, since the opening of the
Cavendish Laboratory, been kept at Cambridge in the custody of
the Secretary. Mr Glazebrook has now left Cambridge for Liver-
pool, and the Committee at a meeting in London agreed that
Mr Glazebrook be authorised and requested to retain the custody
of the standards. In consequence, the various standards will in
the course of the autumn be installed in the Laboratoiy of
University College, Liverpool.

At the Toronto Meeting the Committee agreed that it was
a matter of urgent importance that the general question of the
absolute measurement of electric currents should be investigated,
and a grant of £75 was made for the purpose.

During the year Professors Ayrton and J. V. Jones have
concluded some preliminary experiments with this object, and
have designed a form of current weighing apparatus calculated to
give results of great accuracy. Drawings of the apparatus have
been Isdd before the Committee and the details of its working
explained to them. The estimated cost of the apparatus is £280 ;
to meet this the grant of £75 made last year remains in hand.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 577

After careful consideration and discussion the Committee, at
their meeting at Bristol, agreed unanimously to the following
resolution: —

The Committee, having heard fix)m Professors Ayrton and
J. V. Jones an account of their preliminary experiments on the
absolute determination of the ampere and their plans for the
construction of an absolute ampere balance, are of opinion that, in
view of the importance of the proposed experiments, application
should be made for leave to retain the unexpended balance, £75,
of the grant made last year, together with a further grant of £225.

Accordingly the Committee ask for reappointment and apply
for the above grant. They recommend that Lord Bayleigh be
Chairman, and Mr R. T. Glazebrook Secretary.

Appendix I.

Comparison of the Standard Coils used by Professors J, Viriamu
Jones and W. E, Ayrton in their determination of the absolute
resistance of Mercury with the standards of the Association.
By R. T. Glazebrook, F.R.S.

These coils consist of two tenth-ohm standards of manganin,
and two tenth-ohm standards of platinum silver.

The method employed in comparing the tenth-ohm standard is
described in the Report of the Committee for 1894 (Oxford
Meeting), Report, p. 128. In certain of the experiments the same
mercury cups were used as in 1894 ; in others, the cups used by
Professor J. V. Jones in his absolute measurements were employed.
The Standard Coils made use of were the following: Elliott,
No. 269; Elliott, No. 270; and Nalder, 3716, the last being a
ten-ohm standard, the others units.

The following values were found : —

Nalder, 4274, ^ 389.
R.T.G.*8 mercury cups iised

))

>»

J.V.J.'s
RT.G.'a

»>

»i

»>

»»

Mean

B. a.

•100049

14-2'*

•100051

141

•100045

143

•100030

134

•100054

15-7

•100050

16-1

•100030

131

•100042

13-7

•100044

137

•100042

141

•100043

141

•100044

14-2

578

PBACHCAL STANDARDS

Nalder, 4275, ^ 390.

R.T.Q.'8 mercury cups used-

»>

J.V?J.'8

'R.'T.QJs

jy.J.'8

»
)»

Mean

•100060

14-4'

■100061

144

•100067

14-5

•100040

13-4

•100063

15-8

•100066

161

•100045

131

•100052

13-7

•100055

13-7

•100054

141

•100055

141

•100054

14-3

The values found in 1894 were respectively

•100050 and 100053

in each case at 15*2°, and the differences are probably within the
errors of observation. If all the individual observations for both

Fig. 1. Ko. 4274.

Fig. 2. No. 4275.

•

t :

•

■^'-

«
> <

*"*:

•

f t

•

•
4-/4

•

hi.

-7^

9rf

-4

^ei:

,•'

Af'

/♦•

^^r

:^;

//•

Results of observations on the 'l-ohm coils used by Professors Ayrton and

J. V. Jones: —

Observations in 1894, thas •

Observations in 1897, thas +

The horizontal divisions are 0*1° C.

The vertical divisions are *000005 ohm.

1894 and 1897 be plotted, they will be found to overlap each
other, and it is difficult to assert that there has been any change.
If any exists it is certainly very small.

This is shown in figs. 1 and 2, in which the observations
indicated by dots give the results of the 1894 experiments, those
indicated by crosses the experiments of 1897. At a glance the
observations do not appear very good, but it must be remembered
that the vertical ordinates are drawn to a very large scale, the

FOR ELECTRICAL MEASUREMENTS

679

division being five-millionths of an ohm. For both coils the
resistance appears to reach a maximum at about 16*6^ C.

The ten-ohm coils were compared in the usual manner on the
Carey Foster bridge with the Standard Coil, Nalder, 3716.

The following are the values found : —

Nalder, 3873, ^ 367.

Date

December 9, 1897
11
13

n
n

i>
II
II

Mean

Value

9-9913
9-9917
9-9891
9-9877
9-9909

9-9901

Temperatare

NaMer, 3874, ^ 362.

December 9, 1 897

II
II
II

II
II
II
11

Mean

9-9907
9 9911
9-9886
9-9873
9-9904

US'*

14-4

13-5

1315

14-2

13-9

14-3°

14-3

13*5

13-1

14-05

13-9

The values found in 1893 and 1894 were as follows : —

For 3873, 99919 at 14•8^

If we take the temperature coefficient as *0028 — ^the value
given by Messrs Nalder — this becomes 9*9894 at 13*9**. Thus the
coil appears to have risen in value by *0007 ohm ;

while for 3874, the value found was 9*9926 at 14*9°.

Messrs Nalder give the temperature coefficient as '003, and
this leads to the value, 9*9896 at 13*9*", agreeing exactly with the
observations of December 1897.

The results of these observations are shown in figs. 3 and 4.
The dots refer to the 1893 observations, the crosses to those of
1897. It appears that No. 3874 has not changed ; with regard to
No. 3873, a change is indicated. As to this change, it appears

37—2

580

PRACTICAL STANDARDS

from the note-book that there was some doubt as to the tempera-
ture of one of the observations in 1893 ; it is recorded as 14° ; the

fig. 4. No. 8874.

Fig. 3

. Mo.

887S.

S'99S9

9'9999

■

9'99U

•

9'S990

9-9990

•

99B»9

^197

j§

/9 t9 it m

Results of obsenrations on the 10-ohxn ooils, ased bj Professors Ajrton and

J. V. Jones: —

Obsenrations of 1893, thus •

Obserrations of 1897, thus +

The horizontal divisions are 0*1° G.

The vertical divisions are *0002 ohms.

observation shows that the temperature must have been about
13'7**. Furthermore, the value of the ten-ohm standard used for

Fig.

3»

. No.

8878.

Revi9ed9

" '" 1

9-99*9

•

. J

99919

•

9 '39 99

99890

9-9990

Observations of 1898, corrected to final value of 10-ohm standard, shown thus •
Observations of 1897, thus +

FOR ELECTRICAL MEASUREMENTS 581

3873 was not definitely determined in 1893. If allowance is made
for these two fSeu^ts, the value of 3893 at IS^"" is raised to 9'9923 ;
thus the curve shown in fig. 3* is obtained, and the apparent
change in value is reduced to about *0003 ohm, or three parts in
one hundred thousand On the whole, then, I conclude that 3873
has changed since 1893 by about this amount, while 3874 has
remained stationary in value. The discrepancy between this
conclusion and that given by Mr Solomon in Appendix II. depends
on the different values employed for the temperature coefficients.

The values of these coefficients obtained over so short a range
are not of much importance. Still, in view of Mr Solomon's
determination, they may be given. They are : For 3873, '000283 ;
and for 3874, •000277. These values are relative to the standard
coils of the Association.

Appendix II.

On the Determination of the Temperature Coejffictente of Two
lO'Ohm Standard Resistance Coils (Nos. 3873 and 3874) iLsed
in the 1897 Determination of the Ohm, By M. Solomon.

In the determination of the ohm made by Professor W. E.
Ayrton and Professor J. Viriamu Jones in 1897 (Report, 1897,
p. 212), four standard resistance coils were used, two of which had
a resistance of 10 ohms each, and two of O'l ohm each. Values
for the temperature coefficients of these coils had been calculated
, from four accurate determinations of their resistance made, two by
Mr Qlazebrook in 1894 and 1897, and two by the Board of Trade
in 1896 and 1897 {The Electrician, vol. XL. p. 39). The values
thus obtained neither agreed with one another nor with the
coefficients given by the makers, Messrs Nalder Bros. & Co. It
therefore became necessary to make as accurate a determination
as possible to endeavour to find the correct values for the co-
efficients. The following Paper gives the results of the tests made
on the two 10-ohm coils (Nos. 3873 and 3874), the tests on the
other two coils being not yet completed. These two coils are of the
B. A. pattern, and cure made of platinum silver wire. A preliminary
series of tests made on one of the coils showed that to attain the
required accuracy special precautions would have to be taken to
keep the coils at steady temperatures. Each coil was therefore

682 PRACTICAL STANDARDS

placed in an oil bath, the temperatare of which was automatically
regulated. In making the determination of the temperature
coefficient of one coil, the other was used as a standard, and was
kept at a constant temperature throughout the whole series of
tests. The coil under test was maintained at a steady temperature
for some time, and a measurement of the difference of resistance
between it and the standard was then made by means of a Carey
Foster bridge. The temperature of the coil being tested was then
altered and a fresh measurement taken, this being repeated for
several temperatures.

The apparatus used in the measurements was arranged in the
following manner. The standard coil was placed in an oil bath
with two vessels, in the inner of which the coil itself and a
carefully standardised thermometer were immersed. In the outer
bath was the bulb of an alcohol thermometer, the mercury index
of which, when the temperature rose too high, completed the
circuit of an electromagnet and battery, and caused the gas which
heated the bath to be put out. On the bath cooling the
circuit of the electromagnet was broken, and the gas turned on
and relighted by a bypass. This thermostat was very sensitive,
the temperature of the inner bath rarely varying so much as
O'CS"" C. in a day, and in a run of ten days undergoing a maximum
variation of 0*3"" C. The thermostat in which the coil under test
was placed was not so sensitive, but was designed to work over
a greater range of temperature. The coil and thermometer were
placed in an inner bath, and in the outer bath was a large brass
bulb filled with alcohol. The expanding alcohol either passed into
a small reservoir, or, when the passage to this was closed by
shutting a stop-cock, it expanded into one arm of a glass U tube,
thereby forcing a mercury index at the bottom up the other arm ;
this index out off the gas supply by closing the aperture of the
inlet tube. On cooling the index sank; the gas was turned on
and relighted by a bypass. Regulation of the temperature accord-
ingly did not take place until the path leading to the reservoir
was closed, so that regulation at any desired temperature could be
obtained by leaving the stop-cock open until that temperature has
been reached. In this case, as also in the other thermostat, the
bath was not heated directly by the gas jet, but a baffle plate was
interposed. The daily variation of temperature with this apparatus
was about 0*2° C, but the changes were so slight and so slow that

FOR ELECTRICAL MEASUREMENTS 583

the probable error introduced would be less than that caused by
error in reading the thermometer. The bath was always kept at
a constant temperature for some hours before readings were taken.
With these arrangements it was safe to assume that the tempera-
ture of the coil was the same as that read off from the thermometer.
The terminals of the coil dipped into mercury cups in one end of
a pair of stout copper rods, half an inch in diameter, the other
ends of which rested in mercury cups on a Carey Foster bridge.
The leads from each of the coils were of very small and approxi-
mately equal resistance, so that no appreciable error could be
introduced by alteration in their resistance with change of
atmospheric temperature. Also, as a part of each lead was inside
the thermostat, heat lost by conduction along the leads would be
withdrawn from this part and not from the coil itself

The measurements were made with a Carey Foster bridge, the
platinum silver slide-wire of which had been previously calibrated.
This wire was 50 centimetres long, and had a resistance of 0*001859
ohm per half centimetre at 13*5^ C, and was graduated in half
millimetres. Correction was made for alteration in the resistance
of the wire due to change in its temperature, an increase of I'' C.
producing an increase of 0*000011 ohm in the resistance of half
a centimetre. Determinations of the difference of resistance
between the two coils were made at intervals of about an hour,
and if two or three quite consistent readings could be taken these
were considered as correct, but where discrepancies occurred the
mean of several results was taken. The slight changes in the
temperature of the standard were easily allowed for, since it could
be assumed that for such small changes the two coils had the same
temperature coefiBcients. So if the standard, instead of being at
the temperature t, were at the temperature t + S, and if the coil
under test were at the temperature t\ it was assumed that the
standard was at temperature t, and the coil under test at the
temperature t' — S.

There are four principal sources by which error can be intro-
duced, viz. error in obtaining the correct position of balance, error
in the value of the temperature coefficient of the slide-wire, error
in reading the temperature of the standard coil, and error in
reading the temperature of the coil under test. As regards the
first of these, the sensibility of the arrangement was such that
a change of half a millimetre in the position of the slider produced

584

PRACTICAL STANDARDS

a deflection of about a centimetre on the galvanometer scale, so
that balance could easily be obtained correct to 0*05 mm. The
error due to not knowing the temperature coefficient of the slide-
wire with certainty will not be great, as all the measurements
were made at temperatures near to 13*5'' C, at which temperature
its resistance was known. The greatest error is introduced in
reading the thermometers which were graduated in tenths of a
degree, each division being about 0*6 mm. in length, so that the
temperatures could not be read with certainty to less than 0*02° C.
If all these errors should be made in one direction in making one
determination of difference of resistance, and all in the opposite
direction in making a second, there is a possible maximum error
of about 3 per cent, in the value of the temperature coefficient
calculated from these two determinations. This is, however, highly
improbable, and, moreover, makes no allowance for taking the
mean of several readings. The error in the temperature coefficient
is probably not greater than 1 per cent., if as great.

The following summarises the results of the experiments : —

Ten-Ohm Standard Coil, No. 3873.

A series of tests was made on this coil in the manner above
described, lasting from March 22 to April 1, 1898. Determinations
were obtained of the difference between the resistance of No. 3873
at six different temperatures, and the resistance of No. 3874 at
16*70** C, with the following results : —

Temperature of
No. 3873

Exoess resist, in ohms of Change of
No. 3873, above No. 3874, resist, per
atl6-70«C. PC.

(a) 16-3rC.
(6) 19-33'' C.

(d) 22-43" C.

(e) 25-43'' C.
(/) 26-22° C.

-0-001896 ^^ 0-00307

+ 0007380 -^-=£---1--:. 0-00291)
+0-01545 -^::'r^-* 0-00291$
+0-01639 *^1^^=- 0-00278)
+0-02470 *^cr---* 0'00277J
+0-02678 -^- 0-00274

From readings a, 6, c, and e, and from the measurement of the
resistance of the coil made by Mr Glazebrook in December, 1897,
giving jBi8-9» = 9*9901 ohms, we get

R^ = 9-9398 (1 + 0-000397^ - 0000002 (4) e«).

After testing this coil the other coil (No. 3874) was tested.

FOR ELECTRICAL MEASUREMENTS

585

and then three check tests were made on this coil with the
following results: —

Temperature of
No. 3873

Excess resistaooe in ohms above No. 8874,

at 16-70° C.

(g) 19-62'* C.
(h) IS-SS" C.
(k) 18-13^ C.

+0008398
-0-003074
+0-003704

These points lie well on the curve obtained in the former tests
(see fig. 5). Prom the formula given above the coil will have the
correct resistance of 10 ohms at 17*0** C.

■CIS
#2#

Fig.

£kC4S4 r9^i$U»e9MI9k»A

'f*A

orit

JB4.^a

C9i

fon

^f*0

r/r;

•014'

•9m

'4f/£

)pjm

nc»

•MM

9*^

M»t\

'B/t

/■M

k*'

>OfP

\t9n

•99i

^k..

Im»

ft»

•>M9

mtf

tot

r<*

094

fsr

to H m if m m 00

V it OS 29 OS to tf

ToM^orotorot'o *C

Carve showing ohaDge of resistanoe of 10-ohm standard ooU, No. 8873,
with change of temperature. Ordinatea give excess of resistance of
8878 above 8874 at 16-70° C.

To compare the temperature coefficient here obtained with
those previously determined we have four measurements of resist-
ance, as follows : —

A, Mr Qlazebrook in March, 1894.

B, Board of Trade in Nov. 1896.
(7. Board of Trade in Aug. 1897.
D. Mr Olazebrook in Dec. 1897.

Resistance = 9*9923 ohms at 14*8'* C.
» 9-992994 „ 14-86''C.
« 10-00712 „ 19-3* C.
= 9-9901 „ 13-9' C.

»»

586

PRACTIOAL STANDARDS

These furnish data for calculating the temperature ooeflScient,
and we have also the value given by the makers, Messrs Nalder
Bros. & Co. : —

Observer

Range of
Temperatare

Temperatare

coefficient

per rC.

Coefficient
from these

tests for
same range

Messrs Nalder Bros. & Co....
Tests A and C

17-0° -22-0°C.
14-8° -19-3°C.
14-86°- 19-3° C.
1315°- 14-4° 0.
13-9° -19-3°C.

0-000276
0000331
0-000320
0-000299
0-000317

0-000303
0-000315

o-ooa3i5

0-000330
0-000317

„ -Sand C

„ D

„ Dand C

This table shows that the coefficients calculated from tests B
and G and from tests D and C are both in very close agreement
with those I obtain for the same range of temperature.

Tm-Ohm Standard CaU, No. 3874.

A series of tests on coil No. 3874, lasting from May 19 to
May 31, 1898, were made, and in addition we have one result
fix>m the tests on eoil No. 3873. Altogether we have the
following : —

Temperature of
No. 3874

Excess resist, in ohms of Change of
No. 8874, above No. 387S, resist, per
at 17-25° C. rC.

(a) 16-70° C.
(6) 17-46° C.

(d) 19-09° C.
(c) 21-37° C.
(/) 22-22° C.
{g) 24-46° C.

-0-00095 —v. 0-00334

+0t)0363 _ _ _ 0-00323

+000689 -= =f - = _ 0-00306

+ 0-01382 — ^^ 0-00286

+0-01641 — --rr^=- 0-00278
+0-02264 "^

From readings a, c, e and g, and Mr Glazebrook's determination
of the resistance in December, 1897, which gave iiisir = 9*9896
ohms, we get

Rt = 9-9313 (1 + 0000481^ - 0*000004 (2) ^).

FOR ELECTRICAL MEASUREMENTS

587

Check tests were made on this coil after those on coil No. 3873
had been made, and gave the following results : —

Temperainre of
No. 8874

Exeess resistance in ohms above No. 3873
at 17-26« C.

(h) 1877'C.
(k) 20-15'* C.

+000583
+0-01040

All the nine points lie on a smooth curve (see fig. 6). These
tests make the coil correct at IG'd"* C.

i»€0$SrtStgt90^

Pig.

, 6.

(^^

»nn*

^'i

'iji

'/^

•

'WMW

'WWW

«fr

#•*

•wim

'0ii

•

J»M

efi

f»

•0t9

■#P#

.. J

«Br

V¥

'09€

kJt

0s4

■00€

kr\

!!*»

•##1

tji

0
"9tt

4-/'>

-mm.

M H m n m t9

m » n u t^ u

Carve showing change of resifltanoe of 10-ohm standard coil. No. 8874,
with change of temperainre. Ordinates give ezcess of resistance of
8874 above 8878 at 17-86<>C.

For purposes of comparison we have a similar set of data to
those used for the other coil. The four measurements of resistance
gave the following results : —

A. Mr Glazebrook in March, 18d4.

B, Board of Trade in Nov. 1886.
a Board of Trade in Aug. 1897.
D. Mr Qlazebrook in Dec 1897.

Re8i8tauce:= 9*9926 ohms at 14*9* C.
» 9-993213 „ 14-91^0.
« 10-00775 „ 19-3' C.
= 9-9896 „ 13-9' C.

n
»»

From these we get the following values for the temperature
coefficient : —

588

PRACTICAL STANDARDS

Observer

Range of
Temperature

Temperatare
coefficient
perl°C.

Coefficient
from these

tests for
same range

Messrs Nalder Bros. & Co....
Tests A and C

17-0" -22-0'C.
14-9° -19-3''C.
14-9r-19-3'*C.
13-r -14-3'C.
13-9*' -19-3°C.

0-000300
0-000346
0-000333
0000279
0-000338

0O00316
0O00336
0-000336
0-000365
0-000341

„ B a.nd C

„ I>

„ DandC

Here again the same two sets of tests, viz. tests B and C, and
tests D and C, give values for the temperature coefficient very
nearly equal to those I obtain for the same range of temperatures.

Since for both coils the temperature coefficients that I obtain
agree with those calculated from the three last measurements of
resistance — namely, the two metisurements by the Board of Trade
and Mr Glazebrook's last test — these experiments seem to show
that the coils have not changed since 1896, but that the resistances
as measured in 1894 were a little lower than those that would
now be obtained at the same temperatures.

This conclusion may be better illustrated by calculating what
would be the resistances at the temperatures of the various tests,
on the assumption that the coefficients I obtain are correct, and
that Mr Glazebrook's last (in December, 1897) is correct. We
then get the following: —

Temperatnre of test

Besistanoe as
ineasared in ohms

Besistance as cal-
culated in ohms

/U-S" C.
Coil No. 3873 }9!P'''c.'

{n-9° C.

14-9° C.
Coil No. 3874 . Ig.^ °c;

llS-G" c!

(••1)
(B)

iC)

(.B)
(0)

9-9923

9-9930

10-0071

9-9901

9-9926

9-9932

10-00775

9-9896

9-9930

9-9931

10-0071

9-9901

9-9932

10-0079

9-9896

Thus we see that the 1894 measurements (-4) are too low by
as much as 7 parts in 100,000 in the case of coil No. 3873, and
6 parts in 100,000 in the case of No. 3874. In the case of the
other two measurements the calculated results only differ from
the observed results by 1 or 1*5 parts in 100,000.

FOR ELECTRICAL MEASUREMENTS 589

These experiments were carried out in the laboratory of the
Central Technical College, South Kensington, and I am much
indebted to Professor Ayrton and Mr T. Mather for their valuable
guidance and advice.

Appendix III.

An Ampire Balance, By Professor W. E. Ayrton, F.R.S.,
and Professor J. ViRiAMU JoNES, F.R.S.

The Report of the Committee on Electrical Standards for 1897
ended with the following paragraph: ''It thus appears to be a
matter of urgent importance that a redetermination of the electro-
chemical equivalent of silver should be made and that the general
question of the absolute measurement of electric currents should
be investigated...." This work we were asked by the Committee
to carry out, and a grant of £75 was voted in its aid. We were
thus led to examine into the methods which had been employed
by Lord Rayleigh, Professor Mascart, and others, for determining
the absolute value of a current, as well as to consider some other
methods which have not, as far as we know, been hitherto used.

After much consideration we decided to adopt a form of
apparatus which, while generally resembling the t)rpe employed
by some previous experimenters, possessed certain important
differences, and, before expending any part of the grant of £75, to
construct, without expense to the British Association, the following
preliminary Ampere Balance.

On a vertical cylinder about 17 inches high and 6*8 inches in
diameter we wound two coils, about 5 inches in height, separated
by an axial distance of 5 inches. The coils consisted each of
a single layer of about 170 convolutions of wire and were wound
in opposite directions. From the beam of a balance there was
suspended, inside this cylinder, a light bobbin about 4 inches in
diameter, on which was wound a coil about 10 inches long
consisting of a single layer of 360 convolutions, and the whole
apparatus was so adjusted that when the beam of the balance was
horizontal the inner and outer coils were coaxial and the top and
bottom of the inner suspended coil were respectively in the mean
planes of the outer stationary coils.

This arrangement was adopted because with coils consisting of

690 PRACTICAL STANDARDS

only one layer the geometrical dimensions could be accurately
determined, and because the shapes of the coils lent themselves to
the use of the convenient formula, readily expressible in elliptic
integrals, for the force, F, between a uniform cylindrical current
sheet and a coaxial helix, viz. : —

where 7 is the current per unit length of the current sheet, 7^ the
current in the helix, and Mi and if, the coefficients of mutual
induction of the helix and the circular ends of the current sheet*.

The value of a particular current of about 0*63 ampere having
been determined absolutely by means of this apparatus, the rate at
which it would deposit silver under specified conditions was
ascertained indirectly, by observing its silver value on a Kelvin
balance which had been kept screwed down in a fixed position for
several years past and which had been calibrated many times
during that period by reference to the silver voltameter.

The result of this preliminary investigation showed that the
silver value of the trus ampere was so nearly equal to the reputed
value, viz. 1*118 milligrammes per second, as to require the use of
an apparatus still more perfectly constructed, and therefore of
a much more expensive character, to enable the error, if any, in
this value to be ascertained with accuracy.

We, therefore, started on the design of the instrument, of
which we now submit the working drawings, and for the future
construction of which we would ask for a grant of £300 including
the unexpended grant of £75 voted last year. And we anticipate
that this new piece of apparatus may prove worthy of constituting
a national Ampere Balance, the counterpoise weight for which will
be determined purely by calculation based on the dimensions of
the instrument, the number of convolutions of wire in the three
coils, and the value of the acceleration of gravity at the place
where the instrument may be permanently set up. In this
particular it will differ entirely from the " Board of Trade Ampere
Standard Verified, 1894," which has had its counterpoise weight
adjusted so that the beam is horizontal when a current passes

* See Proceedings of the Royai Society, toI. lxiu. : '* On the CaloaUtion of the
Coefficient of Mutual Induction of a Circle and a Coaxial Helix, and of the
Electromagnetic Force between a Coaxial Current and a Uniform Coaxial Ciroolar
Cylindrical Current Sheet.*' By Profeesor J. V. Jones.

FOR ELECTRIGAL MEASUREMENTS 591

through the instrument, which will deposit exactly 1*118 milli-
grammes of silver per second under specified conditions. In fisM^t,
the proposed Ampere Balance and the existing Ampere Standard
will differ exactly in the same way as do a Lorenz apparatus and
the '' Board of Trade Ohm Standard Verified, 1894."

We have to express our thanks to Mr Mather for taking charge
of the construction and use of the preliminary apparatus, for
checking all the calculations in connexion with the determination
of the electrochemical equivalent of silver that was made with it,
as well as for superintending the making of the working drawings
of the new Ampere Balance. We have also to thank Messrs
W. H. Derriman and W. N. Wilson, two of the students of the
City and Quilds Central Technical College, for their cordial
assistance in canying out the work.

TWENTY-SIXTH REPORT.— DOVER, 1899.

APPENDIX PAOB

I. On the Mutual Induction of Coaxial Helices, By Lord Ratlsioh 593

I I . Proposals for a Standard Scale of Temperature based on the Plati-

num Resistance Thermometer. By Professor H. L. Callendar 595

III. Comparison of Platinum and Oas Thermometers, By Dr P.

Chafpuis and Dr J. A. Harker 597

IV. On the Expansion of Porcelain with Rise of Temperature. By

T. G. Bedford 600

The Committee have been engaged during the year on the
consideration of the details of the new ampere balance, for which
a grant of £300 was voted at Bristol.

Professors Ayrton and Viriamu Jones have completed the
plans and specifications, and the construction of the balance has
been authorised.

An important addition to the plan proposed at Bristol consists
of an arrangement for adjusting accurately the position of the
fixed coils. Sir Andrew Noble has generously undertaken to have
this constructed at Elswick fi-ee of cost, and the Committee desire
to thank him for the oflfer, which they have gladly accepted.

In consequence of the fact that the balance is not yet com-
pleted, the grant of £300 made last year has not been expended,
and the Committee apply for its renewal.

An appendix to the Report contains a proof by Lord Rayleigh
of a theorem due to Professor J. V. Jones, on which the mathe-
matical theory of the new balance is based.

Details of the balance are reserved until it has actually been
constructed.

Professor Callendar has brought before the Committee pro-
posals for the adoption of a standard scale of temperature based
on the Platinum Resistance Thermometer. These are printed in
an appendix and formed the basis of a discussion in the Section.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 593

A Sub-Committee has been formed to consider these proposals and
to report to the Committee.

The ordinary testing of standards has been interrupted by the
removal of the Secretary to Liverpool, and still further by his
proposed removal to Eew. With respect to this the Committee
have passed the following resolution: —

That Mr R T. Qlazebrook, as Secretary of the Committee, be
authorised and requested to retain the custody of the Electrical
Standards of the Association, and to remove them from Liverpool
to London when he takes up his post as Director of the National
Phjrsical Laboratory.

The removal of the Standards and the investigations of
Platinum Thermometry will necessitate some expenditure during
the year.

The Committee therefore recommend that they be reappointed,
with the addition of Sir William Roberts- Austen and Mr Matthey,
and with a grant of £25 in addition to the unexpended balance
(£300) of last year's grant, and that Lord Rayleigh be Chairman
and Mr R. T. Glazebrook Secretary.

Appendix I.
The Mutuiil Induction of Coaxial Helices. By Lord Rayleigh.

Professor J. V. Jones* has shown that the coefficient of mutual
induction (M) between a circle and a coaxial helix is the same as
between the circle and a uniform circular cylindrical current-sheet
of the same radial and axial dimensions as the helix, if the currents
per unit length in helix and sheet be the same. This conclusion
is arrived at by comparison of the integrals resulting from an
application of Neumann's formula ; and it may be of interest to
show that it may be deduced directly from the general theory of
lines of force.

In the first place, it may be well to remark that the circuit of
the helix must be supposed to be completed, and that the result
will depend upon the manner in which the completion is arranged.
In the general case the return to the starting-point might be by
a second helix lying upon the same cylinder; but for practical

* Proe, Roy, Soc. vol. Lxni. (1897), p. 192.
B. A. 38

594 PKACTICAL STANDARDS

purposes it will suffice to treat of helices including an integral
number of revolutions, so that the initial and final points lie upon
the same generating line. The return will then naturally be
effected along this straight line.

Let us now suppose that the helix, consisting of one revolution
or of any number of complete revolutions, is situated in a field of
magnetic force symmetrical with respect to the axis of the helix.
In considering the number of lines of force included in the
complete circuit, it is convenient to follow in imagination a radius-
vector drawn perpendicularly to the axis fix)m any point of the
circuit. The number of lines cut by this radius, as the complete
circuit is described, is the number required, and it is at once
evident that the part of the circuit corresponding to the straight
return contributes nothing to the total*. As regards any part of
the helix corresponding to a rotation of the radius through an
angle d6, it is equally evident that in the limit the number of
lines cut through is the same as in describing an equal angle of
the circular section of the cylinder at the place in question, whence
Professor Jones's result follows immediately. Every circular
section is sampled, as it were, by the. helix, and contributes
proportionally to the result, since at every point the advance of
the vector parallel to the axis is in strict proportion to the
rotation. It is remarkable that the case of the helix (with straight
return) is simpler than that of a system of true circles in parallel
planes at intervals equal to the pitch of the helix.

The replacement of the helix by a uniform current-sheet shows
that the force operative upon it in the direction of the axis (dM/dx)
depends only upon the values of M appropriate to the two terminal
circles.

If the field is itself due to a current flowing in a helix, the
condition of symmetry about the axis is only approximately
satisfied. The question whether both helices may be replaced by
the corresponding current-sheets is to be answered in the negative,
as may be seen fix>m consideration of the case where there are two
helices of the same pitch on cylinders of nearly equal diameters.
In one relative position of the cylinders the paths are in close

* This would be trae bo long as the return lies anywhere in the meridional
plane. In the general case, where the number of convolutions is inoomplete, the
return may be made along a path composed of the extreme radii Tectores and of the
part of the axis intercepted between them.

FOR ELECTRICAL MEASUREMENTS 595

proximity throughoat, and the value of M will be large, but this
state of things may be greatly altered by a relative rotation
through two right angles.

But although in strictness the helices cannot be replaced by
current-sheets, the complication thence arising can be eliminated
in experimental applications by a relative rotation. For instance,
if the helix to which the field is supposed to be due be rotated,
the mean field is strictly symmetrical, and accordingly the mean
M is the same as if the other helix were replaced by a current-
sheet. A further application of Professor Jones's theorem now
proves that the first helix may also be so replaced. Under such
conditions as would arise in practice, the mean of two positions
distant ISO"*, or at any rate of four distant 90°, would sufBce to
eliminate any difference between the helices and the corresponding
current-sheets, if indeed such difference were sensible at all.

The same process of averaging suffices to justify the neglect of
spirality when the observation relates to the mutual attraction of
two helices as employed in current determinations.

Appendix II.

Proposals for a Standard Scale of Temperature based on the
Platinum Besistance Thermometer, To he submitted to the
Electrical Standards Committee, Drawn up by Professor
H. L Callendar, M.A., F.R.S.

The following proposals are submitted in consideration of the
importance of adopting a practical thermometric standard for the
accurate verification and comparison of scientific measurements of
temperature. The gas thermometer, which has long been adopted
as the theoretical standard, has given results so discordant in the
hands of different observers at high temperatures, as greatly to
retard the progress of research.

The arguments in favour of the adoption of the platinum
resistance thermometer as a practical standard were given by
Professor H. L. Callendar, in a paper " On the Practical Measure-
ment of Temperature," communicated to the Royal Society in
June 1886, and published in the PhU. Trans, in the following year.
These arguments have since been confirmed and strengthened by
the work of many independent observers.

38—2

596 PRACTICAL STANDARDS

The Electrical Standards Committee of the British Association
have done so much in the past with reference to the adoption of
the present electrical standards, and more recently in connexion
with the adoption of the joule as the absolute unit of heat, that it
would appear to be the most appropriate authority for the
discussion and approval in the first instance of proposals relating
to an electrical standard of thermometry.

The suggestions for the standard scale of temperature here
proposed may be embodied in the following resolutions: —

(1) That a particular sample of platinum wire be selected,
and platinum resistance thermometers constructed to serve as
standards of the platinum scale of temperature.

(Note, — A degree centigrade of temperature on the scale of
a platinum resistance thermometer corresponds to an increase of
resistance equal to the hundredth part of the change of resistance
between 0" and 100° C. In other words temperature pt on the
platinum scale is defined by the formula

pt = 100 (iZ - iJ°)/(iJ' - i2°),

in which the letters -R, -R°, and JB' stand for the resistances of the
thermometer at the temperatures pt, 0**, and 100° C, respectively.
The melting-point of ice is taken as the zero of this scale in
accordance with common usage.)

(2) That the scale of temperature t deduced from the standard
platinum scale by means of the parabolic difference formula,

t-pt^d{t/lOO'-l)t/lOO,

which has been proved to give a very close approximation to the
true or thermodynamic scale, be recommended for adoption as
a practical standard of reference, and be called the British Associa^
tion Scale of Temperature.

{Note. — The gas thermometer would still remain the ultimate
or theoretical standard, and the exact relation of the British
Association scale to the absolute scale would be the subject of future
investigation. In the present state of experimental science, the
difference between the two scales over the greater part of the range
is less than the probable errors of measurement with the gas ther-
mometer, and the possible accuracy of measurement with a platinum
thermometer, especially at high temperatures, is of a much higher
order than with the gas thermometer. Measurements directly
referred to the British Association scale would therefore be of

FOR ELECTRICAL MEASUREMENTS 597

greater permanent value, because they could be subsequently
corrected when the relation between the scales had been more
accurately determined.)

(3) That the value of the difference-coefficient d in the
parabolic difference-formula be determined for the British Associa-
tion standard thermometers by reference to the boiling-point of
sulphur as a secondary fixed point in the manner described by
Callendar and Griffiths, Phil. Trans. A, 1891.

{Note. — It is probable that this method gives the best results
over the whole range at temperatures above — 100° C. At very
low temperatures there appear to be singularities in the resistance
variation of metals which require further investigation. The
boiling-point of liquid oxygen would be a more convenient
secondary fixed point to choose for low temperature research,
especially for testing thermometers the construction of which did
not permit their exposure to a temperature as high as that of
boiling sulphur.)

(4) That the temperature of the normal boiling-point of
sulphur under a pressure of 760 mm. of mercury reduced to 0° C,
and latitude 45°, be taken for the purposes of the British Associa-
tion scale as 444*53° C, as determined by Callendar and Griffiths
(loc. cit), with a constant pressure air-thermometer.

(Note. — Until the relation between the various gas-thermometer
scales, and the expansion of glass and porcelain, have been more
accurately determined, it does not appear that anything would be
gained by changing this value to which so much accurate work
has already been referred.)

Appendix III.

A Comparison of Platinum and Gas Thermometers made at the
International Bureau of Weights and Measures at Sevres.
By Dr P. Chappuis and Dr J. A. Harker.

Professor Callendar in 1886 investigated the method of
measuring temperature based on the determination of the electrical
resistance of a platinum wire.

He pointed out that if Ro denote the resistance of the spiral of
a particular platinum thermometer at 0°, and i{, its resistance at
100°, we may establish for the particular wire a scale, which we

598 PRACTICAL STANDARDS

may call the scale of platinum temperatures^ such that if £ be the
resistance at any temperature r°, this temperature on the

p jy

platinum scale will be ^ jr x 100 degrees. For this quantity

Callendar employs the symbol pt

In order to reduce to the standard scale of temperature the
indications of any platinum thermometer, it is necessary to know
the law connecting pt and T. These are identical at 0** and 100°,
but the determination of the relationship between them at other
temperatures is a matter for experiment.

The work of Callendar established for a particular sample of
platinum the relation

[ T r T \1

^ = ^-^^=nioorioo|J

over the range 0° to 600°, T being measured on the constant
pressure air-scale, and S being a constant.

Later experiments by Callendar and Griffiths showed that this
relation holds for platinum wires generally, provided that they are
not very impure. They propose that the value of S, the constant
employed in the formula, should be determined by taking the
resistance of the thermometer in the vapour of sulphur, and
a new determination by them of the boiling-point of this substance,
under normal pressure, gave 444*53'' on the air-scale.

The present communication gives a short account of some
experiments which are the outcome of the collaboration of the
Kew Observatory Committee and the authorities of the Bureau
International des Poids et Mesures at Sevres, for the purpose of
carrying out a comparison of some platinum thermometers with the
recognised International Thermometric Standards. A full account
of the work will shortly appear in the Philosophical Transactions
of ike Royal Society, and in the Travaxuc et Mimoires du Bureau
International des Poids et Mesures.

A new specially designed resistance-box, together with several
platinum thermometers, and the other accessories needed, were
constructed for the Eew Committee, and after their working had
been tested at the Kew Observatory, they were set up at the
Sfevres Laboratory in August 1897. The resistance-box in its
general design was very similar to the one previously described
before this Section by Mr Griffiths, but the plugs were replaced
by a special form of contact maker, and the coils were of manganin

FOR ELECTRICAL MEASUREMENTS 599

instead of platinum-silver. The methods adopted for the stan-
dardisation of the apparatus only differed in a few details from
those of Callendar and Griffiths.

The comparisons made between the platinum thermometers
and the standards of the Bureau may be divided into several
groupa The first group of experiments covers the range (— 23^
to SO""), and consists of a large number of comparisons between
each platinum thermometer and the primary mercury standards
of the Bureau, whose relation to the normal hydrogen scale had
previously been studied by one of us.

Above 80* the mercury thermometers were replaced by a gas ther-
mometer, constructed for measurements up to high temperatures.

We at first attempted to use hydrogen as the gas for these
measurements, but, owing probably to a slow chemical action
taking place between the gas and the glass reservoir in which it
was enclosed, we were afterwards compelled to substitute nitrogen,
which we have not observed to exert any action on the material
of the envelope up to a full red heat.

The comparisons between 80"" and 200'' were made in a vertical
bath of stirred oil, heated by different liquids boiling under
varying pressures. For work above 200"" a bath of mixed nitrates
of potash and soda was substituted for the oil tank. In this bath
comparisons of the two principal platinum thermometers with the
gas thermometer were made up to 460'', and with a third ther-
mometer, which was provided with a porcelain tube, we were able
to go up to bdO"", the glass reservoir of the gas thermometer being
replaced by one of porcelain, whose dilatation had previously been
measured by the Fizeau method. Comparisons of the platinum
and gas scales were carried out at over 150 different points, each
comparison consisting of either ten or twenty readings of the
different instrument&

By the intermediary of the platinum thermometers a deter-
mination of the boiling-point of sulphur on the nitrogen scale was
also made. Three independent sets of determinations of this point
gave the following results : —

(1) Platinum thermometer K. 9, and gkus gas thermometer, 445*27*.

(2) „ „ K. 9, porcelain „ 445-26'.

(3) „ „ K.8, „ „ 446-29*.

The mean of these, 445*27'", representing the temperature on
the scale of the constant volume nitrogen thermometer, differs

600 PRACTICAL STANDARDS

by only 0*7° from that found by Callendar and Griffiths for the
same temperature expressed on the constant pressure air-scale.
If, for the reduction of the platinum temperatures in oar
comparisons, we adopt the parabolic formula, and the value of
S obtained by assuming our new number for the sulphur point, we
find that below 100° the differences between the observed values
on the nitrogen scale and those deduced from the platinum
thermometer are very small, seldom exceeding OOl"*, and that
even at the highest temperatures the difference only amounts to
a few tenths of a degree.

Appendix IV.

On the Expansion of Porcelain with Rise of Temperature.
By T. G. Bedford, B.A., Cambridge.

In direct comparisons of the scales of temperature given by air
and by platinum-resistance thermometers at high temperatures,
the expansion of the porcelain envelope enters as a small
correction.

In the experiments described in this paper, a direct deter-
mination of the linear expansion of porcelain was made at
temperatures from 0° C. to 830° C. The method used was
essentially the same as that described by Callendar (Phil. Trans.
1887, A, p. 167).

On a tube of Bayeux porcelain two fine transverse marks were
made at a distance about 91*3 cm. apart. The tube was heated to
as high a temperature as possible in a gas furnace, and was then
slowly cooled by diminishing the gas supply. During cooling the
variation in the distance between the marks was determined by
a pair of reading microscopes which were mounted on stone blocks
and not touched during an experiment except by the screw-head.
The readings of the microscopes for a standard length (a glass
tube kept in melting ice) were taken at intervals.

The temperatures corresponding to the length measurements
were deduced from the resistance of a platinum wire running fix>m
mark to mark in the axis of the tube and supported on a plate of
mica. The resistances in ice and steam were taken after each
exposure to a high temperature. The sample of platinum wire
frx)m which the piece used in these experiments was cut is known

FOR ELECTRICAL MEASUREMENTS 601

to have a value of B, in Callendars formala, from 1*50 to 1*51.
The value B » 1*505 was used, and thus a direct detennination of
the resistance at the temperature of boiling sulphur was avoided.
An error of *01 in B causes an error of less than 1** in the calculated
value of t at lOOO'' C.

Four main experiments were made and the results were
plotted.

From 0° C. to 600° C. the results are represented fairly well
by the formula

/, = i^ (1 + 34*25 X 10-^ « + 10*7 X 10~'« t').

Above 600° C. the points are more erratic, but still do not depart
far on either side from the curve given by the above formula.

A length of about 6 cm. at either end of the tube was not
directly heated by the furnace. Hence there is an uncertainty
due to the ends (greater at the higher temperatures), since the
coefficient of expansion varies with the temperature.

For cubical expansion the above formula gives

t;« « t; (1 + 102*75 x lO"'^ + 32*4 x 10-"f«).

TWENTY-SEVENTH REPORT— BRADFORD, 1900.

Appendix* — Note on an Improved Resistance Coil, By

Robert S. Whipple p. 604

During the year the resistance coils and other apparatus
belonging to the Committee have been removed to Richmond.
Most of the apparatus has been set up in an outbuilding attached
to the Kew Observatory, which has been fitted by the Committee
of the National Physical Laboratory as a temporary laboratory.

It is interesting to note that the case containing the original
coils of the Association bears the words, " To be deposited at Kew."
After many wanderings the coils have at last returned to their
home.

The Sub-Committee on Platinum Thermometry held a meeting
in the spring, and agreed to the following resolutions :—

(i) That a particular sample of platinum wire be selected,
and platinum thermometers be constructed therefrom to serve as
standards for the measurement of high temperature.

(ii) That Mr Glazebrook and Professor Callendar be requested
to consider the details of the selection of wires and construction
of thermometers for the above purpose, and to consult with
Mr Matthey, who kindly consented to give his assistance.

Since then Mr Matthey has supplied the Sub-Committee with
two specimens of very pure platinum. Portions of these have been
made into thermometers and tested at the National Physical
Laboratory, with the following results, iio being the resistance at
0° and -Bioo at 100°, while S is the coefficient occurring in Callendar's
difference formula : —

Wirel . . 1-3883 . . 1493
„ 2 . . 1-3884 . . 1-498

The question of the selection of a wire for the construction of
the standards is still under the consideration of the Committee.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 603

During the summer a very full comparison has been made of the
unit resistance coils of the Association, and the opportunity has been
taken of comparing these with some coils belonging to the Board of
Trade, and with others which have recently been obtained fix)m the
Reichsanstalt. The coils were also compared with one of the mercury
resistance tubes prepared by M. Benoit in 1885, and which has been
in the care of the Secretary since that date.

The results have not yet been completely worked out, and
publication is, therefore, necessarily deferred. Moreover, the
temperature during July was very high, so that the mean
temperature of the observations is much above that at which
previous comparisons have been made. For the purpose, there-
fore, of connecting these results with the past it will be desirable
to make some further observations in the autumn.

It seemed desirable to set up some mercury resistance tubes in
England, with a view of keeping a check on the variations of the
wire standards.

Preparations have been made for this. A number of selected
tubes of '' verre dur " have been obtained, with the kind assistance
of the officials of the Bureau International, from M. Baudin, while
other tubes of Jena glass have been procured from Schott & Co.
Steps are being taken to have some of the best of these calibrated.

Some advance has been made during the year with the con-
struction of the Ampfere Balance. The Committee greatly regret
the serious illness of Prof. J. V. Jones, which has prevented more
rapid progress. The stand for raising and lowering the outer coils
has been completed. Thanks to the generosity of Sir A. Noble, the
cost of this, estimated at about £100, has been saved the Committee.

During the spring the Secretary, as Director of the National
Physical Laboratory, visited the Bureau International at Paris and
the Reichsanstalt at Berlin. The Committee are glad to put on
record their appreciation of the great courtesy and kindness with
which he was received by President Eohlrausch, M. Benoit, and
the other officials connected with those institutions.

The Committee are informed that at the recent International
Electrical Congress at Paris the two following resolutions were
unanimously adopted by Section I., and confirmed by the Congress
and by the Chamber of Qovemment Delegates : —

1. The Section recommends the adoption of the name of Qauss
for the C.O.S. unit of magnetic field.

604 PRACTICAL STANDARDS

2. The Section recommends the adoption of the name of
Maxwell for the CG.s. unit of magnetic flux.

The question of giving names to the units of magnetic force
and flux has been before the Committee on several occasions. The
Committee therefore were in a position to welcome cordially these
resolutions, and at their last meeting agreed unanimously to a
resolution adopting the two names selected by the Paris Congress.

Of the sum of £25 voted last year, £13. 7s. Id, has been
expended on material for the new platinum thermometers and on
the transport of the apparatus from Liverpool to Richmond. If
the plan of constructing standards for platinum thermometers is
adopted, it will be necessary to purchase a large stock of suitable
wire, the w^hole of which should be made at the same time. For
this a considerable expenditure will be required ; there will also be
incidental expenses connected with the making and standardising
of the thermometers. For these purposes the Committee ask for
a grant of £75.

The Committee therefore recommend that they be reappointed,
with a grant of £75, and that Lord Rayleigh be Chairman and
Mr R. T. Glazebrook Secretary.

Appendix.

Note an an Improved Standard Resistance Coil,
By Robert S. Whipple.

The coil in question consists of a bare wire wound on a mica
frame.

The form of coil possesses the following advantage over the
ordinary resistance coil : — (1) The coils can be. annealed to a dull
red heat in situ, thus relieving the wire of any strain caused by
the winding. (2) The heating of a wire immersed in oil is less
than one silk-covered and varnished. (3) The temperature of the
wire can be accurately determined by means of a thermometer
placed in the oil surrounding the wire. German physicists have
adopted a form of coil in which the wire is silk-covered and
varnished and then placed in a metal case perforated with holes.
The whole coil is placed in an oil bath when in use. This form of
coil is open to the objection that it cannot be annealed above

FOR ELECTRICAL MEASUREMENTS

605

140° C. without causing injury to the silk covering on the wire,
and there is a certain amount of lag in the oil obtaining the
temperature of the coil.

By request of the Electrical Standards Department of the
Board of Trade the Scientific Instrument Co., Cambridge, have
designed and made two standard 1-ohm coils the wires of which
are bare and immersed in oil; a modification suggested by Mr
Horace Darwin was also fitted for obtaining the temperature of
the coils. The coils proper consist of 0'036 in. PtAg wire wound
on mica frames, the ends of the wires being attached to stout
copper terminals in the usual manner. A 0*08 in. platinum wire
is wound alternately with the platinum-silver wire, and is attached
similarly to stout copper leads. Both coils are adjusted to a resist-
ance of 1 ohm at 16'5° C. Owing to the difference in the tempera-
ture coefficient of the two wires (PtAg 0*00024, Pt 0*00350), a
small change in the temperature of the coil causes a comparatively
large difference between the resistances of the two coik. This
difierence being known, the temperatures in degrees Centigrade
are given by the adjoined table. The table is calculated from the
difference in the temperature coefficients of the two wires

000360 - 000024 = 000326
for 1" C.

Temperature of

Difference in resistance

gtandard coil of the coils

10-0'C -0-01793v

110*C.

0-01467

Plat

11-9*' C. .

0-01141

13-0" C. .

000815

■

14-0" C. .

0-00489

15-0" C. .

... -0-00163/

15-6°a ,

0-00000

16-0" C.

... +0-00163\ p, .
0-00489 ^^^
0-00815 •

IVO'^C.

18-0" C.

190' C.

0-01141

20-0" C. .

... +0-01467/

Platinimi coil having a
lower resistance
than the platinum-
silver cou.

Platinum coil having a
higher resistance
than the platinum-
silver cou.

As the temperature coefficient of platinum is about fifteen
times as great as that of platinum-silver, the resistance of this coil
may be measured to one significant figure less than the standard
coil without affecting the value for the temperature of this coil.
In measuring small resistances the determination of the last figure
to O'OOOOl ohm requires considerable care, and the advantage of

606 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

not being compelled to measure to such a high degree of accuracy
is apparent. The two wires being wound on the same frame
•alternately with each other and immersed in oil are at the same
mean temperature. Any temperature gradient in the oil influences
both wires similarly, thus doing away with the necessity of a
stirrer. The platinum wire is also useful for testing the insulation
of the windings of the PtAg coil one from the other. The coils
are placed in a glass vessel in order that the behaviour of the
insulating oil with time may be studied.

TWENTY-EIGHTH REPORT— GLASGOW, 1901.

Appendix. — Note on a Comparison of the Silver deposited in Voltameters

containing different Solvents. By S. Skinner . . . .p. 608

During the year a number of comparisons have been made at
the Eew Observatory among the standard coils of the Association.
The temperature conditions, however, in the temporary laboratory
are not sufficiently satisfactory to make it desirable to report fully
on the results ; it is perhaps sufficient to say that no evidence of
any very marked change in the relative values has shown itself.
It is hoped that the coils and other apparatus will be moved to
Bushy House during the autumn.

In the room which has been planned for their reception
arrangements will be at hand for controlling the temperature, and
the work of inter-comparison and control of the standards can go
on as in former years at Cambridge.

Meanwhile some progress has been made in the preparations
for the construction of mercury standards. A number of tubes of
*' verre dur " have been examined, and some of these have been
calibrated ; when the apparatus is set up at Bushy House this work
will go forward rapidly. There has also been during the year some
demand for the issue of standards of capacity : this it has not been
possible to comply with, but the air condensers will be set up
again as soon as possible, and then capacity tests can be made.

With regard to platinum thermometry, Mr Matthey supplied
the Committee with a further specimen of wire, for which he had
made a large stock. This was tested carefully, both at Kew and
under Mr Griffiths's directions, by Mr Green at Cambridge, and the
values for the constants were as under : —

S = 1-496 ± -005.

The wire has proved in every way satisfactory, and the money
voted to this Committee last year (£45) has been spent in pur-
chasing it.

608 PRACTICAL STANDARDS

Mr Matthey, however, is retaiDing for the present, for the use
of the Committee, some more of the wire, and it is, in their
opinion, desirable that they should purchase it also. It is essential
for the success of the scheme approved by the Committee at their
last meeting that they should have a sufficient stock of the wire for
a very long period, and they are anxious not to lose the present
opportunity of acquiring such a stock.

Expense will also be incurred in the preparation of the mercury
standards.

The illness and death during the year of Professor J. Viriamu
Jones have prevented any great progress being made with the
ampere balance. Some part of the apparatus, however, has been
constructed, and is in Professor Ayrton's hands, and the Committee
have good hopes that further progress may be reported shortly.

The Committee desire to put on record ^their sense of the loss
which Physical Science has suffered by the deaths of Professors
J. V. Jones and G. F. FitzGerald, who for many years had been
members of the Committee, and had contributed in a marked
degree to its work ; and by that of Professor Rowland, whose re-
determination of the absolute value of the B.A. unit was practically
the starting-point of the work of the present Committee. Pro-
fessor Bx)wland had on more than one occasion been a valued
visitor at meetings of the Committee.

A paper by Mr Skinner on a pyridine voltameter is printed as
an appendix. Professor Callendar's paper on the variation of the
specific heat of water is closely connected with the work of the
Committee.

In conclusion, the Committee recommend that they be re-
appointed, with a grant of £50 ; that Lord Rayleigh be Chairman,
and Mr B. T. Glazebrook Secretary.

Appendix.

Note on a Comparison of the Silver deposited in Voltameters
cafUaining different Solvents. By S. Skinner, M.A., Demonstrator
of Experimental Physics, Cambridge.

In 1892 Schuster and Crossley* showed that when the same
current is passed through two silver voltameters containing silver
nitrate in aqueous solution, one voltameter in a vacuum and the

* Proc, R, S. L. p. 844.

FOR ELECTRICAL MEASUREMENTS 609

other in air, about 0*1 per cent, more silver was deposited in the
vacuum than in air. This result was confirmed by Myers*. These
results clearly prove that there is an uncertainty in the action of
the silver voltameter depending on the presence of air or oxygen,
and consequently on the freshness of the solution. Wemer+ found
that a silver nitrate solution in pyridine gives by the rise in the
boiling-point of the solvent a nearly normal molecular weight for
the salt ; and Kahlenberg^ found that the solution was an electro-
lyte, and could be used in the silver voltameter ; but that, contrary
to what follows, more silver was deposited firom aqueous solution
than fix^m pyridine solution by the same current. In the following
experiments a comparison has been made of the deposits produced
by the same current in silver voltameters containing aqueous and
pyridine solutions of silver nitrate.

The platinum bowls used are those numbered I. and V. in the
paper on the Measurement of the Electromotive Force of the Clark
Cell § by Mr Glazebrook and myself. The anode for bowl I. was a
silver disc, 5 cm. in diameter, hung by a silver rod, and a silver
cylinder was used for bowl V. The dimensions of the bowls are
given in the paper mentioned above. 100 c.c. of solution was used
in each case, and the pyridine solution contained 10 per cent, of
silver nitrate, whilst the aqueous solution contained, as usual,
15 per cent, of the salt.

The areas of the exposed surfaces were approximately as

follows : —

Bowl I. Bowl V.

Cathode surface ... ... 76 sq. cm. 67sq. cm.

Anode surface 19*6 sq. cm. ISsq. cm.

The conditions of current density in the two bowls may be
regarded as practically identical.

The deposit of silver from the aqueous solution was crjrstalline,
and the character of the crystals appeared to vary with the current
density. The deposit was washed by standing in distilled water
for several hours and dried over an alcohol flame. The deposit
from the pyridine solution is continuous, and forms a hard coating :
it is washed with water in which both pyridine and silver nitrate
are soluble. It is sometimes slightly coloured, but on drying
becomes white. On further heating over the alcohol flame it

* Annalen, 66, p. 288. f Zeits, Anorg. Chem, 1897, 16, p. 23.

X Jowm. Physical Chem. 1900, p. 849. § PhiL Tratu. 1892, A.

B. A. 39

}

610

PRACTICAL STANDARDS

develops a pearly lustre, and in this condition it has been
weighed.

A Weston ampere ojeter was included in the same circuit, and
served to indicate the constancy of the current. The reading of
the amp^ meter is given in the second column of the table. The
variations of the current were very small. In the table the result
of every experiment which I have made is given.

Date

Carrent

Weston
Meter

Weight de-
posited from
Pyridine
Solution

Weight de-
posited from
Aqaeous
Solution

Difference

Milli-

Per-
oentage

Dif-
ference

1
Notes

Aug. 15
„ 16
» 14

» 21

„ 10
„ 19
„ 20

0-07

0-076

013

0-263

0*265

0-368

0-375

0-416

0-52

lOO

•8115

•8696
1-2665

•7866
2-2796
1-1390

•9630
14226
2-0010
2-0180

•8106

-8686
1-2626

•7820
2-2730
11340

•9600
1-4200
1-9982
2-0165

1-0
1-0
4-0
4-6
6-6
5-0
3-0
2-6
2-8
2-6

•124

•116

•318

•576

•30

-44

•41

•276

•14

•12

(a)
(ft)

TotAl deposits

13-5570

13-5242

32-8

•24

(a) and (6). — In these two experiments the aqueous solution was in a partial
vacuum (8 cm. pressure), and •! per cent, has been added to the percentage
difference to make them comparable with the other experiments.

(e). — Fresh solutions were used in this experiment, and the same solutions
were used on all subsequent dates. A few particles of silver were lost from the
aqueous voltameter in this experiment, August 14.

The first result of these experiments is clearly that all the
deposits firom the pyridine solutions weigh more than those from
the aqueous solutions.

In the measurements of the E.M.F. of the Clark cell by Mr
Glazebrook and myself the same current was sent through two
systems of silver voltameters in series, and 15*5123 grammes were
deposited in the bowls which received the greater deposits, as
against 15*5065 grammes in those which gained the smaller
deposits. This gives a mean percentage difference of •044, which
may be compared with the mean percentage difference of '24 in
the present experiments. It is obvious that this difference is of
a much higher order, but this difference is a mean of experiments

FOR ELECTRICAL MEASUREMENTS 611

which differ much more between themselves. On that account
I think it is better to discuss the experiments in groups. The
experiments divide themselves roughly into two groups. There is,
first, a group consisting of those in which the current was about
*07 ampere and from '5 to 1 ampere. This contains the extremes
as regards current, and in it the mean percentage difference would
be just over '1 per cent. So that for these values of current the
deposit firom pyridine would weigh almost the same as Schuster
and Crossley found for a vacuum, which, it will be remembered,
was '1 per cent, higher than in air.

The second group consists of those experiments in which the
current value lies between *13 and '41 ampere, and here the mean
percentage difference is much larger, %.e. '38. Over this range
one of the deposits seems to be uncertain, and I think these
experiments may be considered to indicate that between these
values of current in the given bowls one of the two voltameters
is irregular in its action. The character of the silver crystals
appeared to be variable, whilst the hard film of silver firom the
pyridine solution had always the same texture. The aqueous
voltameter seemed to work best with the large currents *5 to 1
ampere when the crystals were small, hard, and closely packed.
At the lower values of current the silver crystals were thin, long,
and friable. At the lowest value they were again small and hard.
One explanation of the variation may be that particles of silver are
more easily lost during the washing, when the crystals are of the
second character.

Conclusions : —

(1) That Faraday's law holds to within *24 per cent, in the
mean for silver nitrate when dissolved in two different solvents.

(2) That for current values of '07 and -5 to 1 ampfere in
the given bowls the amount of silver deposited fix)m a pyridine
solution of silver nitrate is nearly the same as that deposited from
an aqueous solution in a vacuum.

(3) That for current values between '1 and '5 ampere more
silver is obtained in the pyridine voltameter than in the aqueous
voltameter.

39—2

TWENTY-NINTH REPOKT— BELFAST, 1902.

Appendix.— Ow the Defimtian of the Unit of Heat , . p. 615

During the past year the apparatus belonging to the Association
has been removed to and set up at the National Phjrsical Labora-
tory at Bushy House. A room in the basement has been fitted
for accurate resistance work. By means of a thermostat the
temperature can be kept under very complete control, and the
room has proved very suitable for its purpose. In it the resistance
standards of the Association have been set up, and a number
of comparisons have been made by Mr F. E. Smith. Particulars
as to the results of these comparisons can best be given at a later
date, when the mercury standards now in course of construction
have been set up.

The work of setting up these mercury standards of resistance
has been further advanced. A number of tubes, both of verre dur
and of Jena glass, 16'", have been calibrated by Mr Smith. When
the final corrections to the weights used have been obtained from
the Bureau International it will be possible to complete these and
to determine the values of the platinum-silver and manganin
standards in terms of the mercury unit.

From the resistance-room a cellar — formerly the wine-cellar of
Bushy House— opens, and in it work requiring an extreme
constancy of temperature can be carried on.

In this room Mr Smith has set up a number both of Clark
and also of Weston cells, and comparisons between these have
been carried on systematically.

Discrepancies of a considerable amount have been found
between cells set up in the same manner, but from materials
supplied by various makers, and these have been traced to the
mercurous sulphate. The observers at the Reichsanstalt have
come to a similar conclusion*. Dr Carpenter and Mr Smith are

♦ Th&tigkeit der Phys,'Tech,-Reich$anMtalt, 1901-1902.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 613

now engaged in experiments at the National Physical Laboratory,
the results of which, it is hoped, will enable them to specify a
method of preparing mercurous sulphate which will lead to
consistent results for the E.M.F. of the cells.

The air-condensers belonging to the Committee have been set
up, and a number of determinations of their capacity have been
made by Mr Campbell. The results of these, though at present
they are only to be treated as provisional, show that the condensers
are in good order, and have suffered no damage by their journeys
first to Liverpool and then to Richmond. The capacity of one
is nearly the same as when at Cambridge; that of the other
has altered very slightly. With a view of establishing a
standard of capacity a number of other comparisons between
the standards of the Association and those of Dr Muirhead are
in progress.

In *thi8 comparison work some difficulty has arisen from the
fact that all the resistance boxes belonging to* the Association
are of platinum-silver. The small temperature coefficient of
manganin gives that material a very distinct advantage, and the
Secretary has been endeavouring to use it whenever possible.
It would be of great service for this part of the work to have
a subdivided megohm box in manganin, and the Committee
trust that funds for this may be forthcoming. They hope in
their next report to give a detailed account of the condenser
experiments.

The construction of platinum thermometers as standards for
high temperature thermometry has made some progress. The
National Physical Laboratory was not opened until March, and
the work of setting up the apparatus, carrying out the necessary
calibrations, etc., has occupied most of the time of the assistants
since then.

After some further experiments, however, to test the purity of
the wire it was proposed to use had been carried out, a stock
of eight ounces of wire of the highest purity and of a thickness
varying from six to eight mils has been bought from Messrs
Johnson & Matthey, while four ounces of the same wire, but of
twenty-two mils in thickness, suitable for leads or for drawing
down to special sizes, have also been secured; and six thermometers
are in course of construction in the workshops of the Laboratory
under Dr Harker's supervision.

614 PRACTICAL STANDARDS

Of these* six thermometers two of five ohms fundamiental
interval will be hermetically sealed in glass tubes, and will serve
as standards for low-temperature work ; a second pair, having an
interval of one ohm, in tubes of hard glass — ^probably Jena —
59'" — will serve for temperatures up to 550° C, while the third
pair, also of one ohm interval, in porcelain tubes, will be employed
up to llOO"* or 1200** C. It is hoped by the use of quartz to
extend the range of temperature considerably, and some experi-
ments are in progress with this object.

Two electrical resistance ovens have been built by Dr Harker
for high-temperature work, and these serve their purpose
admirably.

The grants voted during the past two years have been
expended on the purchase of the materials for the platinum
thermometers, and additional sums are necessary to complete
their manufacture.

With regard to the construction of the ampere balance the
Committee are sorry that they cannot report progress ; they have
learnt with extreme regret of Professor Ajrrton's ill-health during
part of the year, but are glad to know that he believes he will
be able to continue his investigations into this important
question, and they have therefore reason to hope the matter will
be advanced.

In this work the late Principal Viriamu Jones was closely
associated with Professor Ayrton, and it is a source of great
pleasure to the Committee to know that, through the generosity
of the Drapers' Company, his name and connexion with Electrical
Measurements will be perpetuated at the National Physical
Laboratory. The Company had promised to Principal Jones the
funds for the construction of an improved Lorenz apparatus for
the determination of the ohm, and they have intimated to the
Committee of the Laboratory their intention to place £700 at the
Executive Committee's disposal for the construction of such an
apparatus in his memory under the superintendence of Professor
Ayrton and the Director of the National Physical Laboratory.
The Secretary states that steps have already been taken to obtain
designs for the instrument.

At the Meeting in Belfast Sir William Preece drew the
attention of the Committee to the work of the Standardisation
Committee of the Engineering Societies, and expressed the hope

FOR ELECTRICAL MEASUREMENTS 615

that in his capacity as chairman of the electrical branch of that
committee he might have the assistance of the Electrical Standards
Committee. The Secretary was instructed to afford all the
assistance in his power.

Reference was also made to the definition of the unit of heat,
and the Secretary was requested, with the assistance of Mr Griffiths,
to draw up an Appendix to the Report dealing with this. The
Committee expressed the strong hope that any unit of heat
formally accepted by engineers should be based on the C.O.S.
system of units.

In conclusion the Committee recommend that they be re-
appointed, with a grant of £75, to be used for the establishment
of a standard of capacity and for the construction of standard
platinum thermometers; that Lord Rayleigh be Chairman and
Mr R. T. Glazebrook Secretary.

Appendix.
On tiie Definition of the Unit of Heat,

The question of the definition of the unit of heat has been
before the Committee on various occasions.

In 1896, at the Liverpool Meeting, after an exhaustive
discussion and the consideration of letters from scientific men in
all parts of the world, the following propositions were provisionally
approved : —

Proposition I, — For many purposes heat is most conveniently
measured in units of energy, and the theoretical c.G.8. unit of
heat is one erg. The name ''joule" has been given by the
Electrical Standards Committee to 1(K ergs.

For many practical purposes heat will continue to be measured
in terms of the heat required to raise a measured mass of water
through a definite range of temperature.

If the mass of water be one gramme, and the range of
temperature 1**C. of the hydrogen thermometer from 9*5** C. to
10*5^ C. of the scale of that thermometer, then, according to the
best of the existing determinations, the amount of heat required
is 4*2 joules.

It will therefore be convenient to fix upon this number of
joules as a secondary unit of heat.

616 PRACTICAL STANDARDS

This secondary thermal unit may be called a ** calorie."

For the present a second proposition is —

Proposition II. — The amount of heat required to raise the
temperature of one gramme of water 1° C. of the scale of the
hydrogen thermometer at a mean temperature which may be
taken as 10** C. of that thermometer is 4*2 joules.

If further research should show that the statement in IE. is
not exact, the definition could be adjusted by a small alteration
in the mean temperature at which the rise of V takes place. The
definition in I. and the number (4*2) of joules in a calorie would
remain unaltered.

These propositions, it will be observed, while reaffirming the
names "joule" as the equivalent of 10' ergs, and calorie, the
equivalent of 4*2 joules, as the amount of heat required to raise
the temperature of one gramme of water one degree centigrade
on the hydrogen scale, leave undetermined the mean temperature
of the water so raised. Proposition II. states that this may be
taken as 10"" C, but it is pointed out that if the heat required to
raise one gramme of water firom 9*5° to 10*5° C. should prove not
to be 4*2 joules a readjustment in the mean temperature employed
in the definition could easily be made.

Accordingly in the Report, 1897, made at Toronto the Com-
mittee wrote : —

*'At the Liverpool Meeting the Committee agreed that the
* calorie,' defined as the heat equivalent of 4*2 x lO' ergs, should be
adopted as the unit for the measurement of quantities of heat,
but the question as to the exact part of the absolute thermo-
dynamic scale of temperature at which this quantity of heat could
be taken as equal to one water-gramme-degree was for the time
being left open.

"This resolution has made it incumbent on the Committee to
consider carefully —

"1. The relation between the results of measurements of
intervals of temperature by accepted methods and the absolute
scale.

" 2. The specific heat of water in terms of the erg and its
variation with temperature.

" With regard to the first point there appears to be no reason
to doubt that the scale of a constant-volume hydrogen thermo-
meter is very nearly identical with the absolute scale. The

FOR ELECTRICAL MEASUREMENTS 617

Committee have therefore decided to recognise the standard
hydrogen thermometer of the Bureau International des Poids
et Mesures as representing, nearly enough for present purposes,
the absolute scale. This convention has at least the advantage
of giving a definite meaning to statements of the numerical
value of intervals of temperature within any range for which
comparison with the hydrogen thermometer is practicable. If
future investigation should show that it is inaccurate to any
appreciable extent, corresponding corrections can be applied when
necessary."

As regards the second point further research has shown that
an alteration in the temperature of measurement is required.
The present position has been summed up by Principal Griffiths
in the Rapports prisentia an Congris Interfiational de Physique,
Paris, 1900, tome i., and in his Lectures on the Thermal
Measurement of Energy*. They are also summarised by Pro-
fessor Everett in the latest edition of his work, C,0,S, Units and
Constants.

The following table (p. 618), taken fix^m Professor Everett's
work, gives the results adopted by Principal Griffiths.

From this it follows that the heat required to raise a gramme
of water 1** on the hydrogen scale is 4*2 joules when the range of
temperature is from 7'2** C. to 8*2*' C. Thus according to this the
10*" C. of Proposition II. should be T'T** C, and a calorie would be
the heat required to raise a gramme of water 1° of the hydrogen
scale from 7-2'* to 8-2° of that scale.

The results of a series of observations on the heat required
to raise a gramme of water from 0° C. to 100"* C. were published
by Re3aiolds and Moorby in 1897. The quantity necessary is
proved to be 4*184 joules. Thus the mean heat required to raise
a gramme 1°C. for temperatures between 0** and 100° is 4*184
joules. This number is not far from the 4*2 joules adopted in
1896 as the number of heat units in a calorie. Accordingly the
suggestion has been made that it would be convenient to change
the definition of a calorie and take it to be one-hundredth part of
the heat required to raise one gramme of water from 0° C. to
100° C. In this case, according to the mean number adopted
by Griffiths (see table), one calorie would be equal to 4*1854 joules,
while according to Reynolds and Moorby it would be 4*184 joules,

* Cambridge University Press, 1901.

618

PRACTICAL STANDARDS

while the degree centigrade through which it would be necessary
to raise one gramme of water in order to absorb an amount of heat
equal to one calorie would be from 17° to 18° C.

Tables of Mechanical Equivalents in Joules.

1 joule =10^ ergs.
Rowland, reduced by Day. Hydrogen scale.

5°

[4-206] 1

13°

4-191

• 21°

4-180

29°

4-174

6°

4-203

14°

4189

, 22°

4179

30°

4-174

4-201

16°

4-188

1 23°

4-178

31°

4-174

8°

4199

16°

4-186

24°

4177

32°

4-174

9°

4198

17°

4-185

' 26°

4-176

33°

4-174

10°

4196

18°

4-184

26°

4-176

34°

4-174

11°

4*194

19°

4-182

27°

4-175

35°

4-175

12°

4192

20°

4-181

28°

4-175

36°

4-175

0°
6°
10°
15°
20°
25°
30°
36°
40°
45°
50°
65°
60°
65°
70°
75°
80°
85°
90°
95°
100°

Day
Hydrogen

[4-205]
4-196
4-188
4-181
4-176
4-174
4-175

Barnes
Air

4-210
4-198
4-189
4-184
4-180

4
4
4
4
4
4
4

178
177
177
178
180
182
184
4-187
4-190
4-192
4-195
4-198
4-201
4-204

Barnes
Hydrogen

4-213
4-200
4191
4-185
4-180
4-178
4-177
4-177
4-178
4-178
4-181
4-183
4-185
4-188
4191
4-195
4-198
4-201
4-205

Mean, giving half-weight to 0° and 100*

Griffiths
adopted

[4-219]
4-206
4*195
4-187
4-181
4-176
4-174
4-173
4-173
4-173
4*174
4-176
4-178
4-181
4-184
4-187
4-190
4-193
4-197
4-201

[4-205]

4*1864

If this view were taken then instead of Proposition II. of 1896
we should read : —

FOR ELECTRICAL MEASUREMENTS 619

One calorie is the amount of heat required to raise the
temperature of a gramme of water from 17'' C. to 18° C. on
the scale of the hydrogen thermometer, and is equal to 4*184
joules.

It should be noted, however, that the Committee have not as
yet taken any resolution on the point, and that formally the
propositions accepted in 1896 and reprinted above are those
which they have approved.

It should also be mentioned that in deference to international
representations the use of the word ** therm '* was withdrawn in
1896, the name being replaced by the word "calorie."

THIRTIETH REPORT— SOUTHPORT, 1903.

APPENDIX PAOB

I. On the Values of the Renstance of certain Standard CoUe of

the British Association, By F. E. Smith. {From, the
Ifatumal Physical Laboratory) 627

II. On some neio Mercury Standards of Resistance. By F. £. Smith.

{From the National Physical Laboratory) • . . « 636

III. On the Platinum Thermometers of the British Association. By

J. A. Habksr, D.Sc. {From the National Physical La-
horatory) 638

IV. Tahle of the Resistance found for Pure Annealed Copper • 646

During the year a very complete comparison of the resifltance
standards belonging to the Association has been carried out, and
the standards have been compared with those of the Reichsanstalt
and of the Board of Trade.

The various units discussed in the Report are : (1) The ** ohm,"
10® C.G.S. units of resistance ; (2) the international ohm — viz., the
resistance at O^'C. of a column of mercury of uniform section
106'3 cm. in length and 144521 grammes in mass; (3) the
original B. A. unit ; (4) the Board of Trade unit, supposed to
represent the international ohm, but constructed in 1891 so as
to be equal to 1*01358 B. A. units ; (5) the N.P.L. unit defined
as No. 4, as deduced from the wire standards of the Association ;
(6) the Reichsanstalt unit, constructed at the Reichsanstalt to
represent the international ohm; (7) the mercury tubes, con-
structed at the National Physical Laboratory to represent the
international ohm.

A full account of this comparison is given in Appendix I. to
the Report, by Mr F. E. Smith, of the National Physical Laboratory.
It appears from this that changes have shown themselves in all
the original platinum-silver coils. The relative values of these
coils are discussed in the Reports of the Committee for 1888,
1890, and 1892. The 1888 Report contains a very complete

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 621

•

comparison of all the coils, not merely those of platinum-silver ;
and it is there shown that they then agreed with the values
assigned to them by Fleming in 1881. The conclusion is also
drawn in the same Report that, with the exception of the platinum-
iiidium coils A and B, no really certain variations could be traced
in the other coils between the results of Matthiessen and Hockins's
comparisons in 1864 and 1867, those of Chrystal in 1876, Fleming
in 1881, and the present Secretary in 1888*. A postscript to the
Report for 1888 recorded, however, an appreciable change in the
coil F in the autumn of that year.

In Appendix I. Mr Smith starts with the values given in the
1888 Report, which are, as nearly as we can tell, the original values
of the coil&

Changes in the three standards F, (?, H have already been
recorded in previous Reports (1890 and 1892). The standard coil
Flat remained unchanged in value until 1901 — 1902. Between
the observations recorded in these years it increased in resistance
by 17 X 10~* B. A. U., and has not varied since.

The alterations in the other coils since the comparisons in 1888
have been as follows : —

^+97xlO-»B.A.U.

(? + 33xl0-» .,

jy+18xl0-» „

It should however be noted that, while between 1888 and
1890 the change in F was +64x 10-» B. AU., that in (? was
- 27 X 10-», and in iT - 13 X lO"'. Since 1890 the same coils
changed by + 33 x lO"', + 54 x lO"', and + 31 x 10"* B. A U.
respectively, while between 1901 and 1902 Flat, as has already
been stated, rose by 17 x 10~' B. A. U.

It is not easy to trace the causes of these changes. In the
case of Flat the observations in 1901 were made at Eew, those
in 1902 at Bushy House, and the change may in some way be
connected with the removal of the coils. The changes in F^ 6, H
first showed themselves after the coils had been subject to a
very low temperature, and may have been started by strains due
to this.

Appendix I. gives the details on which these various state-
ments are based. It appears also from the same Appendix that

* It is possible that coil i^ is an exception to this statement.

622 PRACTICAL STANDARDS

the new platinum-silver ohm standards of the Association have
retained their values since 1898 practically unchanged

The comparison between the standards of the Association and
those of the Reichsanstalt leads to the result that the unit of the
Association (No. 5 of those defined above) is less than that of
the Reichsanstalt (No. 6) by '000105 ohm. This result is deduced
(Table IX. of Appendix I.) fix)m a series of extremely concordant
me€U3ures on coils of value 0*1, 1, 10, 100, 1000, and 10,000 ohms ;
thus both the unit and the multiple coils agree in giving the same
difference between the Reichsanstalt and ourselves.

By the kindness of Mr Trotter a comparison has been made
between the Board of Trade unit and those of the Association,
with the result that, as deduced fix>m the unit coils, the Board of
Trade unit is less than that of the Association by '00006 ohm.
This result, however, is not confirmed by a comparison of a
1000-ohm coil belonging to the Association with one of those of
the Board of Trade* ; these coils show no differenca

The above statements are made on the assumption that the
various changes in the coils which have undoubtedly occurred
have been rightly interpreted, so that we can now recover the
absolute C.G.S. value of the coil Flat, and hence of the standard
ohm as originally determined at the Cavendish Laboratory, cmd
defined by the Committee in the Edinburgh Report, 1892.

That this is the case is borne out by the results of the experi-
ments on the specific resistance of mercury, a summcuy of which
is given in Appendix II. These are not yet complete. Mr Smith
has, however, constructed and calibrated eleven mercury tubea
The mean cross-section of each of these has been determined by
at least four different sets of measurements. In nine cases the
greatest difference between any measurement and the mean is not
more than '001 per cent.

The values found for the resistance of each tube do not differ
by more than '001 per cent.

If we assume as above that the values of the wire standards of
resistance of the Association are known in terms of the absolute
C.G.S. unit, then it follows that the length of the column of
mercury, one square millimetre in section, which would have a
resistance of 10* C.G.S. units, would be 106*291 centimetres.

* If the view be accepted that the laboratory unit is the same as in 1891, the
Board of Trade standard has fallen since that date by '00006 ohm.

FOR ELECTRICAL MEASUREICENTS 623

The value found for this same quantity by the Secretary
(Mr Glazebrook) and Mr Fitzpatrick in 1888*, was 106-29 centi-
metres. We infer then that we still can recover from our standard
coils the absolute CG.s. unit of resistance.

Again, the length of the mercury column constituting the
international ohm has been defined as 106*3 cm.

But we have seen that the absolute CG.s. unit as deduced
from the wire coils of the Association has a resistance equal to
that of 106*291 cm. Thus the absolute unitf is smaller than the
international ohm by '009 per cent. Again, it has been stated
above that the unit deduced from the standards of the Association
is smaller than that of the Reichsanstalt by OlOo per cent.

Thus the mercury standards of the Reichsanstalt, constructed
to represent the international ohm, exceed those just made for
the Association by Mr Smith by *001e per cent., or 1*5 parts
in 100,000.

Again, if these results be accepted, since the Board of Trade
unit, as derived from the wire standards, is less than that of the
Association by *006 per cent., and the Association unit is too
small by "009 per cent., it follows that the Board of Trade unit
is too small by *015 per cent. This difference arises in part from
the fact that the standards of the Association, from which the
Board of Trade standard was copied by the Secretary in 1891, are
too low ; in part from the fact that the Board of Trade standard
has diverged slightly from that of the Association since 1891.

Thus, to sum up this part of the Report, it may be stated
that : —

(a) The original B. A. unit and the standard ohm based on it
(Nos. 3 and 5 of the units concerned) can be recovered frx>m the
wire coils of the Association.

(6) The Board of Trade unit (No. 4) is now less than the
Laboratory unit (No. 5) by '006 per cent.

• Phil. Tram. 1S8S.

t The resiBtanoe taken for a oolnmn of meroary 1 square mm. in Beotion,
100 cm. in length at O^C. at the Edinburgh Meeting in 1892, wae

•9407 X 10* O.O.B. units.

Mr Smith's experiments give, asBuming the yalues of the wire coils known, the
result -9406 x 10* o.o.s. units.

624 PRACTICAL STANDARDS

(d) The Board of Trade unit is less than the international
ohm by '015 per cent.

(e) The mercury tubes made at the National Physical
Laboratory to represent the international ohm are less than
those made at the Reichsanstalt by *0015 per cent.

This last result must be considered as provisional pending
the completion of Mr Smith's work, but it is clearly highly
satisfieu^tory.

Mr Smith has also made progress during the year with his
investigations into certain of the anomalies shown by Clark cells,
but the results of that inquiry are not yet ready for publication.

The standard condensers of the Association have been
frequently in use during the year; about fifteen condensers
have been compared with them. They retain their value in a
satisfactory manner, and are convenient to work with, though
possibly some improvement in the insulation might be desirable.

A chronograph, purchased with part of the grant made last
year, will enable the time measurements requii*ed in the measure-
ment of capacity to be made with greater accuracy, and hence will
permit of greater rigidity in the inquiry as to the permanence of
the standards.

The platinum thermometers made fix)m the stock of wire
purchased from Messrs Johnson and Matthey, which at the time
of the last Report were in course of construction, have been
completed, and the behaviour of some of them investigated
throughout the past year. The resistance-box available was the
old Callendar-Griffiths box used in the work of Dr Chree at Kew
Observatory, having coils of platinum-silver on the binary system.
The contacts are an old form of the Cambridge Instrument
Company's type of plug contact, the cheeks being made of a
special white alloy held in round Doulton-ware cups. In measure-
ments with this box not much significance attaches to the third
figure of decimals representing hundred-thousandths of an ohm,
though the settings could be made to this amount at the lower
temperatures. The box resistance-coils were intended for use
with platinum thermometers of 1 ohm fundamental interval only,
and therefore the two high-resistance thermometers, of 5 ohms
fundamental interval, could not be measured at the sulphur-point;
their systematic investigation has therefore been temporarily post-
poned. The want of a better box for this work is seriously felt

FOR ELECTRICAL MEASUREMENTS 625

Of the original six thermometers made in August 1902, Nos. 1
to 4 are of 1 ohm fundamental interval, Nos. 1 and 2 being in
porcelain and 3 and 4 in specially thin Jena glass tubes of internal
diameter 8 to 9 mm. and 38 to 40 cm. long. Nos. 5 and 6 are of
5 ohms fundamental interval, and in somewhat wider tubes of
specially thin glass, through which the four leads are hermetically
sealed. The heads of all these thermometers are of the design
used by Chappuis and Harker, the contacts to the solid ends of
the copper fiexibles being made by. fusible metal cups. With
reasonable care these contacts prove very satis&ctory, both as
regards the constancy of their resistance and their mechanical
strength.

In the construction of all these thermometers special care was
devoted to adjusting their fundamental intervals to be very close
to their nominal values, and after completing this adjustment all
were subjected to repeated annealing in air at a bright-red heat,
thermometers Nos. 3 and 4 being temporarily placed in porcelain
tubes for the purpose.

The remaining four constructed last summer, and one of later
date, all of 1 ohm fundamental interval, have had their constants
determined from time to time during the year. One of them —
B. A. j — was selected as a representative platinum thermometer
for use in an investigation made to determine the relation between
the platinum scale and that of the gas thermometer of the
National Physical Laboratory at temperatures up to lOOO'^C.
During" the time occupied by two sets of experiments with this
instrument, extending over about three months in all, its constants
altered by an amount only just greater than their probable error,
showing that it is quite possible to use properly constructed
platinum thermometers up to temperatures slightly over 1000° C,
for long periods without fear of serious changes.

The summary of the life-history of the different thermometers
is given in Appendix III. The chief fact apparent is that there
seems to be a small but real difference between the £ of thermo-
meters 1 and 3 on the one hand, and 2, 4, and 7 on the other, the
maximum divergence being about 0*02.

Prolonged electrical heating in air of the wire of one of the
thermometers was not found to sensibly change the value of
the S. The cause of the small differences found is not obvious,
and further investigation is being made on this point.

B. A. 40

626

PRACTICAL STANDARDS

A change in S from 1*50 to 1*51 would make at the sulphnr-
point a diflference of 0153' C, and at 1000** C. one of 0-9°.

The question of the resistance of copper has been raised lately
by the work of one of the sub-Committees of the Engineering
Standards Committee. For commercial purposes the resistance
of copper is defined at a temperature of 60'*Fahr. (15'55**C.).
A table in Appendix IV. gives the values that have been found by
various experimenters.

It is clear that copper i^ now prepared of a higher degree of
purity than in the time of Matthiessen. Taking the mean of the
figures in the table for modem electrolytic copper, we have as
the value of the resistance of 1 metre of copper wire weighing
1 gramme the value 0'1485e ohm at 15*55'' C, but the figures of
which this is a mean range from 0*1475 to 0*1492. The value found
by Matthiessen, as deduced from his paper in the Phil. Trans.
for 1860, is 0*1500 ohm. Thus the conductivity of modem pure
electrolytic copper is 1 per cent, better than Matthiessen's.

The Committee on copper conductors, which investigated the
question in 1899, adopted the number 0*1508 ohm as the
resistance of a metregramme of commercial annealed high-
conductivity copper. This figure has been accepted by the
Eng^eering Standards Committee.

Mr H. A. Taylor has recently placed in the hands of the
Secretary two resistances of gold-silver wire made by Matthiessen
himself, to represent the resistance at 15*5"' C. of 100 inches
of pure annealed copper, having the weight of 100 grains. The
resistances of these coils have been determined by Mr Smith, and
the results are given in the following table : —

Coil No. 1

Coil No. 2

Resistance of 100 inches of copper weigh->'

ing 100 grains, as given by Matthies- ...

sen in B. A . units at 15-6' C.
Resistance found in 1903 in B.A. units\

at 15-6'C. /•••
Resistance found reduced to ohms at 1 5*5° 0. . . .
Resistance deduced of a metregramme)

in ohms at IS-S^'C. J-

•1616

•15136
•14938
•1499e

•1514

•15138

•14929
•14994

Thus Matthiessen's value for the resistance of annealed copper
at 15*55'' C. (eO'^Fahr.), as deduced from these coils, agrees very

FOR ELECTBICAL MEASUREMENTS 627

closely with the value calculated by the Secretary from the figures
in his 1860 paper.

The Committee have had under consideration the drawings
and specifications for the Ampere Balance as designed by the late
Principal Viriamu Jones and Professor Ayrton. The electrical
parts of the instrument need construction under skilled super-
vision. Tests of various kinds have to be made continually, and
the Committee have come to the conclusion that this supervision
can best be secured by having the instrument constructed in the
workshop of the National Physical Laboratory, under the care of
Professor Ayrton and the Secretary, who, as Director, will be able,
with the assistance of the staff of the Laboratory, to control the
work in an efficient manner.

The Committee are of opinion that further expenditure will
be required in completing the set of platinum thermometers,
in particular in providing a satis&ctory resistance-box and in
carrying out the researches on the Clark cell. They consider that
it is of great importance that these researches should be brought
to a satisfactory conclusion.

For these reasons they recommend that they be reappointed,
with a grant of £60, that Lord Rayleigh be Chairman, and
Mr R. T. Glazebrook Secretary.

Appendix I*

On the Values of the Resistance of certain Standard Goils of the

British Assodation. By F. E. Smith.

{Frofn the NcUumcU Phytical Laboratory.)
[The Report covers the period 188S— 1903 inclusive.]

Changes of very considerable magnitude have taken place
since 1892 in the old B. A. standards. The removal of the coils,
first to Liverpool, then to Eew, and finally to Teddington, has
resulted in the comparisons being incomplete in some years. In
consequence the difficulty of locating diSerences has correspondingly
increased.

The observations recorded are in terms of B. A. Flat. Owing
to a change in Flat taking place, however, the 1903 comparisons
were made chiefly with Nalder 3715.

* See also Report for 190S.

40—2

628

PRACTICAL STANDARDS

In Table I. the approximate differences in B. A. U. between
Flat and the B. A. unit coils F, G, H of the Association are
given.

Table I.

Year

Flat

1888

+47xlO-»

+91x10-*

+ 77xl0-»

1890

-17

+112

+90

1891

—

1892

-18

+ 108

+92

1894

—

1897

—

1898

-36

+ 99

+69

1900

-47

+92

+63

1901

-42

+92

+ 70

1902

-33

+90

+76

1903

-33

+ 75

+76

Table II. gives the differences in ohms between* (1-01358
X Flat) and other platinum-silver coils. Temperature of observa-
tions, 16** C.

Table IL

Year

(101358 X Flat)

Nalder

Elliott

8715

264

209

270

1888

_—

1890

—

1891

+ 13x10-*

—

1892

—

1894

-17x10-*

-37x10-*

+ 27x10-*

1897

—

1898

-17

+ 9

-46

+ 27

1900

-17

+ 23

-59

+27

1901

-17

+ 23

-54

+ 27

1902

+38

-39

+ 44

1903

-39

+44

* LB.O.T. ohms 1*01858 B.A. U.

FOR ELECTRICAL MEASUREMENTS

629

Table III. shows the percentage diflferences between (1*01358
X Flat) and the unit of two 10-ohm platinum-silver coils of the
Association at IG^'C.

Table III.

Year

(1^01358 X Flat)

Elliott
288

EUiott
289

1897
1898
1902-3

-27xlO-»

-27

-10

+ 7x10-6
+ 7
+ 24

The coils F^ 0, and H are similarly constituted : they are the
old B. A. coils made by Matthiessen. No. 3715 is by Nalder Bros.,
and the remainder of the coils by Messrs Elliott Bros. No. 264 is
a coil belonging to the Board of Trade, and has been returned to
Whitehall ; hence there are no observations for 1903.

Tables I., II., and III. assume Flat to be constant. It will be
observed that the differences between Flat and 3715, 270. 288,
and 289 are constant from 1897 to 1901. From 1901 to 1903 a
change of about 0*017 per cent, is evident in the differences
between Flat and the coils 3715, 264, 269, 270, and again between
Flat and the units of the coils 288 and 289. This suggests a
change in the value of Flat since 1901.

Table IV. Values at 16° C. in terms of (1-01358 x Flat),

assuming Flat unchanged.

Year

Wolff
1690

Wolff
780

1-00002
•99987
•99987

Wolff
381

1-00014
-99999
•99999

Wolff
147

1901
1902
1903

1-00012
•99996
•99995

•99790
•99783
-99783

Since 1901 comparisons between Flat and the manganin
standards of the Association have been made. Table IV. gives
the observed values in ohms.

630

PRACTICAL STANDARDS

The values of 1690, 780, 881, and 147 dimmish by 17, 15,
15, and 7 times 10~' ohms respectively in the interval 1901 — 1902.
No, 147 is known to be a variable coil of very low insulation-
resistance, and may be disregarded for the purpose of estimating
the chtmge in Flat. It is of interest as being a coil brought to
Cambridge by Dr Lindeck in 1892 and left with the Secretary.

Thus the apparent faXU in value of 3715, 264, 269, 270, 288,
289, 1690, 780, and 381 are respectively 017, -015, "020, 017,
•017, -017, -017, -015, and -015 per cent., giving a mean of
•017 per cent.

This justifies the assumption of a rise in resistance of B. A.
Flat of -017 per cent, in the period 1901—1902.

The following tables, V. and VI., are I. and 11. revised. They
take the change in Flat into account by means of corrections
applied to the observations of the years 1902 €uid 1903. The
values given are for 16* C.

Table V. (L Revisei.)
B.A.U.

CoDsiant Flat

Tear

1888

+47 X 10-*

+91xl0-»

+ 77x10-6

1890

-17

+ 112

+90

1891

—

1892

-18

+ 108

+92

1894

—

1897

—

1898

-36

+99

+69

1900

-47

+92

+63

1901

-42

+92

+70

1902

-60

+ 73

+ 59

1903

-60

+ 68

+59

Tables VIL and VIII. being III. and IV., similarly revised,
show no marked change in any of the coils in those tables
excepting 147.

With reference to Tables V. and VI. the data for 1901—1903
show a rise of '008 per cent, for F, -034 per cent, for (?, bjA
'Oil per cent, for H, indicating that they are certainly changing
coils, the resistance for this period increasing with time.

FOR ELECTRICAL MEASUREMENTS

631

Table VI. (II. Bevised.)
Ohms.

Year

Constant (IM

01358 X Flat)

Nalder

Elliott

8716

264

269

270

1888

1890

—

——

—- .

1891

+ 13x10-*

—

1892

1894

-17xlO-»

-37x10-*

+27x10-*

1897

—

1898

— 17

+ 9

-46

+ 27

1900

— 17

+23

-69

+ 27

1901

— 17

+ 23

-64

+ 27

1902

— 17

+ 21

-66

+ 27

1903

"^ 1 /

— —

-66

+27

Table VII. (III. Revised.)
Values at 16^0.

Year

(1*01858 X Flat)

Elliott
288

Elliott
289

1897
1898
1902-3

-•27x10-*

-•27

+ 7x10-*
+ 7
+7

Table VIII. (IV. Revised.)

Year

Wolff
1690

Wolff
780

Wolff
381

Wolff
147

1901
1902
1903

1-00012
1-00012
1-00012

100002
1-00004
1-00004

1-00014
1-00016
1-00016

•99790
•99800
•99800

632

PRACTICAL STANDARDS

From the values recorded for 3715 and 270 we have evidence
that Flat has probably remained constant for the period 1894 — 1901.
Also we infer that 264 is not a coil showing very great changes.

Between the years 1892 and 1898 the differences between
Flat and the coils F, (?, and H alter by the amounts '018 per cent.,
•009 per cent., and '023 per cent, respectively. The dissimilarity
of these percentage-differences is further evidence that the coils
have changed amongst themselves in this period. Comparing the
amounts with those of the period 1901 — 1903, they represent
quite normal increments of resistance. The balance of evidence
in consequence is in favour of the constancy of Flat over the
period 1892 — 1898, and this constancy has therefore been
assumed.

A summarised statement of the platinum-silver coils of the
Association will now be as follows: —

Table IX. Showing the Percentage-ivci^ease in Resistance of
B. A, Platinum-silver Coils from 1888.

Coil
Flat

1888

1890

1891

1892

1891

1
1897

1898 ,

1900

1901

1
1902 1

—

•017

—

'064

—

•065

—

•088 1

•094

•089

•097

1B7 1

—

-•021

—

-•017

—

-•008

-•001

•018

—

-•OIS

—

-•015

—

•008

•014

•007

•018

KMS

8715

—

— ■

ohserv.
oommenoe

} -

864

—

observ.
oonimence

—

•004

-•010

-'010

-•008

269

—

observ.
commence

^■^

•009

'022

•017

•019

•w?

270

—

obderv.
commence

■^^"

288

observ.

commence

289

—

/' observ.
I ' commence

—

It will be observed that a number of the coils are steadily
rising in value. The insulation remains good.

Temperature Coefficients of B. A, Coils,

Some special observations have been made in order to obtain
the temperature coefficients of the coils. These were carried out
by keeping the standard coil constant and subjecting the tested
coil to various temperatures for twelve or more hours so as to

FOR ELECTRICAL MEASUREMENTS

633

ensure no lag. It is interesting to note that the temperature
coefficients of some of the coils are appreciably different from the
old values of 1892.

Table X. Showing the old and new values of the Temperature

Coefficients of Various Coils,

Coil

Flat
F
O
H

3715
264
269
270

Temperatare Coefficient
per PC, OldTalue

•000277 B. A. U.

286

274

271 „
•000260 ? ohm

312

Temperature Coefficient
per 1° C, New value

•000271 B.A.r.
268 „
274 „

280 „
•000307 ohm
283
285
315

99
9»
99

Comparison of the Unit of Resistance employed at the
Reichsanstalt with that of the N, P. L.

By the N.P.L. unit is meant the unit of resistance as obtained
from the old B. A. coils*. Assuming that all the changes have
been successfully interpreted, the unit at present employed in the
Laboratory should be the same as that employed in the Cavendish
Laboratory in 1898 and at Edinburgh in 1892.

A comparison of the two units was rendered possible in the
spring. Two Wolff coils, Nos. 780 and 738, of nominal values
1 ohm and 10 ohms respectively, were despatched to Germany
last winter. Their values were determined in Reichsanstalt units
(termed international ohms) in March, and the coils immediately
returned to the Laboratory. Unfortunately both coils fell in value
two or three parts in the hundred-thousandth figure during their
joumeyings. The values given in the table are those determined
on their return.

In addition, five new coils were received varying in value from
i^^th to 10,000 ohms. These enabled a more complete comparison
to be made. The Laboratory value was deduced by building up
from the unit, and also by direct comparison with coils of similar
value.

* 1 N.P.L. unit =1-01358 B. A. U.

634

PRACTICAL STANDARDS

Table XI. Results of Measurements of various coils at the
Reichsanstalt and at the Laboratory, March 1903.

Coil No.

Laboratory Value
at 17° C.

Value Dedaoed
from Beichsanstalt

Laboratory Value —
Beichflanstalt Value.

Certificate at 17° C.

Percentage Differenee

2352

•100007

-099996

•Oil per cent.

2351

1-00011

1-00001

-010 „

780

1-00001

-99991

•010 „

738

9-99946

9-9985

-0096 „

2450

100-004

99-993

•Oil „

2449

1000-06

999-96

•010 „

2448

10000-9

9999-8

•Oil „

It is evident from these observations that a difference of
•01 Ob per cent, exists, i.e, —

Resistance of Reichsanstalt unit — Resistance of Laboratory unit

= 000105 ohm (A).

Comparison of the Unit of Resistance employed at the Board

of Trade with that of the Laboratory,

The comparison of these two units is not so complete. Two
platinum-silver units and one of manganin have been measured
at both laboratories. The measurements made at Teddington
indicate that no change resulted during the joumeyings of the
coils. In addition one 1000-ohm coil (Nalder 6863) has been
measured.

Table XII. Results of Measurements of various coils at the
Board of Trade Offices and at the Laboratory, February
and March, 1903.

Coil No.

Temperature

Laboratory
Value

Deduced
B.O.T. Value

Laboratory
Value— B.O.T.

Elliott, 270
Elliott, 264
WolflF, 381
Nalder, 6863

160" C.
16-0** C.
160'' C.
15-84" C.

1-00006
1-00008
1-00015
999-13

l-OOOlOft
1-000146
1-00021 7

999I3

- •004ft per cent.
-■0065 „
-■OO67 „

FOR ELECTRICAL MEASUREMENTS

635

The exact relationship between the B.O.T. unit and that of the
Laboratory is therefore somewhat uncertain. However for the
unit coils we have —

Resistance of Laboratory unit — Resistance of B.O.T. unit

= -00006 ohm, a diflference of -006 per cent. (B).

From the two relationships —

Resistance of Reichsanstalt unit — Resistance of N.P.L. unit

= OOOIOb ohm,

Resistance of Laboratory unit — Resistance of B.O.T. unit

= -00006 ohm
we have

Resistance of Reichsanstalt unit — Resistance of B.O.T. unit
« -OOOieo ohm, a difference of -01 6a per cent (C).

The present values of the B. A. coils are as follows : —

Table XIII.

Coil

Flat
F
G
H

3715
269
270
288
289

Temperatare

Resistance

Temperatare Coeffiotent
per 1° C.

16-0^ 0.

1-00060 B. A. U.
1-00083 „
•99975 „
•99976 „
1^00050 ohm
1-00089 „
1-00006 „
10-0060 „
100026 „

-000271 B.A.U.
■000268 „
•000274 „
■000280 „
•000307 ohm
•000286 „
•000315 „

•oa3io „

■0026, „

The Wolff* manganin coils of the Association are also given at
16''C., with a temperature coefficient to be applied for small

Table XIV.

CoU

Temperature

1690
780
381*
147

16^0*' C.

Besistanoe

1-00012 ohm
1-00002
1-00016
•99800

Temperaiiire

Coeffideiii

per 1° C.

-00001 ohm

•00001

•00002

»»

•OOOOlfi „

* No. 381 18 a manganin ooil belonging to the Board of Trade.

636

PRACTICAL STANDARDS

ranges of temperature only, since it is by no means a linear
function.

As has already been explained, the values are given in terms
of the Laboratory unit which represents 10* C.G.S. units of re-
sistance as determined by Lord Rayleigh and Mr Glazebrook
at Cambridge. It has been assumed that the inter-comparison
of the coils enables that unit to be recovered.

Appendices I. and II. of the present Report afford the means
of connecting this unit with those of the Board of Trade, derived
from it in 1891, and of the Reichsanstalt, and also with the ohm
or international ohm — the resistance, that is, of a certain colunm
of mercury.

Appendix II.

The relation between the international ohm (106300 cm. Ug
weighing 14*4521 gms. at 0° C) and the unit of resistance
employed at the N,P.L. Preliminary Note, by F. K Smith.

(From the National Phygical Laboratory.)

The following measurements of six mercury tubes indicate the
progress made in this inquiry.

Conical

CJorreo-
tion

Tube

Theocetical
Length for

6-a

1 Int. Ohm

Calculated

Mean

Length at

(/ti-1)

Besistanoe

Measured

0°C.

xlQO

of Tube.

Besistanoe.

62-0731

Int. Ohm

Lab. Unit

621319

•99905

-99913

-00008

73-5000

73-4759

1-00033

1-00041

•00008

116-507

116-478

100025

1-00035

•00010

65-6338

65-6354

•99997

1-00007

-00010

62-1867

62-2382

•99917

-99926

-00009

68-5199

68-5057

1-00021

1-00029

-00008 1

Thus, Laboratory Unit of Resistance = -99991 Int. ohm

106-291 ^ , ,
= 106-300 ^''*- °^™-

[The above figures are intended as merely provisional.]

FOR ELECTRICAL MEASUREMENTS

637

With respect to the measurements of the cross-sections the
uniformity of the results show that an accuracy of *001 per cent,
may be relied upon. Four methods of measuring the resistance
will be employed. At present only two of these are completed.
The values in each horizontal line refer to different fillings ; they
are very concordant, as the values given in the following table
show : —

Resistance in Laboratory {N.P.L,) Units of

Mercury Tubes,

Tube

Besistanoe

atO^C.

Potentiometer

Benetance

at0«C.

Kelvin

Double Bridge

•99913
•99912
-99914

•99913
•99912
•99914

1-00041
1-00044
1-00040

1-00041
1-00044
1-000395

1-00034
1-00036
1-00035

1-00035
1-00036
100035

1-00007
1-00006
1-00007

1-00007
1-00006
1-00006

•99926
•99927
•99925

•99926
-99926
-99925

1-00030
1-00029
1-00029

638 PRACTICAL STANDARDS

Appendix III.

On the Platinum Thermometers of the British Assodatiofi,

By J. A. Harker, D.Sc.

{From the National Phytical Laboratory,)

The four platinum thenuometers numbered B.A<, to B.A.4, with
which this Appendix chiefly deals, were constructed at the National
Physical Laboratory in August 1902. The wire used for the
" bulbs " is approximately '006 in. ("15 mm.) diameter, and for the
leads '020 in. ('5 mm.).

After ascertaining approximately the length of wire necessary
to give a fundamental interval of 1 ohm, the proper amount for
the four thermometers was cut off from the stock reel, and heated
in one piece to moderate redness (800° C.) electrically when sup-
ported approximately horizontal. The platinum "lead" wires,
which were of the same quality of pure metal as the finer " bulb "
wire, were then measured off and the pairs assigned to each
thermometer accurately matched. After a preliminary anneal in
an oxidising atmosphere at a bright red heat, one of each of these
pairs was looped upon itself to form the compensator, and the
other cut in half for attachment to the ends of the " bulb " wire.
Several kinds of mica from different sources were tested as to
their suitability for use as insulating material for the frame and
washers to support the wires, and it was found that considerable
discrimination was necessaiy in the selection of the mica for this
purpose. Certain qualities which were colourless before heating
became on exposure to only 800** to 850° C. of a marked brown tint,
and it was found in one case this was due to organic material
having been used to fasten together several sheets to build up the
necessaiy thickness, the carbonaceous matter leading to a fall in
insulating power several hundred degrees below the temperature
at which good mica begins to appreciably conduct, which ought
not to be lower than 1150° C. In another case, a specimen which
showed the characteristic silvery white lustre after several hours'

FOR ELECTRICAL MEASUREMENTS 639

exposure to 1100'' C, had lost so much of its mechanical strength
as to be almost unusable. A specimen which before heating was
of slightly green tint was finally selected, and of this the whole of
the mica frames and washers were constructed. The copper wires
connecting the platinum leads to the fusible metal caps were
silver-soldered to the platinum, and for extra safety against
possible strain the wires were screwed into the caps as well as
hard soldered. In order to be protected as fisir as possible fix)m
unsymmetrical heating, which often gives rise to thermo-electric
effects in certain tjrpes of thermometer, these joints between
platinum and copper are arranged so as to be well inside the
brass tube into which the glass or porcelain protection tube is
fastened The thermometer heads are of ebonite, and are of the
design described by Harker and Chappuis in PhU. Trans. 194,
p. 52. They are practically airtight, and will stand vacuum or
pressure for a considerable time. By a small tap, which is
generally kept closed, communication can be made with a con-
venient apparatus for exhausting and letting in dry air while the
thermometer is suitably heated. The effect of electric leakage in
lowering the apparent resistance of a platinum thermometer when
damp is much more easily traced on thermometers of 5 or 10 ohms
FI than on the usual 1 ohm pattern used for high temperatures.
With the thermometers here described, having the enclosed form
of head, none of the determinations of fixed points have been
found to be vitiated by moisture, care having been taken not
to expose any portion of the interior to prolonged contact with
the outside air, after once being thoroughly dried out at a high
temperature.

The mica cross, having serrated edges with teeth of 1 mm.
pitch, being attached to the leads and compensator, the joints
between the " bulb " wire are made in the strongly oxidising flame
of a very small oxy-coal-gas blowpipe without admixture of foreign
material of any description. Autogenous soldering of this kind is
not very difficult, even for very fine wires, and is essential if the
thermometers are intended for use to the highest temperatures
safely measurable, namely, 1150** C, as the copper and silver
contained in any solder which might be employed give off vapour
sufficient to injuriously affect the platinum on prolonged exposure
to a temperature considerably below this. The " bulb " wire when
fastened to the leads is then wound, not too tightly, upon the

640 PRACTICAL STANDARDS

mica frame, and the thermometer is then inserted into its pro-
tecting tube of very thin glass or of porcelain, which must be
glazed on the exterior, and if the thermometer is not intended
for use above about 1000° C, may with advantage be glazed both
inside and out. A measurement is then taken of the fundamental
interval, with a view to ascertain the change on cmnealing, which
is then carried out by heating two or three times to about lOOO"" C.
for several hours, with slow cooling, the thermometers with glass
tubes being temporarily placed in porcelain ones for this purpose.
The fundamental interval is then taken again, and if this is not
considered sufficiently near the desired value, it can be lowered
by cutting out the required amount from the looped end of the
wire and re-Aising, or raised by stretching judiciously with
platinum-tipped pliers the lowest few inches of the wire, which
is unwound for the purpose. Care must be taken after each re-
adjustment to remove any possible new strains introduced by a
thorough re-aimeal before measurement. In the absence of
definite evidence in its favour, it was not deemed desirable for
this first set of thermometers to heat the wire for some hours
electrically to 1400° or 1500° C, as is usual in careful work
with wires of platinum and the allied metals employed for
thermo-junctions.

After the final adjustment of the FI and final anneal,
systematic observations of the zero, steam, and sulphur points
of the four thermometers were made fi*om time to time with the
resistance-box described in the text. A new calibration of the
box-coils and bridge wire was made in February 1903, and the

values of the relation -p- and of the h found since that date are

tabulated for each thermometer. From this summary it will be
seen that there appears to be a small but systematic difference
between thermometers 1 and 3 on the one hand, and 2 and 4

on the other, this being noticeable both on the values of -^
and of S.

The values of § vary fix)m 1*38709 in B.A.1 to 1-38881 in BJl.„

the mean of the four being 1*38786, which is a little higher than
the mean value found by Chree for the group of seven thermo-
meters studied by him, namely, 1*38702.

FOB ELEOTBICAL MEASUREMENTS

641

The mean values of the B are :

Departure
from Mean

B.A.4

B.A.2

1-5124
1-5083
1-4935
1-4912

+ •0110
+•0069
-O079
-0192

1
1

Mean d=

1-5014

The mean S of the six thennometers observed in sulphur in
Chree's experiments was 1*503, the maximum being 1*509 and
the minimum 1*498. The mean values of the R^, Ri, and FI for
the period from February 12 to August 31 are also given. In
view of the uncertainties in the measurement of the temperature
of the box-coils, which are of platinum-silver not immersed in a
liquid, and also of small irregularities in the behaviour of the
plug-contacts, the experiments afford no certain evidence of
systematic change in any of the thermometers, unless it be a
small rise in the fundamental coefficient and corresponding fall in
theSofBA.1.

RA.,.

Date

Feb. 6, 1903

»> 23, „
Aug. 7, „

26, „

26,

31,

i2o
267-905

B A.

1-514
1-605
1-506
1-505
1-514
1-506

Mean 1-5083

1-38688
1 -38702
1-38708
1-38712
1-38722
1 -38722

1-38709

Jfecm Value of Constants

Ri
367-736

99-831

1-6083

Difference of
from Mean

- -00021
--00007
-•00001
+ O0003
+ -00013
+ O0013

1-38709
41

642

PRACTICAL STANDARDS

Thermometer BA.^, which was heated about fifty times during

November 1902 in electric furnaces up to 1050", and again during

April and May 1903 to similar temperatures for prolonged periods,

appears to be hardly perceptibly affected by it, no certain change

of FI occurring during the period February 12 to August 18

covered by the later experiments, and certainly no variation of

the zero of "1" C.

B.A.,.

Date

» 1

Difference of ■^-

-no 1

from Mean

Feb. 12, 1903

1-484

1-38867

- -00014

„ 24,

1-499

1-38877

-•00004

», 24,

1-495

1-38874

-•00007

May 19,

1
1

1-38876

-•0000ft

July 30,

1-497

1-38880

-•00001

Aug. 18,

1-489

;

1-38863

-•00018

» 21,

1-38901

+ ■00020

„ 24,

'* 1

1-38890

+ -00009

„ 24,

1-493

1-38889

+ -00008

„ 26,

1-491

1
i

1-38892

+ •00011

„ 31,

1-488

1-38882

+ 00001

Mean 1*491

1-38881 '

Mean Value of Corutants

J2o

267172

357163

99 991

1-4912

1-38881

To see if the small lack of homogeneity of the wire as shown
by the properties of the different thermometers was due to the
treatment it had received during the successive adjustments of
Fly a new thermometer, named B.A.7, was made up of wire taken
from the inner end of the same reel as the other six. No attempt
was made at adjustment of its FI, which was found after thorough
annealing to be 100*022 box units.

The S was found to be 1*506, an intermediate value. The

wire was then unwound from the mica frame and suspended
freely in air between the ends of the leads, and a current of
2^ amperes, which was sufficient to maintain it at about 1400"" C,
was passed for about 2 hours.

FOR ELECTRICAL MEASUREMENTS

643

Owing to the volatilisation of a considerable quantity of
platinum from the wire, a large increase in the FI was found,
as was expected, but the i remained unchanged, though a rise in

-n- was recorded amounting to 1 part in 1000.

In order to make certain that the differences observed were
not due to defective insulation in the thermometers, the insulation
resistance between the thermometer and compensator leads of each
of the thermometers was measured by a direct deflection method,

B.A.,.

Date

Difference of -=-

from Mean

Feb. 9,1903

1-511

1-38710

+ ■00010

„ 26,

1-511

1-38730

+ -00000

Aug. 10,

1-509

1 -38714

-•00016

,. 18,

1-522

1-38732

+ -00002

„ 24,

1-511

1-38724

-•00008

» 26,

1-515

1-38736

+ ■00006

„ 26,

1-510

1-38731

+ 00001

„ 31,

1-510

1-38738

+ •00008

Mean 1-5124

1-38730

Mean Value of Constants

Ri,

268-367

358-434

100-067

1-5124

1-38730

and found to be in no case less than 700,000 ohms at any
temperature between 0° and 1000" for B.A.i and B.A.2, and 0" and
500" for B.A., and B.A<4. Some experiments were also made on an
imitation platinum thermometer having its coil wound on mica of
standard quality, but cut at the lower end into two parts.

Although the insulation from one part to another was
practically infinite at all temperatures, when only platinum and
mica were present in the heated part of the porcelain tube, the
introduction of a small piece of clean copper wire into the hot
space near the bulb was sufficient after some time to lower the
insulation, even at only about 800" C, to a few thousand ohms.

41—2

644

PRACTICAL STANDARDS

The cause of the dififerences between the individual thermometers
does not, therefore, appear to be leakage.

Neither does the cause of the small differences in values of S
found lie in the method of taking the sulphur point, as the same

B.A«4.

Date

Difference of -s-

-«o

from Mean

Feb. 11, 1903

1*486

1*38816

-•00009

n 23, „

1*499

1*38833

+ •00008

n 26, „

1*600

1*38826

+ •00001

Aug. 10, „

1*497

1*38836

+ •00010

« ^®> n

1*473?

—

»> 24, „

1*497

1*38825

+ •00000

» 26, „

1-494

1*38812

-•00013

»» 31> «

1*603

1*38826

+ •00001

1*4936

1*38826

Mean Value of Constants

^ 257-627

357-616

99*989

1*4935

1*38825

apparatus was used in the same way for all the experiments. The
sulphur is now boiled in an arrangement similar to Callendar and
Oriffiths's well-known pattern, except that, to avoid the necessity
of removing the tube at each reheat after the sulphur has
crystallised, the glass boiling-tube is replaced by one of thin
weldless steel, brazed with spelter into a rather wider end-piece

267*749

367*771

100*022

1-606

Thermometer eleotrically heated to 1400° for 2 honrs
270-036 375-213 106*177 1*606

1*38806

1*38949

FOR ELECTRICAL MEASUREMENTS 645

of thick iron tubing, which is exposed to the direct flame of the
lai^e bunsen used for heating. The level of the liquid sulphur is
always maintained at least 2 inches above the bottom plate of the
apparatus, and the upper level of the vapour to a definite position,
which can be seen through mica windows in the upper part of the
neck. * Under these conditions no measurable superheating of the
vapour has ever been observed, and a comparison of the sulphur
points obtained with this form of apparatus with those got in the
older one, with glass boiling-tube, reveals no measurable systematic
difference.

For the boiling-point of sulphur under normal pressure in
latitude 45^ Callendar and Qriffiths's old value, 444*53^0., has
been retained, as was also the figure deduced by them fix)m

Regnault's experiments for -j- for sulphur, namely, 0*082® C.

per mm., although it has been shown independently, by Chree
and by Barker and Chappuis, that this value for the variation
is considerably too small It is hoped that a redetermination of
this constant for pressures between 700 and 800 mm. will shortly
be undertaken in the thermometric laboratory.

646 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

Appendix IV.

The following table gives the resistance at a temperature of
60*" Fahr. (15*65° C.) of a wire of pure annealed copper 1 metre in
length, having a mass of 1 gramme, as deduced from the most
recent determinations.

In making the reductions, the values for the temperature
coefficient and for the density given by the author, have been
used.

Table giving Resistance at 60*^ Fahr, of a Wire of Pure Annealed
Coppery such that 1 metre weighs 1 gramme.

Authority

Fitzpatrick

Swan and Rhodin

Do. (second sample)...

Fleming* ,

Lagarde

Mean value

Source of Copper

Electrolytic
Swan's Copper

Swan's Copper
Grammont Electrolytic

Reference

B.A. Report, 1890

Froc, R. S,y 1894

„ „ 1894

Pha. Mag., 1893

ffogpiUdierf 1894

Value in
Ohm

0*1475
0*1493
0-1486
0-1487
01488

01486

* In reducing Professor Fleming's result, the density has been taken as
9*91 grammes per c.c.

THIRTY-FIRST REPORT— CAMBRIDGE, 1904.

APPENDIX PAGE

I. On AnomcUiM of Standard Cells. By F. £. Smith. (Fyom the

National Physical Laboratory) 651

II. On the Electromotive Force of Clark's Cell. By A. P. Trotter 661

The Committee desire to record their deep regret at the death
of their colleague, Profeseor Everett. He had been a member of
the Committee almost since its commencement. He attended the
meeting at which the present Report was considered. His work
in connexion with the G.O.S. system of units is of great importance
and has proved of very real value to science.

The Committee are glad to report that during the year con-
siderable progress has been made with the construction of the
Ampere Balance. Mr L. Oertling has constructed the weighing
mechanism, which has, however, not yet been taken over by this
Committee, and the electrical parts of the instrument are nearing
completion in the workshops of the National Physical Laboratory.
The following particulars of progress and of applied tests may be
of interest.

1. The weighing mechanism. The castings, rods, tubes, screws,
etc., intended for this had their magnetic permeability determined,
and no part used in the construction has a permeability differing
from unity by more than O'OOl per cent.

The balance was examined for stability and sensitiveness at
Messrs Oertling's works with satisfSsustory results; a difference of
one-tenth of a milligramme may be detected.

2. The marble cylinders and fittings. Insulation and per-
meability tests were made on various samples of marble early in
the year ; eventually First Statuary Carrara Marble was chosen as
most suitable for the work. An experimental marble cylinder
was wound with a double helix and the insulation satisfSsustorily
carried out ; the results of the tests leave little doubt as to the

648 PRACTICAL STANDARDS

advantages of the double helix. The winding of both suspended
cylinders has now been completed, and it is anticipated that
the fixed cylinders will be finished in September. The linear
measurements and insulation tests have yet to be made. Unless
unforeseen difficulties arise the balance equipment should be com-
pleted, and the whole ready for preliminary observations, by the
end of the year.

During the early part of the year Mr F. E. Smith completed
his researches into the construction of a mercury unit of resistance,
of which some account was given in the last Report. The results
have been communicated to the Royal Society and are being
published in the Philosophical Transactions. The values of the
various tubes (eleven in number) are very accordant, and a
mercury standard of resistance of a high degree of accuracy now
exists. Since the completion of his work the specification of the
Clark cell has engaged Mr Smith's attention, and a detailed
account of his work forms an Appendix to the present Report.
Mr Smith has amply confirmed the result of previous investi-
gators that the greater part of the difficulty in obtaining entirely
concordant results for the various cells set up by different ex-
perimenters is due to the mercurous sulphate. He describes
three methods of preparing the paste which lead to identical
results, and which have the advantage that cells set up with these
pastes have the same E.M.F. within one or two hundred thousandths
of a volt immediately after manufistcture. In the first method
due to Professor Divers and Mr Shimidzu the paste is prepared by
the action of fuming sulphuric acid on mercury; in the second,
following Professor Carhart, it is prepared by the electrolysis of
weak sulphuric acid and mercury; while in the third mercurous
sulphate is dissolved over a water bath in sulphuric acid: the
acid solution is then poured into a large volume of distilled water
and the mercurous sulphate is precipitated in a pure form. In
all cases it is important that, as advised by Mr Swinburne and
Professor Carhart, the salt should be washed, for a Clark cell, with
zinc sulphate, and for a cadmium cell with cadmium sulphate, and
not with distilled water. Mr Smith is continuing his inquiries
and hopes shortly to be able to issue a complete specification
for Clark and cadmium cells. The completion of the Ampere
Balance will enable an absolute determination of their B.M.F. to
be made.

FOR ELECTRICAL MEASUREMENTS 649

The Committee regret to report that no further progress has
been made since their last Report with the experiments to deter-
mine the permanence and reliability of the platinum resistance
thermometers described in that Report.

It was pointed out last year that a special resistance box was
required to enable the work to continue ; unfortunately the funds
necessary for its purchase were not forthcoming, and the work has
remained stationary for a year.

The Committee would consider it most unfortunate if work of
a very real importance on which a start has already been made
and considerable funds expended in the purchase and investiga-
tion of pure platinum wire should lapse for want of support, and
they trust that their recommendation in &vour of the continuance
of the work may this year be accepted.

Meanwhile they would call attention to the very complete com-
parison up to a temperature of lOOO^'C. between the constant
volume nitrogen thermometer, the platinum resistance thermo-
meter, and the platinum — platinum-rhodium thermo-couple com-
municated recently from the National Phjrsical Laboratory to the
Royal Society by Dr Harker.

The Committee have received a cordial invitation to co-operate
in the Electrical Conference at St Louis during the forthcoming
autumn, and have asked Professor Perry and the Secretary, who
are attending as delegates of the Listitution of Electrical
Engineers, to represent their views on two questions of special
interest.

The first of these relates to a proposal by Professor Carhart to
substitute the saturated cadmium or Weston cell for the Clark
cell as a recognised standard of E.M.F. The Committee are aware
that the iact that the temperature coefficient of the cadmium
cell is one-twentieth of that of the Clark cell offers many valuable
advantages, but in view of the &ct that experiments designed to
lead up to a satisfiEUstory specification of the cell are in progress
at the National Physical Laboratory, and that the completion of
the Ampfere Balance would enable the absolute E.M.F. of the cell
to be determined, the following resolution was passed at the last
meeting : —

"The Committee are not prepared at present to displace the
Clark cell and prefer to wait for the conclusion of the experiments
at the National Phjrsical Laboratory, and with the new balance,

650 PRACTICAL STANDARDS

before coming to a decision as to the value to be assigned to the
E.M.F. of the saturated cadmium cell."

The second question relates to certain proposals as to nomen-
clature which are to be brought forward by Dr Eenelley. These
are (A) that a systematic nomenclature should be agreed upon for
magnetic units, and (B) that the prefix " Abs " should be used to
indicate that a unit is given in the absolute cojs. electro-magnetic
system, and '' Abstat " to indicate that the unit in question is in
the absolute CGJs. electrostatic system.

Thus an '' Abe " volt would be the C.G.S. electro-magnetic unit of
E.M.F. and an '' Abstat " volt the C.Q.S. electrostatic unit of E.M.F.

These proposals have been discussed by the C!ommittee, who
have agreed to the following resolution: —

" With regard to the choice of magnetic units the Committee
are of opinion that the only two systems which need to be con-
sidered are the CQJS. system and the Ampere- Volt-Ohm system,
and that the quantities to be named, if any, are

(1) Magnetic Potential,

(2) Magnetic Flux*

(3) Magnetic Reluctance.

Of the above two alternatives the Committee are in fSeivour of the
C.Q.S. system as that on which to base any nomenclature of mag-
netic units, but are of opinion that a system of nomenclature is not
called for."

The Committee disagree with Dr Eenelley's prefixes for the
absolute electro-magnetic and absolute electrostatic systems of
units, and express the opinion that no system of prefixes should
be employed in which each prefix does not bear some definite
numerical signification.

In view of the work still necessary with regard to the Ampere
Balance, the cadmium cell, and the platinum standard of tempera-
ture, the Committee recommend that they be reappointed, with a
grant of £50, that Lord Rayleigh be Chairman, and Dr R. T.
Glazebrook Secretary.

* The name '* Maxwell " waa recommended by the Paris Congress, 1900, as the
name of this unit, and this recommendation was adopted by the Committee at
Bradford.

FOR BLECTBICAL MEASUREMENTS 651

Appendix I.
On Anomalies of Standard Cells. By F. E. Smith.

(From the NcUumcd Phytioal LaborcOory.)

During the past two years certain anomalies of Clark and of
cadmium cells have been under investigation at the National
Physical Laboratory. The work is still far from completion, but
the essential results so &r obtained are given in this paper.

In March 1902 some experiments at Bushy House resulted in
the detection of the depolariser employed in both standards as the
great disturbing element Lord Rayleigh, in his paper in the
PhiL Trans, for 1885, § 44, had shown this to be the case, and
Mr Swinburne arrived at the same conclusion in 1891*, while
recently in America Professor H. S. Carhart and Mr G. A. Hulett
have traced the variations in E.M.F. of the cadmium cell to the
same source. A new specification of the mode of manufacture of
the paste was thought to be desirable, and this problem was the
first to receive attention.

In order to be independent of the variations of the other
elements, cells were constructed of a type indicated by the
arrangement

Hg — Paste — Solution and Crystals — Paste — Hg
(a) (6)

where a and h represent pastes made with different samples of
mercurous sulphate. The Bayleigh H form of vessel was
employed. Preliminary observations showed that when the same
paste occupied the two limbs, such a cell had no measurable E.M.F.
In addition a cell typified by the arrangement

Amalgam

X—l^ Solution _|_fi}
w ^^ and Crystals ^^ oq

Paste (B)
See B. A. fieport, Cardiff 1S91.

652

PRACTICAL STANDARDS

was larg^ely employed, a four-limb vessel, similar to two Bayleigh
H form of vessels crossed, being used to set up the cell. In
this case there is one negative pole and three positive ones, and
the E.M.F. between any two of them may be measured. Such a
cell not only indicates whether a particular paste is abnormal or
not, but each of the three groups of elements may be compared
with an external standard. It is possible, of course, that a change
resulting in one of the pastes may affect the neutrality of the
solution, and so affect the E.M.F. of all three groups. All obser-
vations were made in a constant temperature room, the cells being
immersed in paraffin oil.

Table I.

Date of

Clark Cell, No. 1 (4 limbs)

Clark CeU,

No. 28
(3 Umbe)

Observation

H>R K:>R H>K

H:>K

Sept 8, 1902..
Sept 30, „ ..
Oct 23, „ ..
Dec. 2, „ .
Feb. 24, 1903 ..
June 2^ „ ..
Nov. 2, „ ..
Feb. 6, 1904 .
July 9, „ .

+0-00213

195

150

123

-0-00001

+0-00047
45
46
43
42
37
30
41
51

+0-00166

150

104

- 0-00014

1
+0-00168
104
79

59 ,
No obflerv.
- 5
-32
-50
-70

The earlier results of the investigation are omitted, but the
differences in E.M.F. due to pastes made from purchased samples
of mercurous sulphate are shown by measurements made of cell
No. 1 (4 limbs) and cell No. 28 (2 limbs), the observations covering
a period of rather more than two years. The pastes have been
distinguished by the letters K, H, and R ; all were subjected to
the same treatment and advantage taken of the latest methods for
their preparation.

It is clear that although the effect of each paste is not known
two of them have certainly changed, of which one is K, In the
chart curve H>R shows the change in E.M.F. of the H group,
assuming the R group to remain constant; similarly the J3[>K
curve represents the change in voltage of this group, K being
assumed constant, and like remarks apply to the third curve.

FOR ELECTBICAL MEASUREMENTS 653

There is a sudden break in the directions of the curves H> K
and K > R after the observations of November 2, while none
is shown in H>R; the deflections consequently indicate that
the element K must have changed in an abnormal fashion.
Indeed between November 2, 1903, and July 9, 1904, the E.M.F.
of the K group apparently increased by at least 0*0003 volt; a
rise of exceptional magnitude. A fall in voltage is the usual
feature.

The fact that the E.M.F. of a cell had changed by as much as
0'16 per cent, was very disconcerting. It is true that a difference
between H and R was anticipated, for H was a pale yellow
colour, while R was grey. On the other hand the paste K was
also slightly yellow, yet no such difference is observed between
K and jR. It was thought that the mode of manufacture
of the sulphate might influence the properties of the product.
Mercurous sulphate is often prepared by precipitation, either
Hg,(NO,), and Na5304 or Hg,(N0,)8 and H^04 being em-
ployed ; traces of the resulting nitrate in the final product would
certainly introduce a disturbing element. Again mercurous
nitrate is often associated with a basic nitrate, and basic salts are
to be avoided, as will afterwards be seen. Samples of the salts
were prepared by these precipitation methods and the results
were far from satisfistctory. Two samples of the sulphate were
also obtained from the same manufacturer; the Clark cells pre-
pared with these differed in E.M.F. by 0*0004 volt; both were
subjected to the same treatment and there was no observable
difference in colour.

A more satisfactory specification of the depolariser being
desirable, other modes of manu&cture eliminating the above
troublesome elements were sought. Concentrated sulphuric acid
and mercury react very slowly at ordinary temperatures, but much
more rapidly at temperatures approximating to 300*" C. The
resulting salt is associated with H,S04, which probably is not
very difficult to remove if the salt be in a fine state of division.
Dr Muirhead has prepared mercurous sulphate in this way and
forwarded two Clark cells containing the product to Bushy
House. A second method of preparation due to Divers and
Shimidzu is reported in the Journal of the Chemical Society (XLVii.
639). Briefly, pure mercury and fuming sulphuric acid saturated
with SO, are brought into contact. They react in the cold, though

G54

PRACTICAL STANDARDS

there is no visible evidence for some time owing to the solubility
of Hg^04 in the liquid ; the acid also becomes saturated with SO,.
If the main portion of the liquid be removed from the resulting
salt, this latter may be freed from SOs by warming; mercurous
sulphate associated with HaS04 is thus obtained. The acid is
removed by washing. Dr Carpenter has prepared five samples of
the salt in this way ; they were made from five lots of the fuming
acid from different manufacturers and mercury distilled in vacuo
at the laboratory. These sulphates were examined in four-limb
cells of the cadmium type; the results of the observations are

200

150

100

0-0

-50

100

■\

—zr
O

-J

-o

—

■^

^«»«

■

•"

—

t;?

»««

...

■ ••■

...

-;-

*#t

— Cu-

■*ir

•••

■•••

•••«

'to —

•<

•

ob;

iER^

rATr

,.J

SEP ORG
1()02

FEB

JUNE
IQ03

NOV

FEB

1904

JULY

given in Table II. The standard of reference is cell No. W 17, a
cadmium cell more than two years old and known to have changed
but little since its manufiu^ture.

The pastes 52a, 53a, 54a were prepared with the same sample
of HggS04; it was purchased and prepared in a similar manner
to the sulphates dealt with in Table I. 546 was also a purchased
sulphate. The remaining specimens were prepared by Divers'
method.

It will be observed that all the pastes change so as to reduce
the E.M.F. of the cell ; but whereas the E.M.F. of the cells prepared
with purchased sulphates is greater than that of WIT, those

FOR ELECTRICAL MEASUREMENTS

655

made up with the Nordhausen sulphates have in each case lower
E.M.F.'8. Cell No. 526 is exceptional in the &11 of its voltage.
The difference in the prepared pastes, though small, condemns
part of the method of preparation, and further investigation
became necessary.

Table II. e.u.f. of Cadmium Cells minus e.m,f. of WIT,
Differences in hundredths of a millivolt

Cell No.

CeU No. 54

Date of

Obfier?atiou

1
1 1

a b

n 1 b

h c 1

April 4,1904...

+ 38

-21

-19

+ .38

-14

-13

+ 38

+ 14 -18

April 18, „ ...

+ 35

-.30

-20

+39

-13

-14 +a5 +13

-18

May 3, „ ...

+ .30

-43

-20 +32 1 -14

-14

+ 34 +12

-19

June 13, „ ...

+ 27

-50

-21 1 +32' -14

-14

+ 34

+ 10

-19

July 9, „ ...

+ 23

"54

-21 +30 1 -14

-15

+ 34 +10

-■20 1
1

The method of preparation adopted by Dr Carpenter was at
first repeated. Close observation showed that on formation the
sulphate cakes considerably, and is accompanied at the surfisM^e of
contact with the mercury by a compound of a light brick-red
colour. If without freeing from the acid or SO, the product is
added to distilled water, reduction of part of the sulphate appa-
rently occurs, mercury is precipitated as a black powder, and the
red compound entirely disappears. (The mercury thus precipitated
is a valuable addition to the paste, the conversion of mercuric
sulphate to the mercurous condition being rendered possible by
its presence.) The salt produced by freeing the first product firom
SO2 also loses the brick-red tint, and is finally obtained as a pure
white paste. On prolonged washing with water, however, hydro-
lysis results and the colour changes to pale yellow. Two samples
of hydrolysed mercurous sulphate were thus prepared, the one
being washed for one hour with water and the other for twenty-
four hours. An experimental cell indicated that the more hydro-
lysed product if employed to set up a cadmium cell would cause
the E.M.F. of that cell to be greater by 0*00064 volt than if prepared
with the first sample. The presence of this hydrolysed product is
therefore to be avoided, and washing by water prohibited.

656 PRACTICAL STANDARDS

About this time, through the kindness of Professor Ayrton,
the results of some experiments by Professor H. S. Carhart and
Mr G. A. Hulett, of the University of Michigan, were com-
municated to the Laboratory. Professor Carhart has also sought
a standard method of preparing the depolariser, and suggests that
any prepared sulphate be washed with cadmium sulphate (or zinc
sulphate for Clark cells) in order to prevent hydrolysis. Prior to
this, Mr Swinburne, in a letter to Dr Qlazebrook, suggested the
precipitation of the sulphate intended for Clark cells from saturated
solutions of mercurous nitrate and zinc sulphate, the washing to
be effected with alcohol or saturated zinc sulphate solution.

Omitting the description of further experiments, the final
mode of preparing the mercurous sulphate for standard cells is
here given.

Fuming sulphuric acid saturated with S0| (32 per cent of
SO, is a convenient specification) is added to sufficient pure
distilled mercury to ensure the latter being alwajrs in excess.
The mercury should be contained in a clean glass vessel and
violently agitated by a glass stirrer, so that the product may be
in a fine state of division. After seven or eight hours the reaction
will be sufficiently advanced for the sulphate to be separated from
the acid, but if convenient the action may go on for some days.
Carefully pour off as much of the strong acid as possible into a
large volume of water or into an empty vessel, and afterwards add
the pasty product to thirty or forty times its bulk of distilled
water. Mercury is precipitated and a considerable quantity of
heat is evolved owing to the dilution of the acid. A few minutes
suffice for the sulphate to settle, when the acid liquid may be
decanted and the salt well washed by agitation with acidulated
water (1 part of cone. H^04 to 10,000 parts of distilled water).
Filtering follows, a filter pump being employed to effect ex-
haustion. It is advisable next to pound the damp sulphate
thoroughly in an agate mortar to ensure the absence of small
caked masses, after which acidulated water is again added, filtering
effected, and the salt washed on the filter-paper with two or three
lots of neutral saturated cadmium sulphate solution (or zinc
sulphate solution for Clark cells). The salt is now removed to a
small flask, saturated cadmium sulphate solution added, and the
whole well shaken and then allowed to stand for twenty-four
hours. Filtering follows, then three more washings with cadmium

FOR ELECTRICAL MJCASUREMENTS

657

sulphate solution, removal to a flask once more with (MSOa
solution, and at the end of twenty-four hours the solution should
still be neutral to Congo red. If so, the sulphate may be filtered
and is ready for the manufacture of the paste. The whole of the
operations should be conducted in a room screened firom sunlight.
As thus prepared the mixture of mercurous sulphate and mercury
is of a dark grey colour. Cells set up with paste prepared firom
it require no ageing, and the constancy obtained with pastes made
fix>m materials obtained fix>m different sources is an indication of
the purity of the salt.

Table III. gives the results of comparisons between cadmium
cells set up with pastes prepared in this way and cadmium cell
iri?; The latter in eveiy case has the greater E.M.F. Differences
are expressed in hundredths of a millivolt.

Table III.

CeU No. 1

Cell No.

Cell No. G8

Date of
ObFervation

—

1 May 12, 1904...

-24

-26

-24

M 1^1 »» •••

-21

--20,

-22

n '^2, „ ...

-2O5

-20fi

-21

—

~~^' 1

„ 16, „ ...

——

—

-27

-25

-28

»» 16, „ ...

-22

-21

—

„ 16, „ ...

-21

-20

-21

-20

— ~

June 1.3, „ ...

—

-29

-27

„ 13, „ ...

—

-2.3

-22

-2.3

1 >» !•% ij •••

-21

-206

-21

-20

-206

-19

-21

-20

-19

, July 6, „ ...

-21

-20

-196

-20

-196

-20

-21

-20 !

»» ^j » •••

-206

-20,

-20

-21

-20

-206

-196,

The first set of observations with each cell was made about
five minutes after adding the solution ; the second set of observa-
tions about twenty minutes afterwards ; and the third set three
hours afterwards. For the first two sets of observations the tem-
perature of the four-limb cells was unsteady; for the remaining
observations they and Wn were at the same steady temperature.
Other cells of the Rayleigh H form have been constructed, and
the comparisons are equally satisfactory.

An alternative method of preparing the salt was next sought.

B. A. 42

658 PRACTICAL STANDARDS

This second method is very simple. Any purchased sample of
mercurous sulphate is heated together with mercury and am-
centrated HsS04 on a water-bath for half an hour, the mixture
being stirred occasionally. At the end of that time the remaining
solid is allowed to settle and the hot clear acid carefully poured
into a large volume of distilled water. Mercurous sulphate is
soluble to a considerable extent in hot concentrated H^04;
the result of the dilution is to precipitate the salt. As thus
produced the mercurous sulphate is in a finely divided state
and of a pure white colour. It is well to ab once admix
with a little mercury and filter. The washing is performed as
before. Portions of three purchased samples of Hg^04 were
dealt with in this way, and after treatment gave identical results
with cells dealt with in Table III. The three original samples
prepared in the ordinary way produced cells differing in E.M.F.
from the standard by 40, 160, and 10 hundredths of a millivolt
A third method devised by Professor Carhart does not
necessitate the use of concentrated acid. In order to hasten
the reaction between mercury and dilute sulphuric acid (one to
six) an electric current is parsed from the mercury to a sheet
of platinum foil suspended in the liquid. It is essential that
the liquid be kept well stirred so as to keep the mercury
surface exposed. Professor Carhart has employed a beaker or
crystallising dish to contain the liquids, and used a current of
about 0*3 ampere; the current density, however, is not stated.
At Bushy House the salt so produced has been compared with
those prepared by the two previous methods. Under ordinary
circumstances about three grams of the salt — very grey owing to
the presence of mercury in a fine state of division — is obtained
per hour. The current density at Bushy House has been about
O'Ol ampere. It was gratifying to find that the product (washed
as before) gave identical results with the other methods. Very
violent agitation was maintained during the preparation. When
the liquid is not stirred a yellow compound (apparently turpeth
mineral HgS04 . 2HgO) is also produced, and cells the pastes of
which are prepared with the product have an E.M.F. when first set
up more than a millivolt higher than the normal. Particular
stress must therefore be laid on the instruction to keep the
mercury surface well exposed. The same thing was found to
happen when attempting to form mercurous sulphate by the

FOR ELECTRICAL MEASUREMENTS

659

electrolysis of a saturated cadmium sulphate solution in an H-form
vessel, the electrodes being pure mercury.

It will be observed that the remarks on the depolariser apply
equally to Clark and to cadmium cells. Cadmium cells sJone
were made up in the final tests because of their small temperature
coefficients; but Clark cells have also been set up and similar
results obtained. It is also necessary to add that all purchased
samples of HgsS04 are not so abnormal as those dealt with in
Table I., nor does the E.M.F. of an abnormal cell always fall so
rapidly as is indicated there. (The rate of fall is probably a
function of the fineness of the sulphate.) Evidence of re-
markably stable cells set up with purchased mercurous sulphate
is afforded by six cadmium cells made at Bushy House in April
1902 : these have been in constant use, and in the case of two of
them have fi:^quently been short-circuited through 100 ohms.
One of these cells is taken as a standard in the comparisons. By
reference to a seventh cell made up in June 1904 with a paste
made from sulphate identical with that employed for the previous
ones it is thought probable that the whole six cells have fallen 0*07
millivolt since their manufacture. Table IV. gives the result of
the comparisons.

Table IV.

Date of ObserTation

May 5, 1902...
Sept. 12, „ ...
Feb. 25, 1903...

Feb. 6, 1904...
July 9, „ ...

16>17

10>18

16>19

16>20

16>21
■00000

•00000

+ 1

+ 0,

•4- O5

+ Ofi

+ 05

- I5

Oft

At the present time the E.M.F. of a cadmium cell set up with
a paste made from fuming sulphuric acid and mercury is less than
that of these cells by 0*2 1 millivolt.

With respect to the other elements of standard cells it is
proposed to investigate the cadmium and zinc amalgams, and the
solutions of the sulphates of these metals, in a manner very
similar to that employed for the pastes. Much valuable informa-
tion has fortunately accumulated respecting the influence of
impurities in these, so that probably the task is a light one. It

42—2

THIRTY-SECOND REPORT— SOUTH AFRICA,

1905.

Appendix. On the Preparation of a Cadmium CdL By F. E. Shpth.

(From the National Physical Laboratory) p. 666

The Committee are glad to report that satisfactory progress
has been made during the year with the Ampere Balance.

The weighing mechanism was taken over from the maker
shortly after the last meeting of the Association and the work on
the coils completed at the National Physical Laboratoiy. The
labour involved in insulating the two wires on the large cylinders
was very great. Each wire consists of about ninety turns of about
103 centimetres circumference. Thus each wire is about 93 metres
in length, and the two are along their whole length about one-
tenth of a millimetre apart. In the coils as finally set up the
insulation resistance between these two wires is measured in
thousands of megohms, and is thus amply sufficient.

The cost of the balance has amounted to £302. 6«., the excess
over the £300 granted for the purpose being met out of the general
fund at the disposal of the Committee.

Calculations of the force to be expected between the coils
when carrying one ampere have been made by Mr Mather and
Mr F. E. Smith, of the National Physical Laboratory, and are in
close agreement.

The designs firom which the balance has been made are the
work of Mr Mather, and originally it was contemplated that the
balance would be set up at the City and Quilds Central Institute
in Exhibition Road. At a meeting of the Committee held on
March 31, 1905, however, this decision was modified, and the
following minute agreed to: —

That the Ampere Balance remains for the present at the
National Physical Laboratory, and that a determination of the
ampfere be made with it there under the supervision of Professor
Ayrton and Mr Mather, steps being taken to connect closely with
the determination and with any notification of the results the

PBACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 668

names of the late Professor J. V. Jones, Professor Ayrton, and
Mr Mather, to whom the design of the instrument is entirely due.

Accordingly the balance has been set up at the National
Physical Laboratory and a number of preliminary tests have been
made. Particulars of some of these follow.

Amph'e Balance, The weighing mechanism of the balance
was erected by Mr L. Oertling at the National Physical Laboratory
in March of the present year, and the electrical equipment wad
completed immediately afterwards. The four coils of bare copper
wire wound on marble cylinders have given complete satisfaction,
the ellipticity and conicality of each being very small. The
average ellipticity is such that the diameters of the ellipse differ
in length by about 10 micron, while the average conicality is
approximately represented by a difference between the mean
diameters of two sections 13 cm. apart (the axial length of one
coil) of about 12 micron. The contour of the cylinders is very
well known, and the mean diameter has been measured with a
probable error not greater than O'OOl per cent. This knowledge
enables the calculation of the mutual induction between two coils
to be made with great accuracy.

Preliminary observations with a current of nominal value
1 ampere were made at various times during April and May, the
first set of such observations determimng the temperature to which
the coils are raised by a continuous steady current, the magnitude
of the distmrbances arising from convection currents, the influence
of the silver strips, and the nature of other disturbances. The
convection currents give some trouble, but the experiments already
carried out indicate that a change of 0*001 per cent, in a current
of nominal value 1 ampere will be detectable.

The balance acts conveniently as an indicator of the magnetic
permeability of the marble and metal parts of which it is con-
structed, and it is satisfactory to know that the permeability of
these parts does not differ firom unity by a measurable amount,
and cannot therefore influence the final results.

Early observations showed that the concentric cable employed
in the leads to the balance was faulty, some of the internal strands
being broken, and the variable contacts thus resulting prevented
a steady current firom being obtained. Fresh concentric cables are
being inserted, which should enable the final observations to be
speedily made.

6i64 ' ' PRACTICAL STANDARDS

Apparatas for the determination of "jr." The three half-
second pendulums (the property of the Board of Education, and
which were used in the last Antarctic expedition) have been
swung at Kew and at the National Physical Laboratory in the
room where the balance is erected. The observations are being
continued.

When all the constants have been determined and the observa*
tions with the balance are complete it will be necessary to consider
the means by which the result is to be given to the world. The
current may be reproduced either by means of the silver volta-
meter or by means of a standard cell and a standard resistance.
The silver voltameter is being investigated at the National Physical
Laboratory, and a comparison of the accuracies of reproduction
would perhaps influence the choice.

The Committee were represented at the International Elec-
trical Congress at St Louis last year by Professor Perry and the
Secretary.

The resolutions passed at the Cambridge Meeting of the
Committee (see Report for 1904) as to certain questions proposed
for discussion were laid before the Congress, and after discussion
the following reports were unanimously accepted: —

Committee of the Chamber of Delegates on International

Electromagnetic Units,

The Sub-Committee appointed September 13, 1904, beg leave to suggest
that the Chamber of Delegates should adopt the following report:

It appears from papers laid before the International Electrical Congress
and from the discussion that there are considerable discrepancies between
the laws relating to electrical units, or their interpretations, in the various
countries represented, which, in the opinion of the Chamber, require con-
sideration with a view to secturing practical uniformity.
• Other questions bearing on nomenclature and the determination of units
and standards have also been raised, on which, in the opinion of the Chamber,
it is desirable to have international agreement.

The Chamber of Delegates considers that these and similar questions
could best be dealt 'with by an International Commission representing the
Governments concerned. Such a Commission mighty in the first instance,
be appointed by those countries in which legislation on electric units has
been adopted, and consist of, say, two members from each oountiy.

Provision should be made for securing the adhesion of other countries
prepared to adopt the conclusions of the Commission.

The Chamber of Delegates approves such a plan, and requests its members
to bring this report before their respective Qovemments.

FOB EUECtRICAL MBAdUREMBNTS 66^5

It is hoped that if the recommendation of the Chamber at Delegates be
adopted by the Qovemmenta represented the Commission may eventually
become a permanent one.

Committee of the Chamber of Delegates on Intemationcd

Standardisation.

The Committee of the Chamber of Delegates on the Standardisation of
Machinery begs to r^)ort as follows:

That steps should be taken to secure the co-operation of the technical
societies of the world by the appointment of a representative Commission
to consider the question of the standardisation of the nomenclature and
ratings of electrical apparatus and machinery.

If the above recommendation meets the approval of the Chamber of
Delegates it is suggested by your Committee that much of the work could
be accomplished by corredpondence in the first instance and by the appoint-
ment of a Ceneral Secretaiy to preserve the records and crystallise the points
of disagreement, if any, which may arise between the methods in vogue in
the different countries interested.

It is hoped that if the recommendation of the Chamber of Delegates be
adopted the Commission may eventually become a permanent one.

The first of these Reports, relating to the summoning of an
International Congress on Electrical Units, is now under the con-
sideration of His Majesty's Qovemment. Meanwhile a preliminary
conference of representatives of standardising laboratories and
others interested in the determination of electrical units has been
summoned by the President of the Reichsanstalt to meet in Berlin
in the autumn. Lord Rayleigh and the Secretary have received
invitations to be present.

The object of this Conference is stated to be that the institu-
tions which are concerned in maintaining the accuracy of electrical
measurements in conjunction with those scientists who have
devoted especial attention to this field of work should exchange
opinions, and if possible come to an agreement as to the measures
which must be taken in order to obtain the international uniformity
in electrical units and measurements which is desired. It is thus
preliminary to the more formal consideration of the subject which
would be the work of the International Congress.

One of the important questions which will be discussed will be
the specification of some form of standard cell. Work on this
matter has gone on in America and at the National Physical
Laboratory, and an Appendix to the Report by Mr Smith contains

666 PRACTICAL STANDARDS

a provisional specification. It is suggested that persons in-
terested in the matter might help by setting up cells in accordance
with this specification and submitting them for test at the
laboratory.

Of the grant made to the Committee in 1904 a balance of
£3. 4i8. lOd. remains.

The work which remains to be done on the standard cell, and
with the Ampere Balance, will involve considerable expense, and
to meet this the Committee ask for reappointment, with a grani^
of £25 in addition to the balance now in their hands.

Appendix.

On the Preparation of a Cadmium Cell*. By F. E. Smfth.

{From the National PhywsaJL Laboratory,)

The research on standard cells has been continued at the
National Physical Laboratory on the lines indicated in the last
Report to the Association. Taken as a whole, the results are very
satisfactory, but it is thought desirable to still continue the
observations on some of the older cells before publishing the
results in detail.

Mr G. A. Hulett, of Michigan, has completed a chemical re-
search on mercurous sulphate, which throws considerable light on
the anomalies reported to the Association last year. Veiy slight
changes can still, however, be traced to this salt, but fortunately
they are of no commercial significance. The 12^ per cent,
amalgam also produces slight variations in the E.M.F. of the cell :
these again are commercially unimportant, and a manner of over-
coming them in cells employed at a standard laboratoiy is indicated
in this paper. The latter cells are set up with an amalgam en-
tirely liquid at 0°C.

At this stage of the research it is thought desirable to describe
the methods by which the materials of the cell can be best pre-
pared in the light of present information, and an appeal is made

•

* Speoifioations for the preparation of the Weston (or oadmium) ceU are alsa
given in the Reports for 1907 and 190S. In 1905 the Weston oeU was naed but
little in Great Britain and the notes on its preparation proved to be of oonnderaUe
valae. The changes introduced in 1907 were based on experiments made at the
National Physical Laboratory, and the specification printed in the Beporfc for 190S
was drawn up after consultation with many experimenters. The three specifieationa
are therefore of some historical interest.

FOR ELECTRICAL MEASUREMENTS 667

to those interested in the subject to set up one or more cells by
these methods and submit them for comparison with the standards
of the National Physical Laboratory. More light will thus be
thrown on the slight discrepancies already referred to, and the
degree of accuracy with which the cell can be constructed will be
established. In this way it is hoped to specify a cell for commercial
purposes accurate at all ordin€u*y English working temperatures
to 1 part in 2000, applying no temperature correction, or to 1 part
in 10,000 if the temperature correction be applied.

In the specification which follows there are four methods of
preparing the mercurous sulphate. The first of these is due to
Professor H. S. Carhart, Mr G. A. Hulett and Dr Wolff, jun.;
the main features of the second method were suggested by
Mr Swinburne to Dr Glazebrook, while the third and fourth methods
have resulted from some experiments at the National Physical
Laboratoiy. It is suggested that one only of these methods be
eventually employed; the observations on submitted cells will
largely determine the choice.

Preparation of Materials for a Standard Cadmiwm, CdL

1. Mercury. The commercial mercury should be squeezed
through wash-leather and passed in the finely divided condition
in which it emerges, first through dilute nitric acid (1 to 6 of
water) and mercurous nitrate solution, and afterwards through
distilled water, both liquids being conveniently contained in long
glass tubes. The mercury is then to be twice distilled in vacuo.
Mercury suspected of abnormal contamination should not be
employed.

2. Amalgam, T^jf^ A, This is a 12^ per cent, amalgam,
and is intended for all commercial cells. The method of pre-
paration is practically identical with that used by Professor Carhart.
A current is passed from a thick rod of pure commercial cadmium
to distilled mercury, the intervening liquid being cadmium sul-
phate solution rendered slightly acid with a few drops of H,S04.
The cathode is weighed before deposition takes place, and again
afterwards, the percentage of cadmium in the amalgam being
thus calculable. More than the requisite amount should always
be deposited, and the percentage reduced to 12^ by the addition
of more mercury. The fall of potential from anode to cathode
should not exceed 0*3 volt. To prevent the anode slime having

668 PRACTICAL SrA^DARI)S

access to the cathode it is desirable to surround the anode with a
small porous pot, as in the Richards' silver voltameter, or to place
a small crystallising dish beneath it for the anode powder to
settle in; Contact with the cathode is made with a platinum
wire sealed into a glass tube so as to protect it from direct contact
with the cadmium sulphate solution, and a rough estimate of the
quantity of cadmium deposited is obtained from the readings of
an ammeter placed in the circuit. The amalgam so prepared,
together with the mercury added to reduce the percentage of
cadmium to 12^, is now heated on a water-bath and stirred so
as to ensure homogeneity, some cadmium sulphate solution still
flooding the surface. It is then cooled, and the acid sulphate
removed) neutral cadmium sulphate solution taking the place of
the latter, and consisting of saturated solution plus an equal
volume of distilled water. This 12^ per cent, amalgam is then
ready for use and is entirely liquid at a temperature approximating
to 60** C.

Type £, This amalgam is liquid at the temperature of melting
ice, and is intended for cells of a slightly better type than those
made with the 12^ per cent, amalgam. The cells may be used
at a higher temperature than C C., but they are not intended to
be so used as their temperature coefficient is about — 0'043 per
cent, per rise in temperature of 1*'C. The cells are primarily
intended for standardising laboratories, and their km.f. at 0"" C. is
equal to the E.Bf.F. of the cells prepared with the A amalgam if
this latter E.M.F. is corrected to 0** C. with the temperature coeffi*
cient formula of the cell. This is equivalent to saying that if an,
A cell was in a steady condition at 0"* C. and nothing abnormal
occurred its E.M.F. would be identical with that of a jS cell at
0° C. It is not wise, however, to use a 12^ per cent, amalgam
cell at low temperatures ; an 8 per cent, amalgam may be so used,
but its upward range (with a small temperature coefficient) is
lower than that of the 12^ per cent, amalgam cell. For com-
mercial purposes probably the 12^ per cent, amalgam will be of
most service.

To prepare the tjrpe B amalgam take some of that previously
prepared and add sufficient mercury to reduce the percentage of
cadmium to 3. The amalgam will now be entirely liquid at
ordinary working temperatures. On cooling a crystalline amalgam
separates from the liquid, and will continue to do so as the

FOR SLECTRICAL MEASUREMENTS 6,69

temperature is lowered. Cool the amalgam to the temperature of
melting ice and remove the mother liquid : this is the amalgam
desired. It is important that the temperature be truly that of
melting ice, and that no solid is removed. For convenience the
3 per cent, amalgam may be placed in a tubular vessel well sur-
rounded with ice shavings ; a long very fine capillcuy tube reaches
to the base of this vessel, and through it the liquid at 0°C. is
removed by suction. Some solid must be left behind, or otherwise
there is no certainty of saturation. Throughout all the operations
neutral cadmium sulphate solution must cover the surfietce of the
amalgam and wet all vessels, tubes, etc. through which the amalgam
passes. Otherwise the amalgam will leave a " tail " and its com-
position may possibly be thereby changed.

3. Cadmium SiUphate Crystals and Solution. Procure com-
mercially pure cadmium sulphate CdS04 . 8/3HsO. Dissolve in
about 1^ times its weight of distilled water, agitating either con-
tinuously for about six hours or occasionally for two or three
days. Filter through a fine grained filter-paper so as to ensure a
clear solution, which should then be placed in a number of crjrstal-
lising dishes and evaporation allowed to take place slowly at a
temperature not exceeding SS^'C, when, provided that dust be
excluded, many transparent crystals of CdS04 . 8/3H,0 will result.
These should be prevented as much as possible from adhering to
one another by removing the liquid to other dishes as soon as the
crystals are of such a size that most of them are in contact. In
this way about five-sixths of the liquid may be evaporated, the
mother liquid being employed for a preliminary washing of the
mercurous sulphate, the manufacture of which is afterwards
described. The crystals of cadmium sulphate so obtained should
be washed with successive small quantities of distilled water until
after standing for five minutes no trace of acidity can be detected
with Congo red. The crystals, still moist, may then be transferred
to a stock bottle. To prepare the final solution agitation with
distilled water is recommended as before, the temperature being
preferably 5° or 10** higher than the normal temperature, so as to
ensure saturation. On no account should cadmium hydroxide be
employed to neutralise the first solution, which is invariably acid ;
nor indeed should any attempt be made to neutralise the solution
except by crystallisation.

4. Mercurous Sulphate. The preparation in each case is to be
conducted in a darkened room.

670 PJEtACnCAL STANDARDS

(a) Electrolytic Method. Pure distilled mercury forms the
anode and platinum foil the cathode, the electrolyte being dilute
sulphuric acid (1 part by volume of concentrated acid to 5 parts
of distilled water). The mercury is preferably placed in the base
of a large fiat-bottomed beaker and about twenty times its volume
of the dilute acid added. Contact with the mercury is effected
by means of a platinum wire passing through a glass tube, while
the cathode is suspended in the upper portion of the liquid.
During the electrolysis the electrolyte must be continually stirred,
an L-shaped glass stirrer being most efficient, the L portion being
placed near the surface of the mercury. A current density of
about 0*01 ampere may be employed. The salt so prepared is
treated as per Note A-

(6) Precipitation method, mercurous nitrate and sulphuric
acid being employed.

Add strong nitric acid to a little pure mercury contained in a
crystallising dish and place in a draught chamber until the action
is over. If any mercury remains add more acid and continue to
do so until the mercury has completely disappeared and a strongly
acid solution assured. Prepare dilute HjS04 (1 to 4 by volume),
allow to cool, and then add the acid nitrate solution drop by
<lrop, keeping the mixture violently agitated. Mercurous sulphate
is precipitated, which should be filtered and treated as per Note A.
No more nitrate solution must be added to the dilute H^04 than
will suffice to neutralise 30 per cent, of the H^04 present The
maximum amount it is permissible to add may be estimated by
taking a small portion of the dilute H^SOa and adding the nitrate
solution until no further precipitation results The proportion of
nitrate solution to dilute HaS04 in such circumstances must be
reduced to one-third its value for the preparation of mercurous
sulphate by method (6).

Purchased mercurous sulphate is warmed with strong HbS04
and a little mercury to a temperature of about 150° C. for about
ten minutes, the operation being conducted in an evaporating
dish covered with a clock glass and the mixture kept well stirred.
The suspended matter is then allowed to settle, the hot liquid
cooling sufficiently meanwhile for the vessel to be handled with
comfort. The clear acid is then poured into dilute HJSO4 (1 to 6),
when crystalline mercurous sulphate separates out. About ten

FOR ELECTRICAL MEASUREMENTS 671

times the bulk of dilute acid should be employed, and to avoid
spitting the hot liquid should be poured through a funnel, having
its stem immersed in the dilute acid. The mixture is well stirred,
cooled, and filtered, and the salt treated as per Note A. As the
operation yields but a small quantity of the salt it is advisable
to repeat several times.

(d) By means of Nordhausen sulphuric acid.

Place distilled mercuiy in the bottom of a beaker or bottle to
the depth of about 3 mm. Add about four times its volume of
Nordhausen sulphuric acid and stir well, keeping the mouth of the
bottle closed as much as possible, as the acid fumes are very
unpleasant. Mercurous sulphate is formed in the cold and appears
in the crystalline form after a few minutes. Allow the operation
to continue until the acid strength has been considerably di-
minished ; warm the product to expel SO, and add to dilute HaS04
(1 to 6). Considerable spitting always occurs, so that caution is
necessary. Proceed with the product as per Note A. •

Note A. The mercurous sulphate obtained by the foregoing
methods is first agitated with dilute EL^SOa (1 to 6) and distilled
mercury. It is then filtered (a small Gooch crucible and filter
flask are convenient), and the greater part of the mercury removed
as it interferes with the filtering. The salt is next washed with
small quantities of saturated cadmium sulphate solution until
free from acid. For the first few washings some of the first acid
solution may be employed, but the final washings must be made
with a little of the neutral solution. Trouble is often ex-
perienced in ridding the salt prepared with Nordhausen sulphuric
acid from all trace of acidity, and it is preferable to wash five or
six times with the cadmium sulphate solution, and then place in
a bottle together with a little of the solution, shaking from time
to time and filtering again in a few days. The acidity of the
washing liquid should be tested with Congo red. Instead of
washing with cadmium sulphate solution, sulphuric ether (water
free) may be employed.

The Mercurous StUphate Paste. Some cadmium sulphate
crystals are ground in an agate mortar with a little cadmium sul-
phate solution ; about one-quarter their bulk of pure mercury is
then added and two volumes of the acid-firee mercurous sulphate,
the whole being well mixed with cadmium sulphate solution so as
to form a thin paste.

672

PRACmCAL STANDARDS

The Form of Cell, The H form of cell due to Lord Rayleigh
is the most convenient^ and is in general use. Two patterns have
l;>een adopted. In fig. 1 a form is shown in which the electrodes
are sealed into the lower ends of the two vertical limbs, while in
the form shown in fig. 2 the electrodes pass through glass tubes
into the lower ends of which they are sealed. Form 1 can be

Fig. 1.

Fig. 2.

V«^ «M •

AT ;= Mercury.

A ss Amalgam.

P := Paste.

C s Cadmium sulphate crystals.

S = Saturated solDtlon of cadmiam

sulphate.
JC=Cork.
O = Marine Olae.

hermetically sealed, and is intended to be immersed in an in-
sulating liquid. Form 2 is sealed with marine glue, and may be
immersed in ice or water. The hermetical sealing of form 1 was
suggested by Lord Rayleigh * and later by Professor Carhart f- The
glass tubes through which the electrodes are introduced in form 2
pass through corks which have been previously boiled in vrater
and soaked in cadmium sulphate solution ; in addition to the hole
which allows of the passage of the electrode, a second hole is bored

* Phil, Tram, 176, §42, 1886. f St Louis Congress, 1904.

FOR ELECTRICAL MEASUREMENTS 673

through these corks for the passage of small glass pipettes. After
the cell is filled these additional holes are fitted with small corks,
and the cell is finally sealed with marine glue. The position of
the various parts is shown in the figure. (Both forms of glass
vessels are stocked by Mr A. C. Cossor, of 54, Farringdon
Road, E.C.)

In filling the vessels it is convenient to use small pipettes
made of two glass tubes, the one about 3 inches long and ^ inch
in diameter, and the other about 2 inches long and ^ inch in
diameter. If the larger tube has one end drawn out in the form
of a cone, a junction is easily made. The amalgam of type A is
melted over a water-bath (its surface being flooded with dilute
cadmium sulphate solution), and is introduced by means of a
previously warmed pipette into one of the limbs. After the
amalgam has solidified, this limb should be washed out with a
little firesh cadmium sulphate solution. If the amalgam of
type B is used this washing is unnecessary. The mercury is
next introduced into the other limb, then the paste, using if
necessary a tiny glass rod as a piston in the pipette, and
afterwards a thick layer of finely pounded crystals is introduced
into each limb. Saturated cadmium sulphate solution is finally
added. The cells are then to be exposed in a warm room for a
week or more to allow some of the liquid to evaporate, and so
loosely cement together the fine crystals. This crystalline plug
keeps the contents in their proper places^ and enables the cell to
be transmitted through thie post. The sealing of the cells is next
completed, care being taken not to abnormally heat the contents.

Cells which are submitted for comparison with the standards
of the National Physical Laboratoiy should be accompanied with
the following particulars : —

1. Maker's name and address.

S. Name* of the firms firom whom the chemicals used in the manufacture
of the materials were purchased.

3. Number of the method employed in the manufacture of the mercurous
sulphate.

4. Type of cadmium amalgam used.

5. Notes on any peculiarities observed in the preparations.

* This information is only required bo that the number of different sonrces of
the materials can be estimated.

B. A. 43

THIRTY-THIRD REPORT— YORK, 1906.

Appendix. On Methods of High Precision for the Comparison of
Resistances. By F. £. Smith. {Irom the National Physical
Laboralory) p. 676

In the last Report reference was made to a conference of
representatives of standardising laboratories which had been
invited to meet in Berlin as a preliminaiy to the more formal
Conference on Electric Units suggested at St Louis.

The question of this preliminary Conference was brought
before the Committee at a meeting on October 19, 1905, and
attention was called to the importance of Clause (2) of the pro-
visional programme, viz, —

" Shall the three units, the Ohm, Ampere, and Volt, be defined
independently, or shall only two be defined, and, if so, which ? "

and it was agreed unanimously that two units should be defined
independently, and that these two should be the unit of resistance
and the unit of current. The Secretary was instructed to report
this to the Conference at Berlin.

This Conference took place in October last at the Beichsanstalt
in Charlottenburg, and was attended by representatives fi:om
America, Austria, Belgium, England, France, and Germany.

Agenda prepared with great care by the President of the
Reichsanstalt were veiy carefully discussed, and, as a result, the
Conference expressed the wish that an International Convention
should be summoned in order to arrive at agreement in the electric
standards which are in use in the different countriea

The following resolution was further adopted : —

" In view of the &ct that the laws of the different countries in
relation to electrical units are not in complete agreement, the
Conference holds it desirable that an official conference should be
summoned in the course of a year with the object of bringing
about this agreement."

PRACnCAL STANDARDS FOB ELECTRICAL MEASUREMENTS 675

The Conference further expressed the opinion : —

1. That the information before it is not sufficient to enable
it to propose any alteration in the formerly accepted value for the
ampere.

2. That the information before it is not sufficient to enable it
to lay down exact directions in respect to the silver voltameter and
the standard cell

3. That if a proposal for a change in the accepted value of
the ampere is to be brought from any source before a formal con-
gress to be held later, an agreement in writing on the point should
be come to previously between the parties interested. If differences
of opinion in the matter cannot be removed, a new preliminary
conference should be held.

The same procedure should be observed in regard to the speci-
fication for the silver voltameter and the standard cell, in the event
of such specifications being submitted to a formal conference from
any quarter.

The following formal decisions were agreed to : —

1. That only two electrical units shall be chosen as funda-
mental units.

2. The international ohm, defined by the resistance of a
column of mercury, and the international ampere, defined by the
deposition of silver, are to be taken as the fundamental electrical
units.

3. The international volt is that electromotive force which
produces an electric current of one international ampere in a con-
ductor whose resistance is one international ohm.

4. The Weston Cadmium Cell shall be adopted as the
standard cell.

Recommendations were also made as to realising the ohm, and
some particulars as to the Cadmium cell were agreed upon.

These results were laid before the Board of Trade, and a
Departmental Committee, of which the Secretary was a member,
drew up a report recommending that an official conference
should be invited to meet in London, and it is understood that
negotiations are now on foot with a view to summoning such a
conference.

During the year the work in connexion with the absolute
ampfere balance has been in progreas. and is practically complete.
Under the supervision of Professor Ayrton and Mr Mather a large

43—2

676 PRACTICAL STANDARDS

number of determinations have been made, and are most satis-
factory. Detailed particulars are reserved until the work is
complete; but there is little doubt that the balance is a most
excellent absolute instrument, and that the probable error of a
determination of current by means of it is only a few parts in
100,000.

The investigation of the silver voltameter has been extended
beyond the limits originally thought to be necessary. The results
so &r obtained are very valuable, and appear to indicate that a
satisfactory form of silver voltameter is realisable. It is hoped
that the publication of the results will take place at the same
time as those of the ampere balance.

An appendix by Mr F. £. Smith describes the methods of
comparing resistances in use at the National Physical Laboratory,
and discusses the sources of error and the accuracy attainable.

The grant of £25 made in 1905 has been expended in
materials for the work on the ampere balance and the silver volta-
meter. In connexion with the latter a large amount of work
involving considerable expense remains to be done. For this
purpose the Committee ask for reappointment with a grant of £50.
They recommend that Lord Rayleigh be Chairman and Dr & T.
Glazebrook Secretary.

Appendix.

On Methods of High Precision for the Comparison of Resistances.

By F. E. Smith.

{From th€ National Physical LahorcOory,)

The object of the author is to give a brief account of the high
precision methods used at the National Phjmical Laboratoiy for
measuring standard resistances. Up to and including the year
1903, the standard unit coils of the British Association were com-
pared by Carey Foster's method, the Fleming circular wire bridge
being used. The probable error of such comparisons is of the
order O'OOl per cent. The build-up of a 10-ohm coil bom the
unit was very conveniently effected by a process suggested by
Loid Rayleigh*. Three 3-ohm coils are arranged in parallel, and

* Phil, Traru, 18S8, 174, 810. See also B. A. Report, 1888.

FOR ELECTRICAL MEASUREMENTS 677

their combination value determined by comparison with a unit
resistance. They are then placed in series ; by the addition of a
unit coil to the series formation, the " build-up" is complete. The
probable error of this build-up is also small, but when combined
with the error of comparison of nominally equal coils, the observed
value of a 1 to 10 ratio may be in error by 0002 per cent. The
use of this ratio for the evaluation of resistances of 10^ units
results in a possible error of n x 0*002 per cent.

The resistance standards of the National Physical Laboratory
are of three kinds — mercury, platinum-silver, and manganin.
When comparing standards of mercury and of platinum-silver,
comparatively small currents must be employed, because the tem-
perature coefficients of these materials are large and the resistances
are surrounded by bad thermal conductors. The manganin coils
are wound on brass cylinders, have small temperature coefficients,
and may be immersed in oil; the maximum permissible current
is therefore much greater. The accuracy of all methods of com-
parison is directly proportional to the current employed, from which
it follows that for all building-up processes, manganin coils are to
be preferred. The question of preference for permanency is not
discussed in this paper.

In order to compare the various methods of measurement it is
necessary to give the formulae for sensitiveness. In presenting
these latter I do not wish to suggest that they are new. The
subject has been previously treated by Mr 0. Heaviside*, Mr T.
Gray f. Lord RayleighJ, Professor Schuster^, Professor A. Gray||,
Dr Jaegerf , Dr St Lindeck, Diesselhorst, and others, and some of
the formulae are given in text-books. In the present paper the
considerations of many of these writers have been extended.
Professor Schuster first pointed out that it is the heating of the
conductors which puts the limit to a measurement of resistance,
and the formulae derived by him are in terms of the current con-
veyed by the resistance to be measured. Dr Jaeger has recently
discussed the question of sensitiveness from the same point of
view, and in this paper the subject is similarly treated. The

* PML Mag, 1S78, XLV. p. 114. f Ibid. 1881, xn. p. 288.

:: Proc. Roy. 8oe. 1891, 49, 208. § PhU. Mag. 1894, p. 176.

II Abiolute MeasuremenUf vol. i. p. 881.

t ZeiUchr, IrutrumenUnk. Biaroh 1906, 86, 69. See also Jaeger, St Lindeok,
and DieeseUiorat, Zeitsehr. Instrumentenk. 1908, 88.

878

PRACTICAL STANDARDS

formulae may be derired in several ways, as will be seen on
reference to the authorities quoted Many of these ways are long,
and it may not be out of place to give a well-known rule, which,
if applied to any sjrstem of conductors, will quickly give all the
desired information.

" In any network of conductors the current in one arm due to
an electromotive force in another arm is equal to the current id
the latter when an equal E.M.F. is placed in the former."
(This rule results from an application of Eirchhoff's Laws.)
The most complicated system of conductors considered in the
present paper is that known as the Kelvin double bridge*, and this
is dealt with here by way of example. Let the current through P
be t, and through -K, i\ and let P/Q^^R/S^a/fi. Also let the

Fig. 1.

applied E.M.F. remain constant. On completing the galvanometer
circuit the distribution of the currents will remain unaltered
Let P be changed to P + SP. The current through it will change
to i — Si, and the change in p.D. of P is %BP — PBi ; of Q it is QSi.
If the galvanometer circuit is now completed, the current through
it will be equal to that produced by an G.M.F. iBP — PSi in P and
an E.M.F. equal to QBi in Q. If an E.M.F. equal to the latter is
placed in the galvanometer branch, the current through Q is
PQSif(P + Q)r, where r is equal to

g^ (P + .fi)/(Q + g)
a + fi^ P + R^Q + S "*"^'

i.e. the resistance of the " external galvanometer circuit" plus that
of the galvanometer. Similarly the current through P due to an

* W. Thomsoo, PhiL Mag. 1862, 24, 140.

FOR ELECTRICAL MEASUREMENTS 679

RM.F. PS* in the galvanometer branch is equal to QPS%/{P + Q) r.
Hence, by the rule, the current through the galvanometer due to
an E.M.F. QBi in Q is equal to the current through the same due to
an E.M.F. P8% in P. As these must be in opposite directions
through 0, we have only to consider the current due to an E.M.F.
iSP in P. The current through 0 due to this E.M.F. is found in a
similar manner and is equal to

gff (F + It)iQ-^S)'P + R + Q + S ^^^

^■^a + ^'^'P+ii + G+fif

This, therefore, is the current through the galvanometer when the
balance of the bridge is disturbed by an alteration in P of SP,

In galvanometers, the coils of which are wound in similar
channels, and contain the same mass of wire, the electromagnetic
force on the needle, and hence the deflection, is proportional to

X V&*, where x is the current through 0. In the case considered
the deflection is proportional to

ygtsp Q+s ,^^

^^a + fi^ P + R + Q + 8

1 his IS a maximum when (/« — r~5 + ^ — p \X — ^, %.e. the

ft + P i^+/v + y + o

resistance of the " external galvanometer circuit," and the value

of this is the most suitable galvanometer resistance. Substituting

this value for 0 in (B), an expression is obtained which, from the

conjugate condition of the arms of the bridge, may be reduced to

the simple form

iA VP/2 y^M^A±«) ((7)

in which A = SP/P.

If in (B) we write g for the best galvanometer resistance and
Ng for the resistance of the galvanometer used, the deflection is

proportional to ^/Ngl(N + 1) ^/g, and the ratio of this to the maxi-
mum (N-l) is 2V^(iV+l). Prof. Schuster, in the paper
referred to, gives a table showing that if N^20 or 0*05, the
sensitiveness is 0*426 times the maximum.

The derivation of the formulae being so simple, the results
alone are given for the other methods considered.

* Absolute MeasuremenUf A. Graj. vol. n.

680

PRACTICAL STANDARDS

Wheatstone Bridge (fig. 2). — If a ^13^0 in the expressions
obtained for the Kelvin double bridge, the values are those for
the Wheatstone bridge*. In this case, expression (C) may be
written

»AVP/2y(l+|)(l+|). (D)

The best conditions for sensitiveness are here clearly indicated.
The resistance R should be small compared with S and with P,
%.e. P should be connected to a comparatively large resistance Q
and a small resistance R, If i is the maximum permissible
current through P, Q must be a resistance of large cooling surface
and small temperature coefficient ; if it is of the same type and

Fig. 2.

dimensions as P, then it should be of the same nominal value.
In the latter case, which is the general one for precision measure-
ments, P = Q = iJ = S, and the sensitiveness is proportional to

tAVP/4. It is generally recognised that for coils of the same

type and dimensions iVP is constant.

The Potentiometer (fig. 3). — Let the resistances of the two
circuits be P + i2i and Q + R^* If i is the current through P, the
current through the galvanometer is

iAP

0 + PR,I{P + R,) + QR,I{Q + R^) '

* The Talues usually given for the Wheatstone bridge (see J. J. Thomson,
Elementi of Elec. and Magnetitm, p. 805 ; Fleming, Handbook of Elec. Laboratory,
▼ol. I. p. 238 ; A. Gray, Ab$. MeasuremenU, toI. i. p. 888) inyoWe the resistanoe of
the battery arm and the e.h.v. of the battery. If, for the latter i(P+Q) is sub-
stituted, the resistance of the battery may be taken as zero, and on substituting,
the value given in this paper is obtained.

FOR ELECTRICAL MEASUREMENTS

681

and the best resistance for the galvanometer is

PR,I{P + A) + QR^{Q + 12,).

The sensitiveness is therefore proportional to

2ViJ,/(P + ii,)+giVP(Q + i2,)'
In the case of precision measurements, Ri and B^ may be made
very great compared with P and Q respectively. If this is so, the

sensitiveness is proportional to iA VP/2 Vl + QjP. If Q is small

compared with P, this becomes t'A VP/2, and the best resistance
for the galvanometer is P. Unless P and R are nominally equal

Fig. 3.

the galvanometer resistance cannot be the most suitable for both
observations, and the sensitiveness of one of the measurements
must be less than that stated. If P = JR and Q « fif, the latter
being comparatively small, the sensitiveness is twice that of the
Wheatstone bridge with equal arms. It has to be remembered,
however, that the current in the potentiometer is continuous and
the heating effects more marked than in the bridge in which a
tapping current only is employed. A great practical advantage of
the bridge method is the rapidity of measurement.

Differential Galvanometer Method (fig. 4). — ^If 0 and g are
the resistances of the galvanometer coils, the difference of the
currents through them is % {Pg - Q6)/6 (Q + flr). If P = Q and
O^g, the difference of the currents is

682

PRACTICAL STANDARDS

and the best galvanometer resistance is G » P » Q. The sensi*

tiveness is then proportional to iA VP/2 V2. If the currents
through the galvanometer are comparatively large, convection
currents are produced in the space containing the suspended
magnets ; also, the resistance of the coils is subject to small but
rapid changes. There is, therefore, a maximum permissible value
for the currents through the galvanometer coils, and in general
some ballast resistance must be added to the galvanometer arms.
This reduces the sensitiveness.

— ^

Fig. 4.

0©

■n/\/\/^\/\/^

Mercury Standards of Resistance, — The Eohlrausch differential
galvanometer (see p. 696), the Kelvin bridge, and the potentio-
meter have been employed* for the measurement of resistance of
mercury standards with current and potential leads of compara-
tively high resistance. These methods are recommended in the
Report of the Conference on Electric Units at Charlottenburg
(1905). The current used in the measurement of such resistances
is limited by the condition that the mercury shall not be sufficiently
warmed to produce appreciable error.

In the Standards Department of the National Physical Labora-
tory no favourable opportunity has arisen for an exhaustive test of
the Eohlrausch method. As used at the Phjrsikalisch-Technische
Beichsanstalt it is very satisfactory; but, strictly speaking, it is
not a null method, as observations of deflections have to be made.
From particulars published f, a favourable arrangement for the
measurement of mercury standards is when 0 = g^6 ohms,
P = Q ss 1 ohm, and the ballast resistance in each galvanometer
arm is 10 ohms. In this case the sensitiveness is proportional to

t'A V12/34 = 0098iA.

With the Kelvin double bridge, if JB = S«1000, P«Q = 1,

* WUsen»e?iaft. Abhand, d, Phy$,-Teeh. Reichsanstalt, 414, Band n. ; see alao
Phil, Tram, 1904, A, 878, 67.

t Wiigefuchaft, Ahhand, d, Phy$,-Tech. Reiehsanstalt^ Band in.

FOR ELEGTRIOAL MEASUBEMENTS

683

a S3 /3» 100, and (7 a 1000, the sensitiveness is proportional to
OOlliA. If E = flf=100, P«Q = 1, a « /8 = 100, the sensitive-
ness is more than doubled, being equal to 0'025 tA. This latter
case is convenient in practice.

In the potentiometer, if P = Q, and O^P-hQ, the sensitive-
ness is proportional to 0'35tA. The current is continuous, and
hence the maximum permissible value of t is not so great as with
the differential galvanometer and Kelvin bridge.

At the National Physical Laboratory the Kelvin bridge and
the potentiometer were employed up to March of the present year.

Fig. 5.

With the former method a tapping current of 0*2 ampere was
necessary in order to measure a difference of 1 x 10~~* ohm with
certainty. With the latter method the current used was 0*08
ampere, but the method was &,t less convenient. At the present
time a modification of the Wheatstone bridge is used, and proves
to be the most satis&ctory and most sensitive of all the methods
discussed. The arrangement is very similar to that suggested by
R H. Housman for the evaluation of small resistances (p. 691).
In fig. 5, P is the mercury standard, of which r and / are the
current leads. R principally consists of a 1-ohm manganin coil
which is shunted with a resistance X, usually of the order 30 to
60 ohms, and a resistance X' of several thousands of ohms.

684 PRACTICAL STANDARDS

The latter is varied in the final adjustment of this arm of the
bridge so as to obtain a very accurate balance. Q and S are
1000-ohm coils of manganin. R' consists of two unit coils in
series; the value of these in terms of other unit coils is known
with great accuracy (see build-up method, p. 688). £ is a thick
copper conductor in series with 8. The current through P is 0*03
ampere. The operations are as follows: The bridge-piece B is
placed in position so that R is out of circuit, and the shunts X
and X' are adjusted until

R,/P^{8 + B)l{Q^r\

Ri being the shunted value of R. The galvanometer lead at a is
removed and connected to b, and the battery lead at c is placed
at a. In practice this change is effected with a rocking com-
mutator. The position r){ B ia altered so as to include B^ as one
of the arms of the bridge, and a balance is obtained by shunting
-B', when, if B/ represents the shunted value of B\

R,'/(P + iJ,) = (S -h B)I(Q + B).

Combining this with the previous equation, we have

B,' = P[(8 + B)(Q^S + r^B)l(Q^ry].

The value of r is obtained with considerable accuracy by
moving the galvanometer lead At e to d and balancing. In an
analogous manner the value of B may be obtained ; the correction
due to i3 is usually less than 1 part in 10,000,000. The ratio of
8 to Q may be eliminated from the last equation by inter-
changing Q and 8 in the bridge and repeating the operations
indicated above. If Q and 8 are not very different from their
nominal values, then

i?/+iJ,' = P[4-6(r-5)/Q].

where B^' represents the second shunted value of R,

With a galvanometer resistance of 2 ohms, with P= jB = l,
and Q = 8^1000, the sensitiveness of the arrangement is pro-
portional to 0-35 %A. With P = i2« 2, and Q = S = 1000, this is
increased to 0*47 lA, the values for i being the same in the two
case& If a greater current value than 0*03 ampere is permissible,
then Q = S may be made equal to 100 ohms, and the increase in
sensitiveness is approximately proportional to the increase in the
current.

The following observations were made on May 30, 1906, the

FOR ELECTRICAL MEASUREliENTS 685

mercury standaxd, F, being used, and two coils in series (Coil 1)
evaluated : —

1st observation, P^Y= l-00027o int. ohms Q= 1000-18 approx.

5=100019 „
Z = 401
Z' = 30900

Shunt on -B' = 31600. r = 0033. B = 000007

2nd observation, Q and S interchanged

Z=401

Z'= 28400
Shunt on 12' » 16400.

Hence iJ^' + jB,' = l-00027o [4-6 (00000329)]

= 400088,,

.-. 212' = 400088« + 2 (1/31900 + 1/16400),

.-. Coil 1 = E' = 2000534 int. ohms, t « 1721" C.

Comparing the various methods as practically employed, the
sensitivities are proportional to

0'025tA for the Kelvin double bridge.

0*098 1 A „ Kohlrausch differential galvanometer.

0*35 lA „ Wheatstone bridge.

0*85 tA „ Potentiometer.

The maximum permissible values of i are the same for the first
three methods. For the potentiometer a smaller current must be
used. Possibly the arrangement considered for the differential
galvanometer might be modified so as to make the method more
sensitive.

Oomparison of Unit Coils. — Manganin coils with potential
leads are alone considered. Platinum-silver coils without such
leads are compared with manganin ones by substitution in one of
the arms of the bridge.

The method adopted is analogous to that of Carey Foster.
The coils are exchanged in position, but the difference of values is
given by the shunts applied to the two ratio coils. Thermal
E.M.F.'s are small, and produce no disturbing effect as the galvano-
meter circuit is continually closed. The self-induction of the coils
is very small indeed.

For coils having potential leads the Kelvin double bridge is
used. P and Q are the coils to be compared, Pjj, P^, Qjj, and Qi^

686 PBACTIOAL STANDARDS

being the resistances of the current leads of these coils. It and
S are 1-ohm standards, a = /8 = 1 ohm and Qr + Pl = d. The
galvanometer is permanently connected as shown in fig. 6, but the

Fig. 6.

battery leads are successively joined to the junctions of P ' Pj^ and
Q'Ql, P'P^ and S'Ql, RPr and QQl^ The coils iJ and iSf
are shunted to effect a balance. Representing the shunted values
of R and S by R^R^R^, SiS^St, etc., we have

m p-Q(^'^Pr) , ^^ (IJi + Pr «\

^^ ^"* S,^Qi, '*"a + )8 + dUi + (2L fir

/ox p_(Q + QL)(JZ. + fj^), dfi (R. + Pn a\

(2) r ^ +_^__|^____-j.

(3) p- Q^ p , dfi f R, a\

^ ^ 8,+ Qj^ ^^'^a + zS + dU + OL fi)'

In practice, the value of d/3l{a + fi-\-d) does not exceed 0-00006
ohm, and the expression accompanjring this is normally of the
order 0*00005 ohm, so that the last term in the above equations is
negUgible. From (1) and (2) Qr, « {R,/8i - iV'8^,)/2, and firom (1)
and (3) Pj, = (R,I8, - iJ,/S0/2.

P and Q are now exchanged in position, when

Q^PiR^ + QMS^-^Pnl

the values of Qi and Pj^ being determined as before. If the coils
are not very different from their nominal values we may now write

P - <2 = i [(2J, - 12,) + (S« - S.) + 2 (P, - <2i)].
a difiPerence readily determined from the shunts employed. Witii

FOR ELECTRICAL MEASUREMENTS

687

a galvanometer resistance of 3 ohms the sensitiveness is pro-
portional to 0'20tA VP. For coils without potential leads, in
which case the method of comparison is simplified, the sensitive-
ness is 0*25 iAVP, the same as for the Carey Foster bridge
emplo3ning equal coils and a galvanometer resistance of 2 ohm&
The latter method is, however, inapplicable to coils with potential
leads, necessitates a calibration and standardisation of the bridge
wire, is more troublesome in practice, and the accuracy is limited,
not by the general arrangement of the bridge arms but by
the openness of the bridge wire and the accuracy of the scale
and vernier.

The following table gives the difference in values of four coils
with potential leads, every possible combination being taken. The

1QAK

Ck>U8

Temperaiare

of
ObaenratioD

Difference at 17^ G.

Mean

IVIPO

1 X 10-' Ohm

July 21.

n
tt
n
n

Sept 8

)t
tt
tt
>t
tt

2361—2205
2361 2206
2483—2361
2483—2206
2483—2206
2206—2206
2361 2205
2361—2206
2483 2361
2483—2206
2483—2206
2206—2206

17-01' C.
16-96' „
16-86' „
17-10' „
17-12' „
17-06' „
17-26' „
17-22' „
17-19' „
17*33' „
17-38' „
17-30' „

492
426
143
636
669
068
476
346
367
831
701
131

493
424
143
636
668
067
474
344
366
832
702
130

492
426
144
637
667
066
476
344
366
832
700
130

492
426
143
636
668
067
476
344
366
832
701
130

differences in the first column result finom the exchanging of the
coils in the bridge arms ; the differences in the second and third
columns are deduced from observations of the two coils with a
common standard. Thus, from the first and second recorded
observations, the difference 2206—2206 is 130 x lO"^ ohm. The
probable error is of the order of 1 part in 10,000,000. The tem-
perature coe£Scients of these four coils are not very different, and
average 0*001 per cent, per 1"" C. The bath used for the comparison
is that described in the Phil. Trans. A, 878, p. 87, 1904.

The differences recorded above indicate that at least three of
the four coils changed between the dates of the observations. In
a similar manner, very small changes have been observed in a few

688 PRACTICAL STANDARDS

coils in an interval of twenty-four hours. Such changes are veiy
interesting, but cannot be discussed here.

Ten, 100, and 1000 ohm Coils and Resistances of a Higher
Value. — By the bridge method the probable error in the evaluation
of a resistance of 10" ohm is n times the error of the 10 ohms
built up from the unit. This latter error must, therefore, be
made as small as possible. The ''build-up" should contain no
variable contacts, and the lines of flow in the coils when these
latter are evaluated singly should be practically identical with the
lines of flow when the coils are in series. At the National Physical
Laboratory three special build-up boxes have been constructed.
The 10-ohm build-up is here described. In this the coils are of
nominal value, 1, 1, 2, 2, 5 ohms, and may be described as la, 1)9>
2a, 2)3, and 5. Each coil is of manganin, is immersed in oil, and
connected by two copper posts to massive copper blocks, the
blocks being provided with side terminals and mercury contacts.
The coils la and 1/9 are evaluated by the Kelvin double bridge as
described for standard unit coils. The leads to the bridge are
from the mercury cups, and the connexions with the shunt
coils a and /9 are from the side terminals. The resistance thus
measured is that between two points Ijdng centrally under the
mercury contacts in the copper blocks. The value of the 5, 2/8,
2a, and 1/9 in series will, therefore, be exactly equal to the sum of
their individual values. The coils la and 1)9; la, 1)9, and 2a; 2a
and 2/9 ; and 1/9, 2a, 2/9, and 5 are compared by forming a simple
bridge, the coils in the other arms being of 10 ohms resistance.
A reversal in position of the two coils enables the difference to
be accurately found. Finally the 5, 2/9, 2a, and 1/9 are employed
to evaluate a 10-ohm coil. 100 and 1000 ohms are built up in a
similar manner.

Let the constructional errors of the 10, 100, and 1000 ohms
build-up boxes be a, 6, and c respectively. Then, if we neglect
the errors of observation, which are small, the error of a 10-ohm
is a, of a 100-ohm {a + 6), and of a 1000-ohm (a + 6 + c). If
the 100 and 1000 ohm coils are evaluated by a Wheatstone
bridge using the 10 to 1 ratio, then the error of the 100-ohm is
2a, and of the 1000-ohm 3a. Hence, if in practice a = 6, and
2a = 6 + c, the probable error of the build-up values must be very
small. Observations show that the differences 2a — (6 + c),
3a — 36, etc., are not measurable with certainty, for not only are

FOR ELECTRICAL MEASUREMENTS

689

the observed differences •very small, but often the sign changes.
The differences resulting in one set of observations is given in the
following table : —

June 11, 1906. Observed Valvsa in Int. Ohms at 17° C.

L-19

2460

2449

From build-up boxes

From 1 to 10 ratio bj use of Wheat- )
stone bridge {
From build-up boxes

9-9997W
9-9997gB

100-0087
100-008e

lOO-OOSe

1000-53o
1000-627
1000-628
1000-62o

The values given on the first and fourth lines are from the
three build-up boxes. The second values of 2450 and 2449 are
obtained by the bridge, using the 1 to 10 ratio from the first
build-up. The third value of 2449 results fix)m the 1 to 10 ratio
from the second build-up, and the probable error is therefore 36.

Low Resistance Staiidards, — A large number of methods have
been suggested for the measurement of small resistances, and as
many of these : are known to be in use, it may be of service to
point out the advantages and disadvantages of each.

(a) Matthiessen omd Hockin*s Method (fig. 7). — By adjuifting
the resistances R and S, a balance is obtained with the galvano-
meter arm connecting R ' S with each of the potential points of P

Fig. 7.

-AyvAAA/-

■• •■

-WVA/W
R

>AAAAAA-

and Q in succession. The value of R + S ia kept constant. The
ratio of jB to £f is necessarily very great in one of the observations,
and the sensitiveness is, therefore, very small (see expression (D),
p. 680). The method is unsuited for accurate work.

B. A. 44

690 PRACTICAL STANDARDS

(6) Method suggested hy Lord Rayliigh (fig. 8). — ^As an alter-
native to the previous method, the following process was suggested
by Lord Rayleigh in 1884* P is the low resistance whose value
is required. Q is a one- or tenth-ohm standard which is shunted
by the resistances b and c, the ratio of c to 6 being approximately
equal to Q/P if the resistance of the galvanometer is compara-
tively great. When the galvanometer is connected across 6, c is
adjusted until the combination gives the same effect upon it as P
does. Then, supposing the resistance of the galvanometer branch
to remain constant,

P =

Q + b-^c + bc/O'

The method may be made a null one by using a differential
galvanometer and an additional resistance 8 (approximately equal

Fig. 8.

Q P

to P) in the main circuit. One coil of the galvanometer is con-
nected across P and the other across S, the resistance of the
galvanometer arm of 8 being adjusted until there is no deflection.
The P galvanometer coil is then joined across b, and c adjusted to
obtain a balance. Small variations in current strength have no
effect, but the current must be reversed and the combination
readjusted in order to eliminate thermal e.h.f.'s. The resistance
of the galvanometer branch is not constant unless the potential
leads of P are equal in resistance to those of b. In order to
neglect the resistance of these leads, and to make bc/0 compara-
tively small, G must be great. This diminishes the sensitiveness.
If Q is made greater than 1, the maximum permissible current
in the main circuit is reduced, and the sensitiveness is again
diminished. Suppose that P = 001, Q = l, 6 = 1, c = 97, and
0 = 100 + x where x is small. Then P = 1/(99 -h 0*97 - O'Ol^).

♦ Cam!), Phil, Soc. Proc, 1884, v. p. 133.

FOR ELXCTRIGAL MEASUREMENTS

691

Hence, if the value of P is desired to be correct within 0*001 per
cent., the value of the galvanometer resistance must be known to
1 part in 1000. Although not so sensitive as other methods
described hereafter, the process is interesting. The combination
of resistances Q, h, and c was used by Lord Rayleigh in the
determination of the ohm by the method of Lorenz.

(c) Housman's Method* (fig. 9). — The first stage in the •
process is to measure the ratio of P to Q by shunting 12 or
{S'\-Q), The second is to shift one galvanometer lead and one
battery lead and measure the ratio of (P + Q) to Q'. Q' is a
1-ohm coil. For precision work the leads connecting Q' to jS and
P to 22 must be known* The great disadvantage of this method

Pig. 9.

Hit-

=x:^

is that the current through P in the second measurement must
be comparatively small. Thus, if P = 00001, Q = 001 and ^ = 1,
the maximum permissible current through P (if P is the usual
type and size of standard resistance) is 100 amperes; through P
and Q in series 10 amperes, and through P -^Q-^Ql in series,
1 ampere. The necessary ratio of the arms S and 12 is also un-
suited for accurate work.

(d) Two-step Method-^ (A. Campbell) (fig. 10).— A suitable
small resistance, whose value need not be accurately known, is
inserted at Uy and is adjusted by shunting until the galvanometer
balances in position a. The galvanometer is then brought into
position h and balance obtained by another shunt at 12 or 8, By
repeating this process a few times the balance is good in both

♦ Electrician, 1897, xl. p. 800.

t Phil. Mag. July 1903.

44—2

692

PRACTICAL STANDARDS

positions. The method is about 60 per cent more sensitive than
the Kelvin double bridge if equally favourable arrangements are
made, but it is much less convenient in practice. The leads con-
necting P to JfZ and Qto 8 have to be evaluated by changing the
position of the batteiy leads.

Fig. 10.

(e) Potentiometer. — With very small resistfimces, if great
sensitiveness is required, two currents of large value have to be
maintained in a steady state. As the probable error is propor-
tional to the variation in the current strengths this necessitates
great care. In practice the sensitiveness may be made greater
than that of any other method. If P « O'OOOl, R = 0-001,
Q = O'OOl, 8 = O'Ol (see fig. 3), and if we suppose the resistance
of the other portions of the circuits to be comparatively great,
then, with G=l ohm, the sensitiveness for one position of balance

is proportional to 0*01 i A VP, and for the second position of
balance 0*031 i A ViJ. If Q = l, S = 10, the sensitivities corre-
sponding are proportional to 0005* A VP and 0*0029 i A Vit

(/) Kelvin Dovble Bridge* (fig. 11). — For measurements of
precision this method is used at the National Physical Laboratory.
Balance is first obtained by shunting 12 or 8, when

QK' Pd (R a\

8' "*'a + /3 + dU' $)'

R and 8' representing the shunted values of i2 -f X and

* W. Thomson, Phil, Mag. 1S62, 34, 149. See alec Jaeger, St Lindeok, and
Diesselhorst, Zeitschr. Imtrumentenk, 1908, •••

FOR ELECTBIOAL MEASUREMENTS

693

S +£' + £". To obtain the value of L the battery lead at P'i
is disconnected and joined to £ * 12 and the bridge again balanced.
L' + L" is similarly evaluated (see example which follows). To
obtain d, a and /3 are disconnected and the galvanometer circuit
completed by connecting to the junction of Q and d and balancing.
The ratio of a to /9 must be known with considerable accuracy
if (2 is comparatively great, a consists of a resistance coil plus a
potential lead of P, and /3 of another coil plus a potential lead
of Q; hence the ratio must be determined with a and /3 in
position in the bridge* The bridge is first balanced in the

Pig. 11.

ordinary way by shunting R or S. The connector which joins P

to Q through l^e arm d is then removed and balance restored by

shunting a or /9. The original arrangement is restored and the

bridge balanced again. Thus, by successive approximations, we

have

P P^-a a R

Q + /8

where R and 3 are the shunted values of R and 8, Thus ^7 is

equal to ajfi within the limits of the errors of measurement. It
does not follow, however, that dPI{a + /8 + d) x {RjS' - ajff) is
negligibly small. It is only so if the value of c2)8/(a + /9 + <2) does
not exceed the value of P. If the value be NP and the probable

694 PRACTICAL STANDARDS

error of an observation is 1 x 10"^, then the error of the final
result is not less than N x 10~^. It will be seen from this that
the current leads of standard resistances intended for measure-
ment on the Kelvin double bridge should have a resistance not
greater than the standard itself. In some commercial standards
the resistance of the current leads plus the connectors necessary
for their measurement is greater than that of the standfiod strip.
In such cases the potentiometer or Kohlrausch differential galvano-
meter should be employed. In the department of Electrotechnics
at the National Physical Laboratory the potentiometer is used.
The sensitiveness of the Kelvin bridge is less than that of the
potentiometer, but it is more convenient in practice. In the
bridge, if P = 00001, Q = 0001,-8 = 1, flf=10, a=l, /8=10,the
sensitiveness is proportional to 0'0034 iA VP. The galvanometer
resistance is supposed to be 2 ohms. An example follows. For
simplicity P = O'l ohm.

P= No. 2484= 0*1 ohm with potentialleads. Value desired.
§=No.2361= 1-0 „ „ „ „ Value= l-OOOOO^at ITO'C.

i2=No. 2483= 1-0 „ „ „ „ Value= 1-000024 „ »

iS= No. 1693= 10-0 ohma. No potential leads. Value =10-00018 » »

a=l /3 = 10 ««17-0°a

Balance was effected by shunting jR with 122,000 ohms. The
connector completing the branch d was then removed and balance
again established by shunting a with 6500 ohms. The balance
still held good when the connector was restored in position. Hence,
if R and a' represent the shunted values of R and a,

the probable error of these ratios being of the order 0*0001 per
cent, in the present instance. The value of d (for measurement
see the following table) is equal to 0*000128 ohm, and is less
than P. Hence

Q(R + L) I'OOOOOe X (100001, + 0*00011,)
iS + i' + X" *" lO-OOOla + 0-0001,

= 0a00011o at 170^0.

The manner in which d, L, and (L' + i") were evaluated will be
seen from the accompanying table. This is a good instance of a
measurement involving a number of connecting pieces which must
be evaluated in position.

FOR ELEOTRICAL MEASUREMENTS 695

Position of Position of
Galvanometer Battery

Leads Leads Balancing Condition Ohm

{1) U'-R a-fi P'L Q'V Shunt on i2 si 22000 Equivalent change =0'00000„

(2) „ „ L'R Q'U .. „ 5= 8160 .. „ =000122,

(8) L"'S „ P'L S'U „ „ R= 7100 „ „ =000014i

(4) L"'R Qd P'L Q'U „ „ 5= 8870 „ „ sO-OOlW,

From (1) and (2) L = 10 (0*0001808)/11 ^O-OOOll,
(1) „ (8) L' +L"= 10 (0-000188) /11= 0-00012,
(1) . „ (4) d = 0-00012,

»i

(g) The Differential Galvanometer*, — This method is usually
used for comparing resistances which are nominally equal. It is
not convenient for their evaluation from the unit by means of a
ratio of 1 to 10.

The difference of the currents through the galvanometer coils
is % {Pg — Q6)/0 (Q + g) where 0 and g are the resistances of the
galvanometer circuits (see fig. 4). This is equal to zero when
PjQ = QJg, When this latter condition holds there will in general
be a deflection owing to want of symmetry of the galvanometer
coils. If P = 1 and Q = 10, then the ballast resistance in circuit
with g or G may be adjusted until there is no deflection. In such
a case, if two other coils, P'=0'1 and Q' = l-0, are substituted
for P and Q {G and g remaining as before), and ten times the
previous current sent through them, there will be no deflection
when PjQ = P'jQ. In general, however, the maximum permissible

current is VlO times that previously employed, and any want of
8}rmmetry in the galvanometer coils does therefore introduce an
error. In addition, the substitution of P' and Q for P and Q
changes the values of G and g, because these latter include the
potential leads of the resistances and also the contact resistances
introduced. If G and g are comparatively large, the error is
reduced, but so also is the sensitiveness. In the same way, errors
are introduced in the comparison of nominally equal resistances.
In this latter case if P and Q are exchanged, P is equal to Q when
there is no change in the deflection, and no error is introduced by
want of symmetry in the galvanometer coils or inequality of the
resistances of the galvanometer circuits, always supposing that
these latter remain constant throughout the observations. Un-
fortunately, the resistances of these circuits do change, for the

* See HeavUide*t Papers, vol. i. Also C. W. S. Crawley, Joum. Irut, of
Electrical Engineers, April 1904.

696 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

reason previously given, and the error introduced may be con-
siderable. Let P = Q =s 0*1 ohm, and let the resistance of the
leads of P = O'OOOl ohm, and of Q = 00002 ohm. (In some low
resistances the potential leads are of the order O'Ol ohm.) Then
i{ 0 = 1, and no correction is applied for the inequality of the
leads, the error of measurement is O'Ol per cent. If G = 100 ohms,
the error is 0*0001 per cent., but the sensitiveness is reduced to
one-fifth of its former value. Such errors are abolished if the
Kohlrausch method of overlapping shunts* is used, of which a
diagram of connexions is given in fig. 13. In fig. 12 let P=^Q.
Then, unless there is symmetry of the galvanometer coils and

Fig. 12.

Fig. 18.

-• •-

equality of resistance of their arms, there will be a deflectioii.
Supposing that 0 emd g can be exchanged in position by sub-
stituting for X a resistance practically identical with it, then the
deflection will be of the same magnitude and of the same sign as
before. In practice P is made equal to Q by shunting one of
them, the equality being determined by the equality in magnitude
emd sign of the deflection before and after interchanging G and g.
The exchange is effected by a six-pole switch as shown in fig. 13.
The resistance of the galvanometer circuits is thus constant, and
it is apparent that the Kohlrausch method of using the differential
galvanometer is the only one so far suggested that can be used for
precision measurements.

* Wied. Ann. 1888, 90, 76^, See also article by Jaeger, ZeiUchr, Imtmmenten'
kunde, 1904, 288.

THIRTY-FOURTH REPORT— LEICESTER, 1907.

APPENDIX PAOB

I. Notsi on the pruent condition of the work on EUctric Units
at the NaHonal Physical Laboratory, By F. £. Smith.
{From the Naiional Physical Laboratory) . . 700

II. Specification for the Practical Application of the Definition of

the IntenuUional Ampire. {From the National Physical
Laboratory) 703

III. Preparation of the Weston {Cadmium) Standard Cell, {FVom

the National Physical Laboratory) . • « . * 707

The main work during the year has been the completion of
the work with the amp^ balance. The general results are
referred to in some detail below. The final measurements confirm
the opinion expressed in last year's Report that an accuracy of a
few parts in 100,000 might be reached. It appears that the
result is probably accurate within 1 in 50,000.

Interim reports on the ampere balance, indicating the progress
of construction, adjustment, and use of the instrument, have been
submitted to the Association since 1904.

The Committee are now pleased to report that the balance con-
tinues to give complete satisfaction. During the past year it has
been much used for determining the E.M.F. of the Weston Normal
Cadmium Cell and the electro-chemical equivalent of silver. A
description of the instrument, its construction and adjustment,
and the results obtained with it in the cadmium cell determina-
tions, has been prepared emd submitted to the Royal Society for
publication in its TramacticfM, by Professor Ayrton, Mr Mather,
and Mr F. E. Smith. An account of the work on the electro-
chemical equivalent of silver is well advanced and will be published
shortly.

In all some 71 observations have been made on a certain
cadmium cell (No. 2), using both sets of coils on the balance, and
13 observations in which one or other of the two sets was em-
ployed. The agreement between the individual results obtained

698 PRACTICAL STANDARDS

with the two sets of coils is remarkable, the average difference
from the mean amounting only to 6 parts in a million. The whole
series of observations extended over a period of nineteen months
(September 1906 to April 1907), and during that time the coils
of the balance were reset five times. No determination made has
been omitted, except those in which the observations were of such
a nature that a decision to disregard the result was arrived at
before its computation. Such occasions were rare.
Of the 71 observations made

7 are within 1 in a million of the mean
14 ,. .. 2

» 99

2o „ „ 5 „ „

o3 •„ „ 10 ,i „

66 „ „ 15 „

70 •„ „ 20 „

9 »

9 99

9 »

Only one determination out of the whole 71, and this one of the
earliest, differs from the mean by so much as 1 part in 59,000.

The above facts constitute important evidence of constancy in
both balance and cell. In fact, both current-weigher and cell
proved to be much more constant and reliable than the standard
resistance, although the latter was very carefully made and annealed
with a view to ensuring permanency.

Expressed in terms of the international ohm as realised at the
National Physical Laboratory, and of the ampere as given by the
new current-weigher, we find that the value of G x R for the
Weston normal cadmium cell is I'OIBSOb at 17** C.

This assumes that the value of g at Teddington is 981*19, a
number probably correct to within 3 parts in 100,000. An un-
certainty of this amount in g introduces a possible error of 1^ parts
in 100,000 in the value of the ampere, and, as all other probable
errors are smaller in magnitude, it is important that a more
accurate determination of ^ be made.

To realise the volt with an accuracy approaching that of the
ampere, as now known, it is necessary that an absolute determina-
tion of resistance of corresponding precision be undertaken.
Through the kindness of the Drapers' Company of London it is
hoped that such a determination by means of a Lorenz apparatus
may be commenced at the National Physical Laboratory before
the end of next year. At the present time the uncertainty in the

FOR ELECTRICAL MEASUREMENTS 699

absolute value of the international ohm approximates to 4 in
10,000.

From the above value of C x i2 for the cadmium cell, together
with the ratio of Clark to cadmium, viz.,

Clark at 15° C. ^
Cadmium at 17" C."*'

Ihe E.M.F. of the Clark cell at 16** C. becomes 1-432,.

The Committee recognise very fully the skill and devotion of
Mr Mather and Mr Smith, on whom the work of carrying out the
experiments has fallen, and have invited these gentlemen to become
members of the Committee.

Papers by Mr F. E. Smith, of the National Physical Laboratory,
dealing with the use of the silver voltameter and the preparation
of the Weston cadmium cell, are nearly ready for publication.

Some preliminary work has also been done on the design for
the Lorenz apparatus, the funds for which are being found by the
Drapers' Company. The proposed design embodies new features
of importance.

With regard to the proposed Conference on Electric Units,
further consideration led to the conclusion that a year's delay was
desirable, and in consequence the meeting was postponed from
October 1906 to October 1907. With a view to a preliminary
agreement on the matters to be raised, correspondence has passed
during the year between the Secretary, acting as Director of the
National Physical Laboratory, emd the heads of standardising
laboratories in other countries. The Conference will probably
deal with the drawing up of an International Convention relative
to Electric Units, which should include the draft of a form of law
which might be adopted generally in the various countries re-
presented, and the consideration of the steps necessary to secure
uniformity in the carrying out of the laws in different countries,
and to arrange for determinations necessary for this purpose.

The necessary invitations for the Conference are being issued
by his Majesty's Government.

To secure uniformity in carrying out the law it will be
necessary that specifications for constructing and using a mercury
unit of resistance, for setting up and working a silver voltameter,
and for preparing a standard cell, be approved either by the Con-
ference itself or by some body nominated by the Conference for
this purpose.

700 PRACTICAL STANDARDS

With a view to aiding discussion, very detailed specifications
dealing with the voltameter and the cell have been prepared by
the National Physical Laboratory and issued to other standardising
institutions. These are printed in Appendices II. and III.

It is not suggested that the final specifications need be so full or
so detailed, but it was thought well that all information necessaiy
to assist in criticising the results should be included.

The work on the silver voltameter and Weston cell still con-
tinues, and, in view of the deliberations of the Conference, it is
probable that further expenditure will be required. The accounts
show that a balance of lOa. Sd. remains from the grant of £50
made last year. The grant has been spent on the purchase of
material emd appliances for the research.

In view of the importance of bringing the work of re-
determining the values of the fundamental units to a satisfetctory
conclusion, the Committee recommend that they be reappointed,
with a grant of £50, and with the addition of the names of
Mr A. P. Trotter, Mr T. Mather, F.RS., and Mr F. R Smith ; that
Lord Rayleigh be Chairman and Dr Qlazebrook Secretary.

Appendix I.

Notes on the Present Condition of the Work on Electric Units cU
the National Physical Laboratory. By F. E. Smith.

{From the NoLtional Pkydcal Laboratory.)

1. The Ohm, (a) Absolute Unit — The value of a resistance
in absolute measure is still subject to considerable uncertainty;
the most satisfa^story value is obtained from the mean of the
results obtained for the ratio of the International Ohm to the
absolute ohm*.

A provisional design has been prepared for the Lorenz
apparatus which the Drapers' Company are kindly presenting to
the National Physical Laboratory, and experiments to test the
more important features of the design are in progress. It is hoped
to realise the ohm in absolute measure to within 1 part in
100,000. The experience gained in the construction of many of
the fittings of the ampere balance will greatly facilitate the work.

(b) International Unit. — Further comparisons of some of the
mercury standards of the National Physical Laboratory were

* See table in the Brit. Asboo. Bep. for 1892.

FOR ELECTRICAL MEASUREMENTS 701

made in October and November 1906. There appears to have
been no change in any of the tubes which affects the resistance
of the contained mercury columns by as much as 1 part in
100,000. The following table gives the observed differences in
1903 (the year of their construction) and in October and November
1906.

Mercary Standards

Obserred Diflerenoe in

Observed Difference in

Compared

Int. Ohms in 1903

Int. Ohmft in 1906

M-P

0-000694

0-000685

M-T

88fl

89s

M-U

947

963

M^V

298

30,

M-X

OI3

023

2. The Amph'e, (a) Absolute Measure. — When the ampere
balance was designed it was hoped by means of it to measure a
current in absolute value to within 1 part in 10,000, but it will
be seen from the report on the balance that the evaluation of a
current of nominal value 1 ampere is subject to an error which
appears to be not greater than 1 part in 50,000.

(b) International Unit of Current — ^The International Con-
ference on Electric Units at Charlottenburg (1905) reaflSrmed
the definition of the International ampere in terms of the deposit
of silver in a silver voltameter or coulometer, but expressed the
opinion that the information before it was insufficient to enable
it to propose any alteration in the formerly accepted value for
the ampere, or to lay down exact directions in respect to the silver
voltameter.

The Rayleigh type of voltameter has been used in a large
number of investigations, but the researches of Rodger and
Watson, Richards, E^hle, and others have shown that this volta-
meter as generally employed gives results which may vary as
much as 1 part in 1000.

In the research at the National Physical Laboratory a repro-
ducible type of voltameter was sought, but after making a large
number of observations on various forms it was found that, subject
to certain easily attained conditions, all the forms give identical
results to within 1 part in 100,000. As the Rayleigh type is the
simplest to erect and produces the least variation in the current
strength, it is proposed that this form be specified. The con-
clusions arrived at in the research differ appreciably from those of
most other observers, and attempts have been made experimentally

702 PRACTICAL STANDARDS

to reproduce the conditions under which they worked. In part we
have been successful, but there are still emomalous results for which
we can at present offer no explanation.

It is certain, however, though the complete chemistry of the
silver voltameter or coulometer is unknown, that a reproducible
type can be specified, and that the International ampere can be
defined in terms of the deposit of silver with very great accuracy,
certainly to 1 part in 100,000.

The Standard Cell. — For the past five years experiments have
been made at the National Physical Laboratory on Clark and on
Weston cadmium cells, and two years ago a provisional specifica-
tion of the cadmium cell was published. It is gratifying to know
that the specification proved of value, for in 1906 fifty-one
cadmium cells were submitted for test at the National Physical
Laboratory, and all of these were prepared on the lines of the
specification. The cells were intended for commercial use, and
they were packed with small crystals of cadmium sulphate to be
more portable ; we have reason to believe that in some cases the
mercurous sulphate had not been properly washed, and in other
cases the solution of cadmium sulphate was slightly acid. Never-
theless the E.M.F. of these cells agreed with the N.P.L. cells to within
about 2 parts in 10,000, the N.P.L. cells having the lower voltage.
Standards more carefully set up have been submitted by two
observers for comparison with the N.P.L. cells in accordance with
the offer made in the British Association Report, 1905. The cells
prepared by one of these observers — ^Mr Tinsley of Beckenham —
differed from the N.P.L. cells by about 0*1 millivolt, or 1 part in
10,000. Mr Mather also submitted a number of cells, and these
had approximately the same mean E.M.F. as those from Mr Tinsley.
The N.P.L. cells were the lower in voltage, and freshly prepared
cells agree with old ones*.

In May 1907 twelve Weston cadmium cells set up by Dr Wolff
at the National Bureau of Standards, Washington, were compared
with a number of the cells of the National Physical Laboratory,
and a mean difference of 3 parts in 1,000,000 was measured.
Dr Wolffs cells were, we believe, set up quite independently of
the N.F.L. specification, which makes this remarkable agreement
all the more gratifying.

* In Mr Mather's oeUs electrolytic mercurous sulphate was used; in Mr Tinal^s
ceUs the mercurous sulphate was prepared by the chemical precipitation method.

FOR ELECTRICAL MEASUREMENTS 703

Appendix IL

Specification for the Practical Application of the Definition of

the International Amph'e.

{From the National Phyncal Laboratory.)

In the following specification the term silver voltameter (or
coulometer) means the arrangement of apparatus by means of
which an electric current is passed through a solution of silver
nitrate in water. The silver voltameter measures the total
electrical quantity which has passed during the time of the ex-
periment, and by noting this time the time-average of the current,
or, if the current has been kept constant, the current itself, can
be deduced.

In employing the silver voltameter to measure currents of
about 1 ampire the following arrangements should be adopted:
The cathode on which the silver is to be deposited should take
the form of a platinum bowl about 10 centimetres in diameter
and 7 centimetres in depth. The mass of the bowl is conveniently
about 80 grams.

The anode should be a plate or disc of pure silver coated with
a deposit of electrolytic silver, the mass of the latter being about
50 per cent, greater than the mass of silver to be deposited on
the cathode. The plate or disc of silver should be of about 6 centi-
metres edge (or diameter) and 3 or 4 millimetres in thickness.
Its total area will thus approximate to 60 square centimetres.
The anode should be supported horizontally in the liquid near the
top of the solution by a silver rod riveted through its centre.
To prevent the disintegrated silver which is formed on the anode
from felling upon the cathode the emode should be inserted into a
cup of filter-paper separately supported

The liquid should consist of a neutral solution of pure silver
nitrate, containing about fifteen parts by weight of the nitrate to
eighty-five parts of water.

The resistance of the voltameter changes somewhat as the
current passes. To prevent these changes having too great an
efifect on the current some resistance besides that of the volta-
meter should be inserted in the circuit. If the value of the
current is desired and the measurement is one of high precision.

704 PRACTICAL STAKDA&DS

this external resistance should be &om 50 to 100 ohms ; in other
cases the resistance shoald not be less than 10 ohm&

Method of maJcing a Measurement

1. The Solution. — The silver nitrate should be purchased as
pure and recrystallised twice; the recrystallisation is preferably
done by evaporating a saturated solution in a flask over a water-
bath. The mother liquor should be drained away and the crystak
dissolved in pure freshly distilled water. Prolonged contact of the
crystals or of the solution with impure air must be avoided. The
solution should be neutral to sensitive litmus-paper.

If the silver nitrate is recovered from much used or con-
taminated solutions, or from an acid solution, the recovered salt
should be fosed (preferably in an electric oven) and afterwards
dissolved, and the solution filtered before the recrystallisation
processes; otherwise it may be necessary to reciystallise more
than twice.

During electrolysis in the voltameter herein specified the
silver nitrate solution does not change in composition as a result
of the electrolysis by an amount which is detectable by any tried
means, but, owing to the presence of impurities in the atmosphere,
the solution should not be used more than once if great accuracy
is desired.

2. The Anode, — The anode should be prepared by cleaning
the silver plate or disc with sand-paper or a scratch-brush. It
should be washed with distilled water and supported so as to
form the cathode of a silver voltameter. The anode of this latter
should be a silver bowl or a platinum bowl coated with silver,
and the liquid should be a 15 per cent, solution of silver nitrate
in water ; this solution need not be specially pure. If the anode
bowl is of platinum coated with silver and of the dimensions
already specified, it is convenient to employ about 350 cubic
centimetres of the solution and support the silver plate or disc
horizontally in the liquid near the top of the solution. A con-
venient current for depositing silver on the plate is 0*3 ampere.
The plate is washed with distilled water and dried in an electric
oven.

The cup of filter-paper should be about 5 centimetres deep and
of a diameter a little greater than that of the silver plate. It is

FOR ELECTRICAL MEASUREMENTS 705

made by folding a large filter-paper (such as Schleicher and Schnll
No. 596, 24 cm. diameter) over a glass cylinder (such as a bottle)
of appropriate diameter and securing the upper portions of the
folds of the paper with sealing wax or with platinum wire. The
cylinder is removed and that portion of the paper which is above
the seals is cut away. The upper parts of the internal folds are
also secured with sealing W€uc or with platinum wire.

3. The Kathode, — The platinum bowl should be cleaned with
a strong solution of sodium hydrate, followed by washings with
water, strong nitric acid, and distilled water. It is then made the
anode of a silver voltameter, the liquid being a 15 per cent,
solution of silver nitrate (an impure solution serves) having a
volume of about 350 cubic centimetres. The kathode should be
a clean silver plate supported near the top of the solution. With
a current of about 1 ampere the circuit should be completed for
ten minutes at least, after which the kathode and liquid are
removed from the bowl. The bowl is washed with water and
afterwards cleaned with strong nitric acid; washings with
distilled water, strong nitric acid, and distilled water follow in
the order named, and the bowl is dried in an electric oven at
about a temperature of 160° C. It is removed to a desiccator
and when thoroughly cool is weighed. A bowl of similar
size and of approximately the same mass is convenient as a
counterpoise.

4. The Circuit. — The platinum bowl is placed in position in
the intended circuit and 300 cubic centimetres of the solution of
silver nitrate are placed in it. The anode is placed inside the
filter-paper cup and the latter suspended by platinum wires,
which are insulated fix)m the anode and from the rest of the
circuit. The anode and filter-paper cup are supported so that the
silver plate or disc is covered by the solution ; the connections to
the remainder of the circuit are then made. Contact is made at
a key and the time noted. The current is allowed to pass for an
interval depending on the precision desired, and the time of
breaking contact must be observed. For measurements of high
precision from 7 to 10 grams of silver should be deposited. During
the passage of the current the voltameter should be covered over,
to exclude light.

5. Deposit of Silver. — The solution is removed from the bowl
and the deposit rinsed with about 100 cubic centimetres of distilled

a A. 45

706 PRAC?nCAL STANDARDS

water. The washing water is poured into a clean glass oiystallising-
dish and the operation of washing is repeated three times. The
bowl is then nearly filled with distilled water and left for at least
three hours; it is rinsed three times, the last of these washing
waters remaining in the bowl for ten minutes. This should give
no milkiness when added to a neutral solution of sodium chloride
in water. The bowl is dried in an electric oven at a temperature
of about 160* C.

If any loose silver is observed in the solution outside of the
filter*paper cup, or in the washing waters, these liquids must be
filtered, the filter-paper dried, and the loose silver added to the
bowl before drying the deposit. The bowl is cooled in a desiccator
and weighed again. The gain in mass gives the silver deposited.

6. CalculatunL — ^To find the cuiient in amperes this mass,
expressed in grams, must be divided by the number of seconds
during which the current has been passed and by 0*001118. The
result will be the time-average of the current, if during the interval
the current has varied.

In determining the constant of an instrument by this method
the current should be kept as uniform as possible, and the readings
of the instrument observed at fi^uent intervals of time. These
observations give a curve firom which the reading corresponding to
the mean current (time-average of the current) can be found.
The current, as calculated from the voltameter results, corresponds
with this reading.

Notes on Observations. — If this specification is carefully followed
the mass of silver deposited for the passage of one coulomb through
the voltameter is constant within the limits of the errors of
measurements of the highest precision. It is certainly constant
to 1 part in 100,000.

The specification is possibly too rigorous for many practical
needs, and for such a simplification is possible. The solution of
silver nitrate may be prepared fix)m purchased silver nitrate, pro-
vided it is free fix)m acid. The anode may be a plate of pure silver
without electrolytic silver deposited thereon. The remainder of
the specification must be followed.

Effect of Pressure. — The observations may be made at any
ordinary atmospheric pressure, or exceptionally low pressures, as
the mass of silver deposited when the silver voltameter is under
a pressure of 76 centimetres of mercury is the same as when

FOB ELSGTBICAL MEASUREMENTS 707

under any lower pressure to 2 centimetres of mercury, and possibly
without these limits.

Effect of Temperature, — ^This specification is based on observa-
tions at or about a temperature of 17'' C. Observations at other
temperatures have been made and are being continued ; if there
is a temperature coefficient to the silver voltameter it is ex-
ceedingly small.

This specification is based on the results of a large number of
measurements made at the National Physical Laboratory.

Appendix III.
PrepanxHon of the Weston {Cadmitm) Standard Cell.

{From the NcUional Phyiical LaborcOory,)

Definition of the Cell. — The cell has mercury for its positive
electrode and an amalgam of cadmium, consisting of 12^ parts
by weight of cadmium to 87^ parts of mercury, for its negative
electrode. The electrolyte consists of a saturated solution of
cadmium sulphate, and solid cadmium sulphate is contained
within the cell. A paste consisting of solid mercurous sulphate,
mercury, and cadmium sulphate rests on the positive electrode.

Preparation of the Materials.

1. Mercury. — Commercially pure mercury should be squeezed
through wash-leather and passed in the finely divided condition
in which it emerges through dilute nitric acid (1 part of acid to
6 parts of water) and mercurous nitrate solution, and afterwards
through distilled water. These liquids are conveniently contained
in long glass tubes. The mercury is then distilled twice in va4)uo.
Mercury suspected of any abnormal contamination should not be
employed.

2. Cadmium Amalgam. — A current is passed from a thick
rod of pure commercial cadmium to distilled mercury, the inter-
vening liquid being cadmium sulphate solution rendered slightly
acid with a few drops of sulphuric acid. The kathode is weighed
before electrolysis commences, and again afterwards; the percentage
of cadmium in the amalgam is then calculated. More than the

45—2

708 PRACTICAL STANDARDS

•

requisite amount of cadmium should be deposited and the per-
centage reduced to 12^ by the addition of mercury. To prevent
the anode slime having access to the kathode the anode should be
contained in a filter-paper cup, as in the Bayleigh form of silver
voltameter. Contact with the kathode is made by a platinum
wire sealed into a glass tube, the wire being thus protected from
direct contact with the cadmium sulphate solution* An approxi-
mate estimate of the quantity of cadmium deposited is obtained
from the readings of an ammeter placed in the circuit. The
amalgam, with very dilute sulphuric acid flooding its surface, is
melted over a water-bath and stirred to ensure homogeneity. It
is then ready for use.

3. Cadmium Sulphate, — Procure commerci&lly pure cadmium
sulphate, CdSOf • 8/3 H,0. Powder in a mortar and dissolve in
distilled water until a saturated solution results: filter the solution
through a fine-grained filter-paper until it is quite clear. The
liquid should then be placed in a large crystallising dish and
slowly evaporated at a temperature of about 35° C, when, pro-
vided that dust is excluded, many transparent crystals of
CdSO* . 8/3 HjO will result. In this way about five-sixths of
the solution may be evaporated (the mother liquor may be
used for a preliminary washing of the mercurous sulphate, the
manufacture of which is described hereafter). The recrystallised
cadmium sulphate should be washed with successive small quan-
tities of distilled water, until after standing for ten minutes no
trace of acidity can be detected in it with sensitive congo-red
paper : the crystals, still moist, are transferred to a stock bottle.
To prepare the saturated solution the crystals are crushed in a
mortar and agitated with distilled water. The latter may be
warmed to 35** C.

4. Mercurous Sulphate, — Add 15 cubic centimetres of pure
strong nitric acid to 100 grams of pure mercury contained in a
crystallising dish, and place on one side until the action is over, or
nearly over. Transfer the mercurous nitrate thus formed, together
with the excess of mercury, to IV beaker containing about 200
cubic centimetres of dilute nitric acid (1 of acid to 40 of water
by volume) ; a clear solution should result. Prepare about 1 litre
of dilute sulphuric acid (1 of acid to 3 of water by volume), and
while the mixture is hot add the acid mercurous nitrate solution
to it. The solution should be added as a very fine stream from

FOB ELECTRICAL MEASUREMENTS 709

the narrow orifice of a pipette and the mixture violently agitated
during the mixing. Mercurous sulphate is precipitated and
rapidly settles to the bottom of the vessel when the stirring
ceases. Decant the hot clear liquid and wash the precipitate
twice by decantation with dilute sulphuric acid (1 of acid to 6 of
water by volume). The precipitate should then be filtered. (A
small Buchener filter funnel and a filter flask is very convenient
for this latter operation.) Wash the precipitate three times with
the dilute sulphuric acid (1 to 6), and afterwards wash six or seven
times with saturated cadmium sulphate solution to remove the
acid. After each washing the liquid should be removed as com-
pletely as possible by the filter pump. When these operations
are complete, the mercurous sulphate is flooded with saturated
cadmium sulphate solution and left for one hour ; the solution is
then tested with congo-red paper. In general no acid will be
detected, and the mercurous sulphate is ready for use. It is
placed in a stock bottle together with some saturated cadmium
sulphate solution, and should be kept in the dark. If acid is
detected, the washing must be continued. When the cells are
required for observations of the highest precision, the apparently
neutral mercurous sulphate should not be immediately used. It is
placed in a bottle with saturated cadmium sulphate solution, and
at the end of one week the latter is tested for acidity. The
sulphate is given another washing with the solution, and may then
be used if only a trace of acid is detected.

One of the following methods of preparation may, if desired,
be substituted for the foregoing : —

(1) Electrolytic Method. — This preparation is conducted in a
darkened room. Pure distilled mercury forms the anode and
platinum foil the kathode, the electrolyte being dilute sulphuric
acid (1 volume of acid to 5 volumes of water). The mercury is
placed in the bottom of a large flat-based beaker and about
twenty times its volume of the dilute acid is added. Contact
with the mercury is made by a platinum wire passing through a
glass tube, and the kathode is suspended in the upper portion of
the liquid. During electrolysis the electrolyte must be con-
tinually stirred, an L-shaped glass stirrer being very efficient, the
foot of the L moving close to the surfisu^e of the mercury. A con-
venient current density is 0*01 ampere per square centimetre of
anode sur&ce. The mercurous sulphate so prepared is filtered

710 PRACTICAL STANDARDS

and the greater part of the mercury removed ; it is then washed
with dilate sulphuric acid and saturated cadmium sulphate solution
in a manner already described for the previous preparation.

(2) By meam of Fuming Sitlphuric ileid-^Place distilled
mercury in a crystallising dish so as just to cover the bottom.
Add sufficient fuming sulphuric acid to flood the sur&ce of the
mercuiy to a depth of about 2 millimetres. Cover with a clock
glass and place on one side for 48 hours. Mercurous sulphate is
formed emd appears in the crystalline form. Carefully add the
salt to hot dilute sulphuric acid (1 to 6) and well agitate. Decant
the hot. liquid. If any caked masses of the sulphate are left,
these should be rejected or crushed in an agate mortar. Wash
three times by decantation with hot dilute sulphuric acid, and
afterwards filter and wash with saturated cadmium sulphate solu-
tion in the manner already described. Set aside with cadmium
sulphate solution for one week at least, test for acidity, and wash
as described for the first preparation.

The Mercurous Sulphate Paste. — ^The mercurous sulphate is
mixed with about one-fourth its volume of powdered recrystallised
cadmium sulphate, and about one-tenth its volume of pure mercuiy.
(When the electrolytic sulphate is used, or that prepared with
fuming sulphuric acid, no mercury need be added.) To the
mixture of mercurous sulphate, cadmium sulphate, and mercury,
sufficient saturated cadmium sulphate solution is added, so that
when well mixed the whole forms a thin paste.

Setting up the Cell, — ^That type of H-form of cell which may
be hermetically sealed is the most convenient; if the lower end
of each limb is slightly constricted, the ccmtents of the cell are
less liable to be disturbed The platinum wires inside the glass
vessel are amalgamated by passing an electric current from a
platinum wire anode through an acid solution of mercurous nitrate
to each of the wires in turn as a kathode. The vessel is washed
out twice with dilute nitric acid, several times trith water, and
finally with distilled water; it is dried in an oven. A small
pipette is used for the introduction of the amalgam, and a small
thistle funnel for the insertion of the paste and crystals. The
main stock of amalgam is flooded with very dilute sulphuric acid,
and it is melted over a water-bath ; a little of it is introduced
into one of the limbs of the H -vessel. After the amalgam has
solidified, this limb must be washed out several times with distilled

FOR ELECTRICAL MEASUREMENTS 711

water, care being taken to avoid wetting the interior of the other
limb. A little distilled water is added and the amalgam is melted
by immersing the limbs of the H-vessel in hot water. After the
solidification of the amalgam, it is washed once more with distilled
water. Into the other limb sufficient mercury is introduced to
cover the amalgamated platinum wire; then the paste is added,
care being taken not to smear the sides of the vessel. Finally,
powdered crystals of cadmium sulphate are introduced into each
limb, and saturated cadmium sulphate solution is added. The
cell may be immediately sealed with the aid of a blowpipe, but
the- contents must not be abnormally heated thereby. The
cadmium amalgam introduced should cover the amalgamated
platinum wire ; the depth of the paste should be from 0*5 cm. to
I'O cm., and the depth of the layer of crystals about 0*5 cm.
Twenty-four hours after the cell has been set up it may be used.
Its electromotive force at 15/" C. is I'OlSs volt. The electromotive
force at any other temperature may be obtained from the formula
given by the Phys. Techn. Reichsanstalt, viz.,

Et^E^" 0-000038 {t - 20) - 000000065 {t - 20)*,

or from the formula obtained at the National Physical Laboratory,

Et = E„ - 0-0000345 (« - 17) - 000000066 {t - 17)».

This specification is based on observations made at the National
Physical Laboratory.

THIRTY-FIFTH KEPOKT— DUBLIN, 1908.

APPENDIX PAOS

I. On the Secular Changes of the Standards of Bemtance at the

National Physical Laboratory. By F. E. Smith, A.R.C.Sc
{From the National Physical Laboratory) .... 716

II. Specificalions for the Practical Realisation of the International

Ohm and International Amphe, and Instructions for the
Prepcuratum of the Weston Cadmium Cdl. {From the
Na,tional Physical Laboratory) 738

The Committee desire in the first place to record their deep
sense of the loss they have sustained by the death of Lord Kelvin.
He was an original member of the Committee appointed at
Cambridge, October 3, 1862, and he continued his active interest
in their work up to the end. His name will always be associated
with the establishment of the absolute system of electrical
measurement and with the determination of the absolute units.
The Reports of the Committee from 1862 onwards contain a large
amount of valuable information in a form which is not generally
very accessible — the reprint of the earlier reports, issued under
the editorship of Fleeming Jenkin in 1873, is out of print — and
the Committee suggest that their reports fix)m 1862 up to the
present time might be reprinted as a memorial to Lord Kelvin.
The present time is in other respects specially suitable for such a
reissue, for it is hoped that the proposed International Congress,
to be held in London in October, will settle in a definite manner
the few matters relating to the fundamental units which are still
outstanding, and will organise a method whereby a close agree-
ment may be maintained among the electrical standards in use
throughout the world.

The electrical measurements of certain of the fundamental
units, which have been in progress for some time at the National
Physical Laboratory, have been brought to a conclusion, and the

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 713

results published in three papers in the PhUoaophicai Transactions
of the Royal Society.

1. * A New Current Weigher and a Determination of the Electromotive
Force of the Normal Weston Cadmium Cell' By Professor W. K Ayrton, F.R.S.,
and T. Mather, F.R.S., Central Technical College, London ; and F. E. Smith,
A.R.C.SC., National Physical Laboratory, Teddingtibn, Fkil. Trans, A, VoL 207,
pp. 463-549.

2. <0n the Normal Weston Cadmium Cell.' By F. E. Smith, Phil.
Trans. A, Vol. 207, pp. 393-420.

3. *0n a Comparison of many forms of Silver Voltameters.' By
F. E. Smith ; and ' A Determination of the Electrochemical Equivalent of
SUver.' By F. E. Smith and T. Mather, F.RS., PhU. Trans. A, VoL 207,
pp. 645-681.

* The Chemistry of the Silver Voltameter.' By F. K Smith and T. M.
Lowry, D.Sa, Phil, Trans. A, Vol. 207, pp. 581-699.

From the first of these it appears that to a very high degree
of accuracy the electromotive force of the Weston cadmium cell,
as set up at the National Physical Laboratory, when expressed in
terms of the ampere (10""^ c.as. units of current) and the inter-
national ohm is l*0183o at a temperature of 17"* C.

The second Paper deals with the preparation of the Weston
cadmium cell, and contains a comparison between cells set up at
the Laboratory and others constructed elsewhere, the general
conclusion being that cells can be prepsired by different persons
in different countries which will agree in E.M.F. to 1 or 2 parts in
100,000.

In the third Paper there is given an account of a comparison
of some six forms of silver voltameters, and it is shown that the
silver deposited by a current of one ampere in all these various
forms is the same if proper precautions are taken, and amounts to

1*11827 milligrammes per second

After this work was completed a comparison was made by
Messrs T. Mather and F. E. Smith, by the kindness of Mr Trotter,
between the ampere standard of the Board of Trade and the
ampere as realised by the new Ajrrton Jones balance at the
National Physical Laboratory. The comparison, an account of
which appears in the Proceedings of the Royal Society, A, vol. 80,
1908, was very satisfisu^tory.

It was found that the Board of Trade amp^ will deposit
silver at the rate of l'1179i milligrammes per second, a value

714

PRACTICAL STANDARDS

which is nearly identical with the number 1*1179« given by Lord
Bayleigh and Mrs Sidgwick in 1884. Indirectly the S.1LF. of the
normal Weston cadmium cell was found to be 1*01879 Board of
Trade volt at 17° C, the Board of Trade volt being defined as the
potential difference between the terminals of a resistance of
1 Board of Trade ohm' when 1 Board of Trade amp^ is passing
through it.

During the year the ten mercury standards at the National
Physical Laboratory have again been set up and interoompared.
An account of this work appears in an Appendix by Mr
F. K Smith, the value of the international ohm, as realised by
the mean of the ten tubes, being taken as unit The following
table gives the values of the individual tubes as found in 1903
and 1907:—

Biercary
Standard

Value in Mean International Ohms

Differenoe

valne
1907-1908 1

1903

1907

0-971705
l-OOOSSr
1000196
0-973497
1-001379
lOOlOSs
1-000267
l-00130e
1-001062
1-000974

0-971699
l-0004Si

l-OOOSQs
0*973488

1 -001371
1 -00106s
1-000266
1-001299
1-001044
1-00097J

-0,
+8i

+li
-0,

-0,
-0,
-0,
-0,
-0, '

Except in the case of Tube P, where there is an apparent
change of 3 to 4 parts in 100,000, the differences are negligible.

Mr Smith has also compared with the mercuiy tubes a large
series of wire standard resistances, including those made by
Matthiessen and Hockin for the B. A. Committee in 1865-67,
and various other old standards kindly lent to the Committee by
their owners for the purpose. The general conclusion is that all
the original coils except D and E, which are made of platinum,
have changed appreciably since they were constructed, though since
1888, during a period of twenty years, for which the coils have
been very carefully watched, the changes also in A, B^ (7, H, and

FOR ELEOTftlCAL IIEASURKMENTS

715

Flat have been small F and 0 have, however, in the same period
changed considerably.

Besigtance at 16*0^ C. in terms of the original B.AM. (1867).

Ck>U Siaterial

1867

1876

1879-81

1888

1908

Maximum
Ghange

Ptir.

lOOOOO

1-00077

1*00056

1-00147

1-00122

0-00147

Pt It.

1-00029

1-00121

1-00080

1-00104

1-00008

0-00002

Au. Ag.

1-00050

1-00141

1-00101

1-00146

1*00173

0-00123

1-00002

0-00000

1-00152

0-00000

PtAg.

—

— —

1-00016

1-00072

1-00160

0-00144

PtAg.

1-00022

1-00030

0-09082

1-00025

1-00175

0-00103

PtAg.

1-00020

— —

1-00042

100044

0-00024

Flat

Pt. Ag. —

1-00070

1-00120

1*00125

0*00046

The mean resistance of six other platinum^silver coils, first
measured in 1888, appears to have increased since that time by
14 parts in 100,000; and five more platinum-silver coils, first
measured in 1894-7, have now a greater mean value by 8 parts
in 100,000.

It would appear also that in many of the variable coils the
changes have occurred mainly, if not entirely, at the soft-soldered
joints, and with a view of testing this the Committee have
authorised the Secretary to open and examine one of the coils.

A comparison, given in the Appendix, has also been made of all
the manganin resistances in the possession of the Standards De-
partment of the Laboratory. The behaviour of the various coils
is somewhat different ; while some have been very constant, others
appear to have changed considerably.

At the Conference on Electrical Units, held at the Beichsan*-
stalt in 1905, it was suggested that the Jena glass 69'" was, fix)m
its good elastic properties, the best glass to employ for mercury
resistances, and accordingly efforts have been made to get some
suitable tubes. Five tubes have recently been secured, after great
difficulty, which will probably do for standards, but the difficulty
of drawing them is a serious obstacle to their use.

A number of tubes of French glass, ' verre dur,' are also on
order.

Progress has also been made during the year in the design of

716 PRACTICAL STANDARDS

the Lorenz apparatus, to be given by the Drapers' Company, and
the manufacture of the bed and the heavy-metal work has been
entrusted to Messrs Armstrong, Whitworth and C!o., who have
kindly undertaken it. The marble cylinders required have been
delivered at the National Physical Laboratory.

Preparations for the holding of an International Congress on
Electrical Units in London in October next are well advanced.

Specifications dealing with the international ohm, the inter-
national ampere, and the Weston cadmium cell, which have been
prepared at the National Physical Laboratory after consultation
with other workers to serve as a basis of discussion at the
Congress, are given in the Appendix with a view of making
them known.

The grant of £50 made to the Committee at Leicester has been
spent in great part in the purchase of materials for the Weston
cells and the silver voltameter research and in obtaining suitable
tubes for use for standards of resistance.

The balance now in hand is £1. Os. 4(i, and the Committee
recommend that they be allowed to retain this for the purpose of
continuing the experiments now in progress.

The Committee therefore recommend that their Reports from
1862 onwards be reprinted, after careful editing, as a memorial
to Lord Kelvin, and that they be reappointed, with a grant of
£100 in addition to the above unexpended balance, for the purpose
of undertaking this work and continuing their researches on the
standards ; that Lord Rayleigh be Chairman and Dr B. T. Glaze-
brook Secretary.

Appendix I.

On the Secular Changes of the Standards of Resistance at the
National Physical Laboratory. By F. E. Smith, A.R.C.Sc.

{From the Natwnxd Physical Laboratory,)

It has long been known that many resistance coils of platinum-
silver, of manganin, and of other resistance alloys do not keep
constant in resistance. The causes of the changes may lie in
some alteration in structure of the alloy, of some change in stndn,
of surface action, of faulty joints, or, as suggested by Dr Rosa, they
may lie in the insulating medium.

FOR ELECTRICAL MEASUREMENTS 71 7

The question of the permanence of manganin standards has
been discussed recently by Messrs Rosa and Babcock* and by
Drs Jaeger and Lindeck-f", and it seemed desirable to bring
together all the information available regarding the changes which
have taken place in the coils of the Association, and of others which
have from time to time been compared with them.

At the National Physical Laboratory the primary standards
of resistance are of mercury, and the secondary standards are of
platinum, platinum-iridium, gold-silver, platinum-silver, and of
manganin. It will be shown that the mercuiy standards have
kept constant, that the platinum coils have probably kept constant,
that the platinum-iridium, gold-silver, and a few of the platinum-
silver coils have changed considerably, while other platinum-silver
coils have kept very nearly constant. Of the manganin coils a
few have kept very nearly constant, but most of them have in-
creased in resistance.

The platinum, platinum-iridium, and some of the gold-silver
and platinum-silver coils are the property of the Association, and
many of them were first compared by Matthiessen and Hockin in
1865-7. Most of the manganin standards were constructed by
O. WolflF, Berlin, but four were built by Mr Melsom at the National
Physical Laboratory. These standards vaiy in nominal value from
one-thousandth of an ohm to 10,000 ohms.

The method of comparing resistances has been dealt with in a
previous Report^, and for the purposes of this communication it
will be sufficient to state that, on all occasions when mercury
standards were erected, the resistance coils were measured in
terms of the mean unit represented by the mercury columns, with
a probable error of about 5 parts in 1,000,000. In the intervals
between the comparisons with mercury standards the values of
the coils in international ohms were at times uncertain within
1 to 2 parts in 100,000, but the relative values of the unit coils of
manganin with potential leads could at all times be determined
with an error not greater than about 2 parts in 10,000,000, and the
one-thousandth ohm and 10,000 ohms manganin standards could
in general be measured in terms of the unit coils within about
5 parts in 1,000,000. In the intervals between the erections of
the mercury tubes a very careful survey of the history of the coils

* The Eleetricianf Jane 14, 1907, and November 15, 1907.

t Ibid., August 2, 1907. t B.A. Report, Section A, 1906.

718

PBACTICAXi STANPARDS

was often necessary to detennine the most probable changes in the
coils, and a slight readjustment of the values allotted to the ooils
was sometimes made when the mercury standards were next
employed. The probable error of the resistance valu^ assigned
to the manganin standards on any date is almost certainly less
than 1 part in 100,000,

Mercury Standards of Registance.

The mercury standards of resistance are 10 in number, and
were constructed in 1902^. The mean international ohm as
I'ealised by the ten standards is taken as the unit, and each tube
is measured in terms of it. In practice two manganin coils act as
intermediaries. The measured values in 1903 and 1907 are given
in the following table : —

Table I.
Owing the Values of the Mercury Standards in 1903 and 1907.

Yalae in Mean International Ohms

Mercoxy
Standard

Difleraaoe
1907-190S

1903
0-971708

1907

0-97 169^

-0^1

1000387

lO0042i

+34

1-000196

1 -000206

+ I1

0-973497

0-973488

-o»

1-001379

1 -001371

-o»

1-001068

l-00106s

-0,

1-000267

1-000265

-0,

l-OOiaOfl

1-001299

-0,

1-00105,

1-001044

-Qs

1-000974

1-000978

-0.

With the exception of P the relative values of the standards
have kept remarkably constant, and in the case of P the increase
in resistance may be apparent only, for only in 1907 has an
increase been noted. It is thought that a very thin film of grease
may be coating a portion of the inner wall of the tube. As the
tubes M, 0, and S are of French verre dur, and the remainder of
Jena 16^'' glass, there is justification for assuming the constancy
of the standards. It is of interest to state that the relative values

FOR ELICTBICAL MEASUREKENTS 719

of the French mercury standards in 1885 and 1905, and of the
mercury standards of the Reichsanstalt in 189S and 1904, are also
in veiy good agreement

Wire Standards of Platinum, PkUinum-Iridium, Oold-Silver,

and Platinum-Silver.

The original coils of the Association are six in number: two
are of platinum, two of platinum-iridium, one of gold-silver, and
one of platinum-silver. They were compared together by Messrs
Matthiessen and Hockin in 1865-67, by Messrs Chrystal and
Saunder in 1876, by Dr Fleming in 1879-«1, by Dr Glazebrook
and Mr Fitzpatrick in 1887-88, and by the author in 1908. In
addition to these six coils, Messrs Chrystal and Saunder examined
a platinum-silver coil marked No. 29 F, and also a coil known as
Flat, while measurements of another platinum-silver coil H are
given in the Report for 1888. These coils — ^in all, nine — have
remained in charge of the Secretary.

In a report to the Association in 1888 Dr Qlazebrook discussed
the probable changes which had taken place in the coils since
1867, and changes in the platinum-silver coils only are discussed
in the Reports for 1892 and 1903. In 1865-67 the probable
error of the comparisons appears to have been of that order which
would be introduced by an error in the temperature of the coils
of about O'l"* C. In 1888 and 1903 the error of the comparisons
corresponds with an error in the temperature of the coils of a
little less than O'l"* C, and in 1908 the error has been reduced so
as to correspond with about 0*02'' C. All of the coils are sur-
rounded by paraffin wax, and it is only by maintaining a constant
temperature for many hours that very accurate observations can be
made. The scale of temperature employed for the 1908 measure-
ments is the hydrogen scale ; that used for previous observations
is almost certainly the Kew glass scale. Dr J. A. Barker has
recently shown* that the difference between these two scales is
negligibly small; hence we may assume that the same scale of
temperature has been used throughout.

The present method of comparing the coils is by substitution
in one arm of a Wheatstone shunt bridge, of which the other three
arms consist of manganin resistances. The high-temperature

♦ Proe. R, 8. A, Vol. 78.

720 PRACTICAL STANDARDS

coefficient coils are kept in a room remaining constant in tem-
perature to 0*01 ** C. over several dajrs, and the temperature does
not differ from 16*0** C. by more than 0*6° C. About 16 measure-
ments, spread over several days, are made of each coil, and the
value at 16*0*" C. is deduced from these measurements. During
1908 approximate values for the temperature coefficients of
resistance of the coils have been obtained by varying the tem-
perature from 14* to 17° C. These values are given in Tables III
and IX,

In 1867 the temperatures are given at which the coils were
1 B. A. unit, and this procedure was in part followed in 1876,
1879, and 1888. The unit of 1867 was, however, probably
different to those of 1876, 1879, and 1888. Messrs Chrystal and
Saunder (1876) assumed one of the coils (B) to have remained
constant between 1867 and 1876, and expressed the values of the
other coils in terms of it. The unit, in terms of which the
measurements of 1879-81 were made, is the mean B.A. unit as
indicated by Fleming on his chart; it is supposed to represent
the mean of the resistances of the six coils A, B,C, Z), JE^ G at
the temperatures at which they were originally correct It is this
unit which was used by Lord Rayleigh in his work on the ohm,
and by Dr Olazebrook since about 1880, and it has been closely
adhered to in all measurements made by the Committee since
that date.

A close examination of the chart at the present day shows
that the mean of the values of the six coils is really about 0*99985
unit; hence if this interpretation be accepted, the mean BA.
unit is really 15 parts in 100,000 less than the unit which has
been taken since 1880; but it has not been thought wise to
attempt any correction on this score, except in the compilation of
Table III. At times a sudden change in a coil has been recorded,
as in 1888, when Dr Glazebrook reported that F had suddenly
risen in value by 000048 B.A. unit*, and that Flat had &Ilen by
1 part in 10,000+. Similar changes may have been observed
when the coils were comparatively new, and it is possible that a
slightly variable coil was disregarded, or a correction applied
because of it, when the chart summarising the observations for
1879-81 was constructed.

♦ B.A. Beport. 1888. t Phil. Trans. A. 1888, p. 864.

FOR ELECTRICAL MEASUREMENTS

721

The chart gives the values of the coils from 0° C. to 26** C, and
the graphs are such that the value of a coil can be read with an
error not greater than 3 parts in 100,000, which is equivalent to a
change in temperature of O'l** C. of a platinum-silver coil. The
resistances of the coils at various temperatures as given by the
chart are given in Table II.

Table II.
Oiving the values of the coils in 1879-81, from FlemiTig's Chart.

Coil

A
B
C
P
E
F
0
Flat

Temperature at
which coil was

stated to be
correct in 1867

Yalae of coil, from

Fleming's chart, at

temperatare given

in 1867

Value of coil, from

Fleming's chart,

atl6 0«>C.

16-0"' C.
15-8' „
16-3' „
16-7' „
15-7' „

15-7 „

1-00011
1-00006
100007
0-99960
1*00010

0-99916

1-00011
1-00036
1-00056
1-00052
1-00102
0-99971
0-99937
1-00034

If we tentatively adopt as the B. A. unit at any date the
exact mean of the resistances of the coils Ay B, (7, Z), E, G at the
temperatures at which they were originally said to be equal, the
values of the coils at 160° C. in 1867, 1876, 1879-81, 1888, and
1908 are as given in Table III. This table has been very easy to
compile, because only the differences between the resistances of
the coils at the various dates, and their temperature coefficients,
were required.

In all the tables of this Appendix the values of high-tem-
perature coefficient coils are given within 1 part in 100,000 ; but
as the errors of observation must often have exceeded the change
in resistance corresponding with a change in temperature of a
coil of 05^ to O'V C, too much significance must not be attached
to an apparent change in resistance, corresponding with a difference
in temperature of a coil of a tenth of a degree.

B. A.

722

PRACTICAL STANDARDS

Table IIL

In this Table it is assumed thai the B.A.U. is equal to the mean of
the coils A, B, C, D, Ey G at the temperatures at whidi they
were found by Hockin in 18G7 to be correct, and that tkU
mean has not altered.

Values at 16-0' C.

Coil

A
B
C
I)
E
F
G
H
Flat

Mat rial

Pt. Ir.

Ft. Ir.

An. Ag.

Pt.

Pt.
Pt. Ag.
Pt. Ag.
Pt. Ag.
Pt. Ag.

Approi.
Tempera-
ture Ou-
efl^ient
(1908)

1867

1876

000148
000148
0-00070
0-00312

o-oosu

0 00027
0-00028
0*Ufl028
000027

1-00000 •
100029 .
1-00050
1-00002
l-00091»j

1-00022
100020

1-00021
1*00069
100085
1-00086
1-00009

0*90074

1879-81 1888

Meaiiof ^,B, C7,i?,J?, (?= 1*00047 1-00047

1-00026
1-00050
100071
100067
100117
0-99086
0-99952

100019

1-00047

I 1-0008S
, 100040

1*001182
I 1-00028

1-00088
' 1*00008
, 0*99061
! 0*9:W78
' 1*00056

1892

1908

1-00051
0-991*25
0*99943
1*00083

1*00047 —

1*00068
0-99975
0-99076
I'OUOSO

1-OOOUi
1-00010'
1-00083,

1-O0004
1-00064 1

1-ooon

1-00087 1
0*99886'
I'OOOST

I
— 1-O0O47

* For a note as to this value see p. 723.

It is clear that changes of very considerable magnitude have
taken place, and the task before us is to select the most constaDt
and the most variable coils. In all such cases a table of difference
values is most helpful. Table IV. gives such values for the six
coils A, jB, C, D,E,0 mix 10-» B. A. units at 160° C.

We conclude from the differences given in column 7 and the
temperatures given in the last column of Table IV. that B and E
have possibly remained constant during the period 1867-1908 and
that C and D are next in order of constancy. The coils D and E
have remained relatively constant since 1876.

Dr Glazebrook in 1888 measured the B. A. unit in terms of

the specific resistance of mercury, and found that the value of the

resistance of a column of mercury, 1 metre long, 1 sq. mm. in

section, at 0°C. was

0-95352 B. A. unit.

For the purposes of the comparison, Dr Glazebrook used the two
coils F and 6r, and their values are given by him as

F = 0-99807 B. A. units at 10° C.
(? = 0-99778 „ „ ,. 10° C.

FOR ELECTRICAL MEASUREMENTS

723

These values were taken from Fleming's chart, and when corrected
to 16° C. they are practically identical with those recorded in
Table II., as they should be. Flat was also used (0*99857 B. A.
units at 10** C), but observations during the two years preceding
1888 showed that it was relatively lower than when examined by
Dr Fleming, and its value was not, therefore, taken from the chart.

Table IV.

Difference Values in 1 x 10

- B.A.U.

Maximum
Difference

The Difference
in Colamn 7
is eqaivalent
to an Uncer-
tainty of

GoUs

1867

1876

1879-81

1888

1908

between

Difference

Values

Temperature
of the Coil
with the
Largest Tem-
perature Go-

-24

efficient of

A-rB

-29

-44

0-6'' C.

A C

-50

-64

-45

-51

0-4* „

A D

-92

-15

-41

.30

147

0-5' „

A E

-91

-78

-91

- 5

-30

O^'' „

A G

-22

122

-53

175

1-2' „

B—C

-21

-20

-21

-42

-75

0-4- „

B D

-63

-17

0-3- „

B E

-62

-34

-67

-48

-54

0-1 rc.

B-0

-77

175

1-2*C.

C~D

-42

123

0-4'' „

C-E

-41

--14

-46 ! - 6

O^'' „

C G

111

119 ; 121

- 2

123

1-8^ „

D E

-63

-50 1 -60

-60

0-2'' „

J)—G

116 , 67

-83

198

0-6'' „

E G

1 .

125

165

127

-23

188

0-6' „

* Hockin (1867 Report) gives the temperatures at which D and E were correct
in 1865, 1866, and 1867. From the values given by him it appears that the
difference D—E was -69x10-' B.A.U. at 160'' G. in 1865, -59xlO-o in 1866,
and 1 in 1867. These differences, taken in conjunction with those given in the
above table, make it practically certain that the difference given for 1867 is
incorrect, and should be replaced by a difference of the order -60.

In 1908 the individual coils were compared with the new
mercury standards set up at the N.P.L. and their values found in
terms of mercury. If we assume that the mean value of the
coils A, B,C, D, E, 0, is the same as when Fleming's chart was

46—2

724

PRACTICAL STANDARDS

constructed, we obtain as the resistance of 1 metre of mercury,
1 sq. mm. in section, at 0** C, the value

0-95333 B. A. units,

an alteration of 20 parts in 100,000 since 188&

If, on the other hand, we suppose that the mercury units set
up in 1908 agree exactly with those constructed in 1888, then the
mean value of the six coils in question has altered by 0*00020
B. A. units. At the present date, assuming as found in 1888, the
resistance of 1 metre of mercury, 1 sq. mm. in section, at 0*^ C. to
be 0*95352 B. A. units, the individual coils have the values given
in Table V., column 3.

Table V.

ValiLea of CoUs at 16'0° C. in 1888 and 1908 obtained from cotn-
parison with Mercury Tubes, assumung ihe Resistance of
1 Metre of Mercury to be 0-96352 B.A.U.

V&lae in 1S8S at Time

CoU

of Determination of
Specific Resistance

Value in 1908

of Mercury *

1-00068

1-00042

1-00026

1-00018

1-00067

1-00093

1-00013

1-00012

1-00073

100072

0-99970

1-00080

0*99936

1-00095

0-99963

0-99964

Flat

1-00023

1-00046

* In Dr Glazebrook's experiments the terminals of the mercury standards were
not exactly at 0° C, and an error of about 4 parts in 100,000 was probably intro-
duced because of this. No correction on this score has, however, been applied, u
the magnitude of the error is only of the same order as the probable error of the
observations.

The apparent changes in resistance of the coils, together with
the alterations in temperature of the coils necessary to produce
equal changes in the resistance, are given in Table VI.

From Tables V. and VI. it appears to be practically certain
that the coils jS, i), E, and H have the same resistance in 1908 as
they had in 1888. The agreement of the values for D and E is

FOR ELECTRICAL MEASUREMENTS

725

very remarkable, for the temperatures at which these coils were
believed to be correct in 1888 are stated to the nearest tenth of a
degree only; an apparent change in resistance 6f 15 parts in
100,000 would, therefore, have been negligible. With respect to
Gy it has risen by over 1 part in 1,000 during the past 5 years
and Flat changed by 17 parts in 100,000 in 1902* The fluctua-
tions in the value of H are believed to have amounted to about
1 part in 10,000 during the period 1888-19081.

Table VI.

ResiBtance of Coil in 1908

Change equivalent to

Coil

mintw

Difference of

Resistance of Coil in 1888

Temperatare of

- 26 X 10 'i B.A.U.

018"'C.

- 7 „

0-06' „

+ 26 „

0-37° „

- 1 ,»

0-00'' „

- 1

0-00" „

+ 110 „

4-0' „

4- 169

6-7° „

+ 1

0-03' „

Flat

4- 22 ,.

0-81' „

Of the four coils J5, D, E^ JT, apparently constant for the
period 1888-1908, we have already concluded from the differences
given in Table IV. that -B, i), and B have remained approxi-
mately constant since 1867. One of the coils Z) — E^ appears, from
Table IV., to have changed in the interval 1867-1876, and the
apparent change corresponds with the change resulting when one
of the coils is lowered 0*2** C. in temperature. It is, however,
practically certain that the change is only apparent. The tem-
peratures at which the platinum coils were stated to be correct in
1865, 1866, and 1867 are given by Mr Hockin in the Report for
1867. They are as follows :

r 1 B. A.U.

at 15T C.

January 7, 1866

Coil No. 35 (Z))... •

1 >»

„ 16-7" C.

August 18, 1866

'■^ «

„ 15-7' C.

February 10, 1867

*■ n

„ 15-5* C.

January 7, 1865

CoUNo. 36(Jgr)....

■^ >J

„ 15-6'C.

August 18, 1866

^1 »>

„ 16-7*0.

February 10, 1867

* B A. Report, :

1903

t IWd.

726

PRACTICAL STANDARDS

In the Report for 1888 the temperature coefficient of D is
given as 000308 B. A. unit, and of jE? as 000302 B. A. unit. These
values agree closely with those given in Table HI., and they have
been used in the compilation of the following complete list of the
difference values (D — E) which now deserves attention :

-1

^=-59x

10-

-6 B.A.U.

16-0'

Year 1865

= -59

1866

1867

»>

= -63

>»

1876

= -50

1879-81

= -60

>»

1888

= -60

1908

The conclusion is obvious. The original difference between
the coils was approximately 60 x 10~~' B. A. unit and has remained
constant ever since. There is little doubt that the difference
recorded for 1867 is incorrect ; it may easily happen that there is
a difference of 0*2** C. between the apparent and true temperatures
of a coil embedded in paraffin wax, and such a difference would *
completely explain the 1867 result.

This conclusion necessitates a revision of the difference values
in Table IV. The corrections are easily made, for the differences
A — E, B — E, C — E and E — 0 should be respectively equal to
the differences A — D, B — D, etc.

We believe that the two platinum coils have remained constant
in resistance since 1867, and that the values in 1867, 1879-81,
1888, and 1908 of these and other coils in terms of the original
B. A. unit (1867) are as follows:

Table VII.

Resistances at 16*0** C. in terms of the origiival B.AM. (1867).
(Values obtained through the two Platinum Coils D, E.)

Coil

Material

1867

1876

1879^1

18*t8
1*00117

1906 ,

Pt. Ir.

1-00000

1*00077

1-00056

1*00122

Pt. Ir.

1*00029

1*00121

1*0008 »

1*00104

1*00096

Au. Ag.

1*00050

1*00141

1*00101

100146

1*00173

Pt.

1*00092

Pt.

1*00152

1*0015S

1*00152

Pt. Aff.

—

1*00016

1*00072

100160

Pt. Ag.

1*00022

1*00030

0*99982

1-00025

1*00175

Pt. Ag.

1-00020

1*00042

1*00044

Plat

Pt. Ag.

100079

1*001*20

1*00125

Mnzimum
Difference

l47x-10*B.A.r
li3

14*

198

»*
46

FOR ELECTRICAL MEASUREMENTS 727

From Tables IV. and VII. it is clear that the maximum
number of coils which can have kept constant is two, and if the
platinum coils have not remained constant then one only of the
other coils can have done so. Since D and E are of pure
platinum, and not of an alloy, it is probable that these would
change least.

If our conclusions are correct, the results are not only of some
value as showing the changes which may take place in the
resistance of certain alloys when embedded in paraffin wax, but
they are also of value because the coils link together so many
determinations of the ohm in absolute measure and of the specific
resistance of mercury. It is not convenient to collect the various
determinations here, but as an instance of the uses to which the
data given in this Appendix might be put we take Lord Rayleigh's
and Mrs Sidgwick's determination in 1881* of the specific
resistance of mercury. It was found that 0*9541 2 B. A. unit was
equal in resistance to a column of mercury 100 cm. long, 1 sq. mm.
in section, at O'^C. Now in Lord Rayleigh's experiments the
terminals of the mercury standards were not at 0** C, but at 5** or
6*'C., and it was shown by Dr Glaze brook f in 1888 that an error
of about 0*00024 was almost certainly introduced because of this.
If we apply a correction of this amount. Lord Rayleigh's value
becomes 0'.96388 B. A. unit as the resistance of 100 cm. of mercury
at 0"* C. The coils F and Flat were used in the 1881 determina-
tion, and the values of these coils were taken fix)m Fleming's
chart. They were therefore:

/*= 0-99971 B.A.U. at 16-0** C. (From Table XL)
Flat= 1-00034 B.A.U. at 160^ C. (From Table II.)

From Lord Rayleigh's observations, therefore,

F at 16-0" C. =0-99971/0-95388 = 104-805 cm. mercury; and
Flat at le-O' C. = l-00034/0-95388 = 104-871 cm. mercury.

At the present time (1908)

F at le-O' C. = 104-959 cm. mercury ; and
Flat at 160*' C. = 104-922 cm. mercury.

Using the 1908 values and the changes in F and Flat, recorded in
Table VII., we conclude that in 1881

F at 16-0" C. was equivalent to 104*808 cm. mercury ; and
Flat at 16-0' C. was equivalent to 104-874 cm. mercury.

• Phil. Trans, Vol. 174. p. 178. t Ibid. A, 1888, pp. 375-6.

728

PRACTICAL STANDARDS

The difference from the values given by Lord Rayleigh is 3 parts
in 100,000, which is less than the probable error of the observations.
We conclude, therefore, that the determination of Lord Rayleigh and
Mrs Sidgwick in 1881 is in excellent agreement with that made at
the National Physical Laboratory in 1908, and this latter has already
been shown to agree with that made by Dr Glazebrook in 1888.

The following is now a very useful summary. The values of
the coils in centimetres of mercury in 1881, 1888, and 1908 are
given in Table VIII.

Table VIII.

Giving ike Values at 16*0° C. of certain Coils in cm, of Mercury
in 1881, 1888, and 1908 obtained from comparisons wiA
Mercury Standards.

1S81

1
1888

1908

Valaes deduced from

Valaes at time of

Valaes

Lord Rayleigh's De-

Dr Glazebrook*8

directly

termination of the

Determination.

Determined

CoU

Specific Besistanoe of

F, Q, and Flat

thrbagh

MA-rfmntn

Mercnry. F and Flat

were nsed ;

N.P.L.

DifFereuce

were used ; for Rela-

for Relative

Mercury

tive Values of Coils

Valaes of Coils

Standards of

see Table VII.

see Table V.

Resistanoe

om.

cm.

cm. j

104-847

104-946

104-918

0H)71

104-872

104-901

104-893

0-029

104*894

104-945

104-972

0-078 i

104-886

104-888

104-887

0-003

104-948

104-951

104-950

0-003

104-805

104-843

104-959

0-154

104-769

104-807

104-974

0-205

104-836

104-837

0-001

Flat

104-871

104-898

104-922

0-051

The preceding comparison strengthens the conclusions already
arrived at respecting the most constant coils. From Table VIII.,
D and E have apparently kept constant in resistance since 1881,
while H appears to have remained constant since 1888.

It is of some importance to note that in 1892 the ratio of the
B. A. unit to the ohm was accepted as being

1 ohm = 1-01358 B.A. unit,

this being based on the values

100 cm. mercury = 09536 B. A. unit,
106*3 cm. mercury = 1 ohm.

FOR ELECTRICAL MEASUREMENTS

729

Other Platinum-siher and Oold'Silver Coils,

In addition to the platinum-silver coils, F, G, H, and Flat,
originally constructed to represent the B. A. unit at a particular
temperature, there are three other platinum-sijver coils, numbered
3716 (Nalder Brothers) and 269 and 270 (Elliott Brothers), made
to represent the ohm = 1 "01 358 B. A. unit. There are also two
10-ohm platinum-silver coils, numbered 288 and 289 (Elliott
Brothers). All these coils are the property of the Association,
and they were extensively used from 1888 to 1903 for the
standardising of other coils. From the results of observations
recorded in the Report for 1903 it appears that from 1894 to
1903 Nos. 3715 and 270 remained constant in resistance, and that
from 1897 to 1903, 288 and 289 remained constant. In 1903 the
N.P.L. mercury standards of resistance were constructed, and
since then the mercury standards have been taken as constant,
and the resistances of all coils expressed by means of them. The
B. A. unit (as obtained frx)m all the platinum-silver coils, taking
the values given in 1888 as correct, and applying corrections for
estimated changes in the coils) was in 1903 found to be equal to
1/1*01367 international ohm. Accepting this ratio for the time
being, the resistances at 16*0** C. of certain coils, compared in
1888, 1894. 1897, 1903, and 1908, are given in the following
table : —

Table IX.

Coil

MAterUU Approz.

OOD-

stmction

Temp.

Co-
efflcient

— |.

•8715

"aw

•270

•288

•280

68 (^)

1C.P.T.

lOC.F.T.

•*
tf

Aa. A|f.
PI. Ak.

Au. Ag.

Pt. Ak.

0*00080

0*00029

0*00032

0*0080

0*0026

0*00013

0*00071

0*00029

000028

000071

OOO.")!

0*0084

0*0080

Resistance

1888

1894

1897

1903

1908

1*00090

100007

—

1*00070

—

1*00089

—

1*00006

—

1*00006

1*00008

—

— .

10-0060

10 0060

10*0096

—

100026

10*0081

0*09076

—

0*99967

000928

—

. —

0-90987

V 999119

—

0*09988

0*99927

1 —

—

0-99941

0*90980

.A.

—

> —

1*00006

9-9996

—

9-9968

9*9941

—

9-9964

9-9984

—

9*9940

Maximum

DifTerenee

(Pkrtsin

100,000)

* The resistances of these coils are gi^en in ohms (1 ohm = 1*01858 B.A. unit).
The remaining ooils have their resistances given in B.A. units.

For the loan of coil No. 64 we are indebted to Professor
Trouton, of University College ; originally this coil was in the

730 PRACTICAL STANDARDS

possession of Professor Carey Foster. For the loan of the coils
numbered 19, 68 (i/), 1 C.F.T., 34, 3, 4, and 10 C.F.T. we are
indebted to Mr H. A. Taylor, of Victoria Street, London. We
tender our hearty thanks to Professor Trouton and Mr Taylor.
All the coils, excepting 19 and 34, are of platinum-silver; 19 and
34 are of gold-silver.

In Table IX. maximum differences of the order 1 to 5 parts in
100,000 may probably be neglected if this maximum difference
does not occur in the period 1903-1908. In 1903 and 1908 the
errors of observation were very small, and a recorded difference
of 1 or 2 parts in 100,000 must be taken as indicating a true
change in the resistance of a coil. The method of measuring a
very small change in resistance will be made clear in the next
section on manganin coils.

The most constant coils appear to be 270, 288, 289, 10 C.F.T.,
3, and 3715. Of these six resistances two only are unit coils ;
the remainder are coils of 10 ohms each. In Table IX. the
values of eight unit coils and of five 10-ohm coils are tabulated,
and of the latter four have kept nearly constant. This fact is
important, as it points to the changes of resistance being largely
due to actions at the soft-soldered joints, and not entirely, if at
all, to the action of paraffin wax (possibly acid) on platinum-silver.
In addition, part of the changes may be due to change in structure
of the alloy.

The values at 160' C. of the coils A, B, C, D, E, F, G, H and
Flat, in terms of the unit of resistance employed for the purposes
of Table IX., are approximately

i4 = 1-00060 F =1-00088

^=l-00026 Q =1-00103

C= 1-00101 H =0-99972

/) = 100020 Flat =1-00053
^=1-00080

Manganin Standards of Resistance.

The manganin standards of the National Physical Laboratory
are in constant use and have proved of very great value. They
not only facilitate electrical measurements, but they bring them
to a far higher degree of accuracy than was formerly attainable.
Nevertheless, the variations in these resistances have in many
cases been a source of trouble, and attempts have been made, and

FOR ELECTRICAL MEASUREMENTS

731

are being continued^ to constract standard coils of manganin which
shall remain practically constant in resistance.

Since 1903 the manganin standards have been intercompared
at least four times eveiy year, and the probable changes have been
deduced from occasional comparisons with mercury standards and
from tables of difference values, due regard being also paid to the
past history of the coils. As an example of the comparisons we
take the case of seven 1-ohm coils which were intercompared in
January, April, July, and October 1906. The observed differences
are given in Table X.

Table X.

Differenoes in 1 x IQ-^ ohm at 17-0'' C.

1690-780

1690-2361

1690-2483

1690-381

169a-Z17

1690-Z18

Jan. 1906

April 1906

July 1906

Oct. 1906

6-82

6-87

7-05

5-93

217

2-36

2-04

1-47

0-26

0-48

0-26

-0-30

-5-80

-6-05

-5-97

-6-27

16-63

13-41

12-69

11-80

16-69

16-05

15-54

14-92

Maximum
Difference

112
0-88
0-78
0-47
3-83
1-67

Any one of these differences was not obtained from a single
observation, but is the mean of six differences. All possible com-
binations of the seven coils were taken — 21 in all — and the
differences observed. From these 21 observations six values
resulted for the difference between any two of the coils ; it is the
mean of these six values which is recorded. The temperature
during the observations was very nearly 17*0** C, and the differences
were corrected to I?'' C. before taking the mean.

An analysis of the figures given in Table X. indicates that the

coils L'l7 and L'18 probably changed most during 1906, and that

the other five coils changed by amounts less than 3 parts in

1,000,000 from January to July 1906. From July to October the

difference 1690-780 changed by an appreciable amount and the

differences in the values for July and October, viz.:

1690- 780 Change- M2x 10"* ohm
1690-2361 „ =-0-57

1690-2483 „ =-0-56

1690- 381 „ = -0-30

>»

732

PRACTICAL STANDARDS

JAN.
03

1908

JAN

JAN.

JAN

1904

I90S

1906

JAM

i90r

JAM
00

tsoc

o
o
o

I .

o
o
o
o

Chabt I: Showing the Variations in Besistanoe of Manganin Standard Coils
of Nominal Values, 1 ohm, 10 ohms, and 100 ohms.

FOR ELGCTBICAL HEASUREMEMTS

733

OAN

- '05 I90S '04

1904

JAN
OS

1905

JAN

'oe

1906

JAN
07

i9or

JAN
08

1909

i -

o
o
o
o

o
o
o

■^ *

V^i^

OHAt

^ijV

— r —

■mi

o-i

OHM

assa

J-

•

O'OI

OHM

"5^93^

^^^^^^^^^^^

O^^*^

»aoO^

d^s>-

2199

«49S

OOOl

OHM

ooo»

OHM

r^H^**"

"r^o

■

-rsS"

--"

/•
It

} ^

Chart II: Showing the Variations in Bedstance of Manganin Standards of
Nominal Yalaes, 0*001, 0*01, 0*1 ohm» and 1,000, 10,000 ohms.

734

PRACTICAL STANDARDS

indicate that 1690 probably fell in resistance in this period by about
6 parts in 1,000,000, and 780 rose by about 5 parts in 1,000,000.
The other small changes are difficult to assign and are possibly
due to variable humidity. The errors in the differences recorded
are certainly less than 1 x 10"* ohm.

The above is only part of the analysis of the differences which
is in general made. Comparisons with coils other than units are
also often desirable, but need not be dealt with here.

Table XI.
Resistances in International Ohms at 170'* C.

Resistant
Standard

O.W. 2196
O.W.2483

O.W. 2200
O.W. 24»2

O.W. 2852
O.W. 2484

O.W. 1690
O.W. 780
O.W. 381
O.W. 248*
O.W. 2351

il7

2.-18

O.W. 788
O.W. 1683

Z--19

i-20

O.W. 739
O.W. 2460

O.W. 740
O.W. 2449

O.W. 2448

Nominal
Value

0*001 ohm

0-01 ohm
tt

0*1 ohm
»»

1 ohm

»»
•»

10 ohms

100 ohms

1000 ohms

10000 ohms

Oct. 1903 Oct. 1904 ' Oct. 1905

00999984

1*000044
0*9999.19
1 -00009a
1 -000004
0*99999,
0-99985a
p'99984o

9•89^74
10-0000,

99*999g
99-9950

1000-15,
1000*01,

10000-24

0*00090996g
O'OOlOOOOOo

0-000e9997a
0*001000014

0*0100014« 0*0100016b
O'OlOOOll, , O'OlOOOllo

0'099998o
0*100009o

l*00002s
0*99994,
1*00009,
l-OOOOOf
0-99998j
O-99F864
0*999858

9-99878

lO-OOt)?,

9-9993i

9 -9995 J

99-999,
lOOOOOa

1000-17,
1000*244

10002*4.

, 0*0999984
O'lOOOU^

l*0000?s
0*999041)
I-OOOO84
1-000024
I'OOOOdj
0*99987s
0-999858

9*99864

10*0001,

9*9997,

9*9096,

99 999a
100-004^

1000-21,
1000*494

lOOOS-57

Oct. 1906

0-OOlOOOOOa

a*ooioouoi«

0-01000207
0*01000097

0*0999984

0*1000114

1*000025

0-99996^
1*00008,
1*000024
1*000014
0*99990,
0-99968s

9*9985,
10-0002,

9*QQUB

9*QUOQ
VtfWff

09988,
100*000,

1000'24«
1000*66,

10003*8,

Oct. 1907

0-001000l4a
0*00100001 4

0*01000414
0-0100009,

0-0909S9,
0*1000174

1*00000;

0-990964
1-00009,
l-OOOO:.',
I*0li002«
0*9090<,
0*99988»

9*90879
10*0004?
10-00037

99-960L
100-013,

1000*26,
1000*81 4

IOOO8-74

Table XI. gives the resistance of a number of manganin coils
in the October of each year from 1903 to 1907, and charts Nos. 1
and 2 show the complete changes in most of the coils from March
1903 to June 1908. In Table XL the resistances are given in the
same month of each year in order to eliminate from the table (as
much as possible) the effects of humidity on the resistances of
the coils.

Resistances L'19, 2448, and 2449 were placed in atmospheres
of varying humidities in the interval, October 1907-April 1908,

FOR ELECTRICAL MEASUREMENTS

735

and hence the curves for these coils are not continued on the
charts after January 1908,

At first limiting our attention to the unit coils, we see from
the charts that these have varied during the past five years by the
following amounts : —

Table XII.

Mazimom Change in

Diflference Value.

Coil

Resistance in

Resistance in 1908 mintu

1690

5 years

Resistance in 1903

3-7 X 10-6 ohm

3-7 X 10-* ohm

780

4-6

2351

7-8

-0-2

2483

4-4 „

3-2

381

2-6

-0-2

Z17

10-2 „

9-6

Z-18

8-8 „

S-8

Mean = 6-0 „

Mean =+4-2 „

If we neglect L'l7 and L'lS the mean value of the other five
coils is 2*2 X 10"» ohms greater in 1908 than in 1903.

Apart from the cause of these changes, it is interesting to
form soine idea of what interpretation of the differences might
reasonably have been applied if mercury standards had not been
the master standards. If the mean value of the seven coils had
been taken as remaining constant, the error in five years would
have amounted to 4*2 parts in 100,000. A comparison with coils
of nominal values differing from unity might, however, be made,
and such might largely influence the result.

The maximum changes which have taken place in the other
resistance standards and the difference values (1908-1903 values)
are given in Table XIII.

The mean difference in the values of all the manganin coils
for 1908 and 1903 is 12*6 parts in 100,000. The oldest coils are
581 (seventeen years old), 780, 738, 739, and 740 (thirteen years
old), and 1093 and 1690 (eight years old), the ages being
approximate only. The remainder of the coils are trom five to
jsix years old.

The most constant coils belonging to various groups are : —

381 — most constant of the unit coils.

738 „ „ 10 ohms coils.

739 „ „ 100

740 „ „ 1,000

736

PRACTICAL STANDARDS

Table XIII.

Resistance
Standard

O.W. 2196
O.W. 2493

O.W. 2200
O.W. 2492

O.W. 2352
O.W. 2484

O.W. 738
O.W. 1693

Z19

X-20

O.W. 739
O.W. 2450

O.W. 740
O.W. 2449

O.W. 2448

Nominal
Value

0001 ohm
0-001

0-01
0-01

01
01

>»

10 ohms
10
10
10

100
100

1000
1000

10000

»>

Maximum

Change since

1903.

Parts in

100,000

22-4
2-0

33-0

1-9

2-0
8-0

2-2

110

8-3

1-2
18-0

11-8
89-4

40-0

1908 Value
minus

1908 Value.
Parts in
100,000

22-4
1-8

33-0
0-2

1-4
8-0

1-7
7-2
9-6*
7-6*

-lO
18-0

11-4
89-4

36-8

Mean

Difference.

Parts in

100,000

+ 121

+ 16-6

+ 4-7

+ 6-5

I + 8-5

I +50-4
+ 36-8

Mean di£ference value (1908-1903 values) = +16*5 parts in 100,000.

Mean difference value (1908-1903 values) including the unit

coil8= +12-6 parts in 100,000.

These are the difference values (1908-1904 values).

In general, therefore, the older the coil the more constant does
it appear to be.

With reference to the sudden changes in resistance, as shown
by the curve for 2351 in 1903, of 381 in 1904-6, and of Z-20 in
1906, we can offer no complete explanation ; but it is possible that
variable humidity of the surrounding medium, such as might arise
from the presence of a small quantity of moisture in the insulating
oil, was responsible for part of these changes.

The breaks in the curves for 2483, 2351, Zl7, Z-IS, 1693, and
2484 are due to these coils being away from the National Physical

FOR ELECTRICAL MEASUREMENTS 737

Laboratory ; they were being compared with the wire standards of
the Reichsanstalt.

The increase in resistance of No. 2449 is phenomenal. The
daily rate of change for 1906 is over four parts in 10,000,000;
that is, in about twenty- two days the coil changed in resistance by
about one part in 100,000. In April 1907 we attempted to
measure the change from day to day, and for this purpose we
compared 2449 and 740 every working day for four weeks. The
results obtained are as follows : —

Day of Obeerration ... 1 S 5 8 10 12 16 17

^"''^pi^^m 100^0^ 1 ^^*^^ ^^'^^ ^^'^ ^^'^^ ^^'^^ ^^'^ ^^'^^ ^^'^
Day of Obfierration ... 19 22 24 26 29

^"'^^^iffl 1 52-75 62-95 53-00 53-00 53.16

The change was, therefore, a very gradual one, and easily
detected. It is of interest to note that the rate of change for the
last six months of 1907 is less than that for 1903-6.

The possible causes of the changes in the manganin resistances
may be classified under the following heads : —

1. Change in structure of the alloy.

2. Surface action.

3. Humidity eflfect.

4. Change in the soldered joints connecting the wires of

high-resistance coils to the current leads.

5. Change at the junctions of the potential leads with the

resistance standard.

Only the first of these appears to fully explain the gradual
rise in resistance. Causes 2 and 4 would have an inappreciable
eflfect on veiy low resistances; yet some of these — e,g., 2196 —
have changed by considerable amounts. Cause No. 5 would have
no eflfect on high-resistance coils, since these are not provided with
potential leads; but Table XIII. shows that all of the high-
resistance coils have changed. Cause No. 2 produces in general
a cyclic change, and, while being without doubt a cause of variation,
it cannot be modified to explain all the gradual increases in
resistance, owing to the negligible eflfect of humidity on very low-
resistance standards. Cause No. 1 appears, therefore, to have been
the chief agent in the cases we have considered.

It is necessary, however, that we should say something about

n. A. 47

738 PRACTICAL STANDARDS

other manganin coila In 1903 the resistances were measured of
some manganin coils (1 to 5,000 ohms) in a box by R W. Paul,
London. The coils could not readily be immersed in oil, and the
measurements were therefore uncertain to about 1 part in 100,000.
The resistances were again measured in 1904, 1906, and December
1907. The maximum change in the resistance of any coil is
5 parts in 100,000, while the mean increase in resistance during
1903-7 is 4 parts in 100,000.

In 1902, and again in 1907, the resistances were measured of
some manganin coils (1 to 10,000 ohms) in box No. 1723 by
O. WolflF, Berlin. The maximum change in resistance during the
period 1902-8 is about 6 parts in 100,000, and a few of the coils
have kept practically constant. Many manganin coils in other
boxes are known, however, to have changed very considerably.

It will be seen that of the manganin standards we have
examined some have kept remarkably constant, while others are
practically useless as standards. It must not be concluded, how-
ever, that all manganin resistances are subject to such changes.
Drs Jaeger and Lindeck have shown that the manganin standards
of the Reichsanstalt keep very constant, and the manganin coils
at the Bureau of Standards also appear to be of a fairly constant
type, though subject to considerable cyclic changes owing to
variable atmospheric humidity. The manganin standards reported
on in this Appendix comprise every standard resistance of nianganin
in use in the Standards Department of the National Physical
Laboratory.

Appendix II.

Specifications for the Practical Realisation of the Definitions of
the International Ohm and International Ampere, and
Instructions for the Preparation of the Weston Cadmium

Cell.

{From the National Physical Laboratory.)

The following specifications have been prepared after consulta-
tion with various authorities, and will form a basis for discussion
at the forthcoming Congress on Electric Units in London. They
have not been authoritatively adopted, and are subject to
amendment.

FOR ELECTRICAL MEASUREMENTS 739

In the last Report specifications for the realisation of the inter-
national ampere and for the construction of the cadmium cell
were given, the processes of preparation, etc., being described with
considerable detail. These specifications appeal to a much wider
circle than the present ones, for the latter are intended mainly to
serve as a guide to the standardising institutions of the various
countries in order to obtain, as feir as possible, complete agree-
ment in the units of electric measurements. Certain instructions,
such as the purification of mercury, have therefore been omitted,
but all which are thought* to be essential for an exact reproduction
of conditions are still included. Instinictions for the erection of
mercury standards have not previously been issued.

The International Ohm.

The international ohm shall be equal to the resistance offered
to an unvarying electric current by a column of mercury at the
temperature of melting ice, 14'4521 grammes in mass, of a con-
stant cross sectional area, and of 106300 centimetres in length,
arranged in accordance with the following specification.

The column of mercury shall be of circular section, or nearly
80, and shall be contained in a tube of suitable glass which has
been carefully annealed. The tube shall be straight to the eye,
and the maximum variation in its area of cross section shall not
exceed 2 parts in 100. The tube is to be carefully calibrated, and
the correction for its conicality determined.

In determining the weight of mercury contained by the tube
when filled at the temperature of melting ice, the column of
mercury is to be bounded by planes at the terminal cross sections
of the tube. The tube should not be unduly heated, and it
should be filled with mercury by exhaustion of air.

The axial length of the tube should be measured at 0^ C. if
possible, otherwise the coefficient of expansion of the glass should
be determined and the axial length of the tube at 0"" C. calculated
from axial measurements made very near to that temperature.
To facilitate measurements of the axial length, the ends of the
tube should be ground very slightly convex.

For the electrical measurements the ends of the tube are to
be connected to spherical bulbs of glass, the slightly convex ends
of the tube forming, very nearly, portions of the internal spherical

47—2

740 PRACTICAL STANDARDS

surfaces of the bulbs. Each bulb is to be provided vdth a current
and a potential lead, the point of entry of the former, and an end
of the tube being at opposite ends of a diameter of the bulb. The
potential lead shall be situated in a plane midway between the
point of entry of the current lead and the end of the tube, and at
right angles to the line connecting them.

Contact with the mercury shall be made by means of platinum
wires.

The diameter of a bulb is to be from 30 to 33 times the
diameter of the terminating section of that end of the tube to
which it is connected.

If 2/ is the axial length in centimetres of the mercury column
contained by the tube at 0° C, W the weight of the column in
grammes, and fi the correction for the conicality of the tube, the
resistance of the column at 0^ C. is

L* 14*4521 L^

^(iWSOO^ ' — W~~ "^ 0001278982/A ^ international ohms.

When the spherical bulbs are fitted to the ends of the tube
and the whole filled with mercury, if r is tHe mean radius of the
tube and ri, r,, the mean radii in centimetres of the terminal
sections, the resistance at O^'C. between the potential leads is

^-2— international ohms.

0-001278982 ^ ^

0-80r«

/*+ — r

correct to 1 per cent, of the added resistance

0-001278982 ^ jo-80r« (p;^)]* -

The electrical measurements are to be carried out at 0°C., the
tube and spherical vessels being surrounded by melting ice and
about 16 centimetres below the upper surface of the ice. The
connecting wires employed for the current and potential leads
must be thin, the flow of heat through them to the mercury
being insufficient to produce appreciable error by the warming of
the mercury.

The insulation resistance between the mercury column and the
ice surrounding the tube must not be less than 10,000,000 ohms.

The current employed in comparing the mercury resistance

* The end correction factor U given in these formalaa as 0*S0 : this Taloe is.
howeyer, subject to amendment.

FOR ELECTRICAL MEASUREMENTS 741

^vith other resistances shall be limited by the condition that the
mercury shall not be warmed sufficiently to produce appreciable
error.

The mean of at least five tubes must be taken to determine the
value of the mercury unit.

The mean of at least three fillings shall be taken as the value
of the resistance of a tube.

Specification for the Practical Application of the Definition of the

International Ampere.

Conditions under which silver is to be deposited to measure
currents from 0*5 to 8 amperes: —

The solution shall consist of frt)m 15 parts to 25 parts by
weight of pure crystallised silver nitrate in 100 parts of distilled
water free from chlorine. It shall be used for one determination
of current only.

In cases in which it is desired to measure a current of about
1 ampere the anode shall consist of a disc or plate of pure silver
about 60 square centimetres in area and 3 or 4 millimetres in
thickness. It is supported by a silver rod riveted through its
centre. The anode shall be inserted into a cup of filter-paper
separately supported.

The kathode shall consist of a platinum bowl about 10 centi-
metres in diameter and 7 centimetres in depth.

About 300 cubic centimetres of the silver-nitrate solution are
to be placed in the kathode bowl, and the anode is to be supported
near the top of the solution and is to be just covered by it. Not
more than from 7 to 10 grammes of silver should be deposited.

(For the measurement of smaller currents, say from i to ^
ampere, a bowl holding about 60 cubic centimetres of solution
may be used, the anode being proportionately reduced in size and
from 2 to 3 grammes of silver being deposited.)

The deposit should be rinsed with distilled water fi^e fit)m
chlorine until the addition of a drop of neutral solution of sodium
chloride in water, to the wash water, produces no milkiness. The
kathode bowl is then nearly filled with distilled water and left for
at least three hours ; it should be rinsed three times, the last of
these wash waters remaining in the bowl for ten minutes. This
last wash water should give no milkiness when added to a neutral

742 PRACTICAL STANDARDS

solution of sodium chloride in water. The deposit is to be dried
in an electric oven at a temperature of about IGO'^C; it is placed
in a desiccator to cool, and is afterwards weighed.

The mass of the deposit, expressed in grammes, divided by
the number of seconds during which the current has been passed
and by O'OOlllS, gives the mean current in amperes.

Preparation of the Weston Cadmium Standard CelL

The cell has mercury for its positive electrode, and an amalgam
consisting of from 12 to 12*5 parts by weight of cadmium in 100
parts of the amalgam for its negative electrode. The electrolyte
consists of a saturated solution of cadmium sulphate, and solid
cadmium sulphate is contained within the cell. A paste, con-
sisting of solid mercurous sulphate, mercury, and solid cadmium
sulphate, rests on the positive electrode.

For the positive electrode, pure distilled mercury should be
uSed.

The amalgam may be made either by electrodeposition or by
mechanical mixing. It should be fused and freed from oxide by
washing with dilute sulphuric acid.

For the preparation of the cadmium sulphate crystals and
solution, commercially pure recrystallised cadmium sulphate should
be dissolved in pure distilled water so as to form a clear saturated
solution. Evaporation at about 35* C. is then allowed to proceed,
when crystals separate from the solution. The crystals are washed
with successive small quantities of distilled water, and part of
them is dissolved in distilled water to form a saturated solution.
The solution should be neutral to congo red.

The mercurous sulphate should be quite pure, and its crystals
should not be so small as to have an abnormal solubility or so
large as to be inefficient as a depolariser. The following is an
example of a method for preparing the salt satisfactorily : —

Add 15 cubic centimetres of pure- strong nitric acid to 100
grammes of pure mercury, and place on one side until the action
is over or nearly over. Transfer the mercurous nitrate thus
formed, together with the excess of mercury, to a beaker con-
taining about 200 cubic centimetres of dilute nitric acid (1 volume
of acid to about 40 volumes of water); a clear solution should
result. Prepare about 1 litre of dilute sulphuric acid (1 volume

FOR ELECTRICAL MEASUREMENTS 743

of acid to 3 of water), and while the mixture is hot add the acid
mercurous nitrate solution to it. The solution should be added
as a very fine stream from the narrow orifice of a pipette, and the
mixture violently agitated during the mixing. Mercurous sul-
phate is precipitated. Decant the hot clear liquid and wash
the precipitate twice by decantation with dilute sulphuric acid
(1 volume of acid to 6 of water). The precipitate should then be
filtered and washed three times with dilute sulphuric acid (1 to 6),
and afterwards 6 or 7 times with saturated cadmium sulphate
solution to remove the acid. The mercurous sulphate should then
be flooded with saturated c€ulmium sulphate solution and left for
one hour, after which the solution is tested with congo red paper.
In general no acid will be detected, and if so the mercurous
sulph&te is ready for use.

To set up the cell the H form of vessel is the most convenient.
The platinum wires inside the vessel should be amalgamated by
passing an electric current to each in turn through an acid
solution of mercurous nitrate. The vessel must afterwards be
washed out twice with dilute nitric acid and several times with
distilled water; it must be free firom stains and scrupulously
clean ; it is dried by the application of heat. The amalgam is
fused and its surface flooded with very dilute sulphuric acid;
sufficient of it to completely cover the amalgamated platinum wire
should then be introduced into one of the limbs of the H vessel.
To firee from acid the amalgam may be remelted and washed with
distilled water. Into the other limb of the vessel sufficient mer-
cury is introduced to completely cover the amalgamated platinum
wire. Then the paste, finely powdered crystals of c€ulmium
sulphate, and saturated cadmium sulphate solution are added in
the order named and the cell sealed.

Its electromotive force at 20° C. is I'OlSj volt.

The electromotive force at any other temperature (t) may be
obtained from the equation: —

Et = 1-0185 - 0000038 (t - 20) - 000000065 (t - 20)«,
the limits of temperature being — (these have not yet been fixed).

THIRTY-SIXTH REPORT— WINNIPEG, 1909.

Appbkdix.— ii^por/ of the JntemcUional Conference on Electrical

Units and Standards, London^ 1908 p. 748

The Committee desire in the first place to record their sense
of the great loss electrical science has sustained since their last
meeting by the death of Professor Ayrton, F.R.S. The revival
o^the Electrical Standards Committee was proposed by him at
the Swansea Meeting in 1880. He had been a member since
that date, and much of the work of the Committee owes its
initiation to his inspiration. He contributed greatly to the
success of the Ayrton-Jones ampere balance, and was deeply in-
terested in the preparations for the Lorenz apparatus now being
erected at the National Physical Laboratory as the giit of the
Drapers* Company in memory of Professor Viriamu Jones. The
Committee will miss in no small degree his keen criticism and his
active help.

The International Conference on Electrical Units and Standards,
referred to in previous Reports, met, on the invitation of H.M.
Government, in the rooms of the Royal Society, from October 12
to October 22, 1908. It was attended by forty-six delegates,
representing twenty-two countries and four British dependencies.
The Report of the Conference is printed as an Appendix to this
Report. In accordance with one of the resolutions passed by the
Conference, Lord Rayleigh, as Chairman, appointed a committee
of fifteen to advise as to the organisation of a permanent Com-
mission, to formulate a plan for and to direct such work as may be
necessary in connection with the maintenance of standards, fixing
of values and intercomparisons of standards, and to comjJete the
work of the Conference.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 74o

The work of this Committee is now in progress, and it is
proposed that representatives of the National Physical Laboratory
and of the Reichsanstalt should visit Washington this autumn.

In their last Report the Committee suggested the republication
of the Reports of the original Committee fix)m 1862 to 1871, and
of the present Committee fix)m 1881, as a memorial to the con-
nection of Lord Kelvin with their work. They are glad to learn
that the recommendation from the Committee of Section A in
favour of this course has been accepted by the Council, and that a
proposal to undertake the projected republication will be made to
the General Committee at Winnipeg.

The Committee are greatly indebted to Mr R. K. Gray for a
generous donation of £100 towards the expenses of this work.

In the Appendix to the Report of the Committee for 1905 it
is stated that slight variations in the electromotive force of the
Weston normal cell can be produced by 12^ per cent, cadmium
amalgam. A preliminary investigation showed that the variations
were generally very small and not easily reproduced. In gen&al
the electromotive force was normal at 0° C. A more exhaustive
investigation has now been completed at the National Physical
Laboratory, and the results show that in general the 12^ per cent,
amalgam may be used from 0° C. to above 60° without any appre-
ciable error, but the E.M.F. of a standard cell containing such an
amalgam may he very abnormal at all temperatures below 12° C.
The limits of temperature for the general use of a 12| per cent,
amalgam are very nearly 12° C. to 62° C. A 10 per cent, amalgam
may be used between 0° C. and 51° C.

Progress with the Lorenz apparatus has been slow but satis-
fectory. The difficulties attending the driving have been over-
come to a considerable extent : an electric motor will be installed.
The iron of the motor has been demonstrated to have no appre-
ciable effect on the mutual induction of the Lorenz system when a
small addition to the electrical system is introduced.

A comparison between the standards of resistance used at the
National Physical Laboratory, the National Bureau of Standards,
and the Physikalisch-Technische Reichsanstalt has been made by
the use of some hermetically sealed standards belonging to the
Bureau of Standards. The following tables give the results
obtained : —

746

PRACTICAL STANDARDS

Table I. — Giving the Results of Comparisons made February —

March 1908. Values at 20° C.

No. of Coil

Resistanoe as determined at

1
DifFerenoe
Parts in 100,000

1
f

N.B.S.
(Jan.)

N.P.L.
(Feb.)

P.T.R.

(Mar.)

N.P.L.-
N.B.S.

N.P.Ii.-
P.T.B.

1. (B.S. 1102a)

2. (B.S. 1102 b)

3. (B.S. 1102c)
12. (B.S. 2415D)

1. (B.S. 3946 E)

2. (B.S. 3946 F)

1. (B.S. 39461)

2. (B.S. 3946 J)

0-99980,
0-99975o
l-OOOOOo
0-999982

99-99QS
99-9858

999-9O2
lOOO-Olg

0-999820
0-999776
1-000022
0-999997
99-9916
99-987,
999-92o
1000-036

0-999816
0-99975e
l-OOOOlo
0-999976
99-991fl
99-9872

2-8
2-6
2-2
1-5
1-2
1-3
1-8
1-6

1-3

1-7

1-2

2-2

-01

Mean... 1-9

1-0

Table IL — Giving the Results of Comparisons made November 1908

to March 1909. Values at 20° C.

No. of Coil

Resistance as determined at

Difference
Parts in 100»000

N.B.S.
(Sept. 1908)

fNo^^iaOft^ (Nov. 1908-
(Nov.1908) Van. 1909)

N.P.L.

(Feb.-Mar.

1909)

N.P.L.-
N.B.S.

1
N.PX.-

p.T.a

L (B.S. 1102a)

2. (B.S. 1102b)

3. (B.S. 1102c)

4. (B.S.1102D)

11. (B.S. 5315 c)

12. (B.S. 5315 D)

1. (B.S.3946E)

2. (B.S.3946F)

0-999997
0-999997
1-OOOOOb
0-999994
0-99999,
0-99998,
99-9916
99-9876

1-000023
1-000022
1-000016
1-00001,
1-000014
l-OOOOOj
99-9946
99-99O7

0-99999»
l-OOOOOg
1-999989

l-OOOOlg
1-000026
1 -00001 1

2-6
2-5
1-6
21
2-9
2-6
2-9
3-2

1-6
1-2
1-8

Mean...

2-5

1-5

FOR BLECTRICAL MEASUREMENTS 747

The unit coils Noe. 1 and 2 were adjusted at Washington on
September 23, 1908, so as to have values more nearly equal to the
nominal. The changes made were + 0*0001 9« ohm and + 0-00024
ohm respectively.

Analysis of all the data relating to the comparisons indicates
that the coil No. 11 (Table II.) changed by about 000001 ohm
during transportation £rom Teddington to Charlottenburg.

No. 12 is a comparatively new coil, having been sealed in
January 1908.

At the Bureau of Standards (Washington) wire coils were
employed as standards in all the comparisons.

The values given by the N.P.L. in Table I. are in terms of the
N.P.L. mercury standards of resistance, which were set up in
November and December 1907. The N.P.L. values in Table II.
are in terms of the mercury standards of resistance which were
erected in February 1909.

With respect to the values given by the Reichsanstalt, in
Table I., Dr Lindeck states, " The last complete series of measure-
ments on the standards of the Reichsanstalt was carried out at the
end of January and the beginning of February. The values given
in the Table (I.) are based upon this series."

In Table II. the P.T.B. values depend upon the values assigned
to a wire standard of the Reichsanstalt which had been kept for
about a year in an atmosphere of constant humidity, and frequently
compared with other standards of resistance.

In conclusion the Committee recommend that they be reap-
pointed for the purpose of continuing their researches on the
standards and carrying out the republications of the Reports if
sanctioned by the General Committee, and that Lord Rayleigh be
Chairman and Dr R. T. Glazebrook Secretary.

748 PRACTICAL STANDARDS

APPENDIX.
International Conference on Electrical Units and Standards, 1908.

Report.

The Conference on Electrical Units and Standards, for which
invitations were issued by the British Government, was opened
by the President of the Board of Trade, the Right Hon. Winston
S. Churchill, M.P., on Monday, October 12, 1908, at Burlington
House, London, W.

Delegates were present from twenty-two countries, and also
from the following British Dependencies — namely, Australia,
Canada, India, and the Crown Colonies.

It was decided by the Conference that a vote each should be
allowed to Australia, Canada, and India, but a vote was not claimed
or allowed for the Crown Colonies.

The total number of Delegates to the Conference was forty-six,
and their names are set out in Schedule A to this Report.

The officers of the Conference were : —

President,

The Right Hon. Lord Rayleigh, CM., President of the Royal
Society.

Vice-Presidents.

Professor S. A. Arrhenius. M. Lippmann.

Dr N. EgorofiF. Dr S. W. Stratton.

Dr Viktor Edler von Lang. Dr E. Warburg.

Secretaries.

Mr M. J. Collins. Mr C. W. S. Crawley.

Mr W. Duddell, F.R.S. Mr F. E. Smith.

The Conference elected a Technical Committee to draft speci-
fications and to consider any matter which might be referred to
the Committee and to report to the Conference.

The Conference and its Technical Committee each held five

sittings.

FOB ELECTRICAL MEASUREMENTS 749

As a result of its deliberation the Conference adopted the
resolutions and specifications attached to this Report and set out
in Schedule B, and requested the Delegates to lay them before
their respective Governments with a view to obtaining uniformity
in the legislation with regard to Electrical Units and Standards.

The Conference recommends the use of the Weston normal
cell as a convenient means of mectsuring both electromotive force
and current when set up under the conditions specified in
Schedule C.

In cases in which it is not desired to set up the standards
provided in the resolutions Schedule B, the Conference recom-
mends the following as working methods for the realisation of the
international ohm, the ampere, and the volt : —

1. For the International Ohm,

The use of copies, constructed of suitable material and of
suitable form verified fix)m time to time, of the inter-
national ohm, its multiples and submultiples.

2. For the International Amphre.

{a) The measurement of current by the aid of a current
balance standardised by comparison with a silver
voltameter; or

(6) The use of a Weston normal cell whose electro-
motive force has been determined in terms of the
international ohm and international ampere, and of a
resistance of known value in international ohms.

3. For the International Volt

(a) A comparison with the difference of electrical
potential between the ends of a coil of resistance of
known value in international ohms, when carrying a
current of known value in international amperes ; or

(6) The use of a Weston normal cell whose electro-
motive force has been determined in terms of the
international ohm and the international ampere.

The duties of specifying more particularly the conditions under
which these methods are to be applied has been assigned to the
Permanent Commission, and, pending its appointment, to the
Scientific Committee to be nominated by the President (see
Schedule D), who will issue a series of Notes as Appendix to this
Report.

750 PRACTICAL STANDARDS

The Conference has considered the methods that should be
recommended to the Governments for securing uniform adminis-
tration in relation to electrical units and standards, and expresses
the opinion that the best method of securing uniformity for the
future would be by the establishment of an international electrical
laboratory with the duties of keeping and maintaining inter-
national electrical standards. This laboratory to be equipped
entirely independently of any national laboratory.

The Conference further recommends that action be taken in
accordance with the scheme set out in Schedule D.

Signed at London on October 21, 1908,

by the Delegates of the Countries above written.

For the United States of America. S. W. Stratton, Henry S.
Carhart, and Edward B. Rosa.

For Austria. — Viktor von Lang and Ludwig Eusminsky.

For Belgium. — P. Clement.

For Brazil. — Leopold J. Weiss.

For Chile. — Victor Eastman.

For Colombia. — Jorge Roa.

For Denmark and Sweden. — Svante Arrhenius.

For Ecuador. — C. Nevares.

For France. — G. Lippmann, J. Ren^ Benoit, and T, de
Nerville.

For Germany. — E. Warburg, W. Jaeger, and St. Lindeck.

For Great Britain. — Rayleigh, J. Gavey, R T. Glazebrook,
W. A. J. O'Meara, A. P. Trotter, and J. J. Thomson.

For Ovatemxda. — Francisco de Arce.

For Hungary. — ^Desir^ Harsanyi and Joisef Vdter.

For Italy. — Antonio R6iti.

For Japan. — Osuke Asano and Shigeru Eondo.

For Mexico. — Alfonso CastelW.

For Netherlands. — Dr H. Haga.

For Paraguay. — Max. F. Croskey.

For Russia. — N. Egoroff and L. Swentorzetzky.

For Spain. — Jose Ma. de Madariaga and A. Montenegro.

For Switzerland. — ^Dr H. F. Weber, P. Chappuis, €uid Jean
Landry.

For Australia.— C. W. Parley and — Threl&lL

For Canada. — Ormond Higman.

FOR ELECTRICAL MEASUREMENTS 751

For Grovm Colonies, — P. Cardew.
For India. — M. G. Simpson.

In the presence of—M. J. Collins, W. Duddell, C. W. S. Crawley,
and F. E. Smith, Secretaries.

SCHEDULE A.

List of Countries and Delegates,

America (United States). — Dr S. W. Stratton, Director Bureau
of Standards, Washington; Dr Henry S. Carhart, Professor of
Physics at the University of Michigan; and Dr E. B. Rosa,
Physicist, Bureau of Standards, Washington.

Austria. — Dr Viktor Edler von Lang, President of the Com-
mission of Weights and Measures, Vienna; and Dr Ludwig
Eusminsky, Inspector of above Commission.

Belgium. — Professor Eric Gerard, Director of the Montefiore
Electro-Technical Institution and President of the Consultative
Commission on Electricity; and Monsieur Clement, Secretary of
the Consultative Commission on Electricity.

Brazil. — Mr L. Weiss, Chief de la Section Technique des
T616graphes, Br^sil.

Chile. — Don Victor Eastman, First Secretary to the Legation
of Chile, London.

Colombia. — Don Jorge Roa.

Denmark and Sweden. — Professor S. A. Arrhenius, Nobel
Institute, Stockholm.

Ecuador. — Seflor Don Celso Nevares, Consul-General.

France. — Professor Lippmann, Member of the Institute and
Professor at the Sorbonne ; M. R. Benoit, Directeur du Bureau
International des Poids et Mesures ; and M. de Nerville, Ing^nieur
en Chef des T^l^graphes.

Germany. — Professor Warburg, President of the Imperial
Physico-Technical Institute; Professor Jaeger, Member of the
Imperial Physico-Technical Institute; and Professor Lindeck,
Member of the Imperial Physico-Technical Institute.

Great Britain. — The Right Hon. Lord Rayleigh, President of
the Royal Society; Professor J. J. Thomson, F.R.S., Cambridge;
Sir John Gavey, C.B.; Dr R. T. Glazebrook, F.R.S., Director of the

752 PRACTICAL STANDARDS

National Physical Laboratory; Major W. A. J. O'Meara^ C.M.G.,
Engineer-in-Chief General Post Office; and Mr A. P. Trotter,
Electrical Adviser to the Board of Trade.

Guatemala. — Dr Francisco de Arce, Diplomatic Representative,
London and Paris.

Hungary/. — Joisef VAter, Director Technique des Postes et des
T^l^graphes, Budapest ; and Dr Desir^ Harsanyi, Director of the
Hungarian Royal Commission for Weights and Measures.

Italy. — Professor Antonio R6iti, of Florence.

Japan, — Dr Osuke Asano, Doctor of Engineering, Official
Expert of the Department of Communications, Tokyo; and
Mr Shigeru Kondo, Official Expert of the Department of Com-
munications, Tokyo.

Mexico. — Don Alfonso CastelW and Don Jos6 Maria Perez.

Netherlands. — Dr H. Haga, Professor at the University of
Groningen.

Paraguay. — M. Maximo Croskey.

Russia. — Dr N. EgoroflF, D.Sc, Director of the General
Chamber of Weights and Measures; and Col. L. Swentorzetzky,
Ing^nieur Militaire, Prof, de TAcad^mie Militaire Nicolas des
Ing^nieurs, St Petersburg.

Spain. — Don Jos^ Maria Madariaga, Professor of Electricity
and Physics at the School of Mines, Madrid; and Don A.
Montenegro, Ing6nieur Professor du Laboratoire de TEcole de
Mines, Madrid.

Switzerland. — Dr Fr. Weber, Professor at the Swiss Poly-
technic School at Zurich ; Dr Pierre Chappuis, Membre Honoraire
du Bureau International des Poids et Mesures ; and Dr J. Landry,
Professor of Industrial Electricity in the University, Lausanne.

British Co Jomes.— Australia: Mr Cecil W. Darley, I.S.O., late
Inspecting and Consulting Engineer New South Wales Govern-
ment ; and Professor Threlfall, M.A., F.R.S. Canada : Mr Ormond
Higman, Chief Electrical Engineer Electric Standards Laboratory,
Ottawa. Cro^Ti Colonies: Major P. Cardew, Electrical Adviser.
India: Mr M. G. Simpson, Electrician of the Indian Telegraph
Department.

Secretaries: Mr M. J. Collins, Mr W. Duddell, F.RS.,
Mr C. W. S. Crawley, and Mr F. E. Smith.

FOR BLECraiCAL MEASUREMENTS 758

SCHEDULE B.

Resolutions.

I. The Conference agrees that, as heretofore, the magnitudes
of the fundamental electric units shall be determined on the
electro-magnetic system of measurement with reference to the
centimetre as the unit of length, the gramme as the unit of mass,
and the second as the unit of time.

These fundamental units are (1) the ohm, the unit of electric
resistance which has the value of 1,000,000,000 in terms of the
centimetre and second; (2) the ampfere, the unit of electric
current which has the value of one-tenth (01) in terms of the
centimetre, gramme, and the second; (3) the volt, the unit of
electromotive force which has the value 100,000,000 in terms of
the centimetre, the gramme, and the second; (4) the watt, the
unit of power which has the value 10,000,000 in terms of the
centimetre, the gramme, and the second.

II. As a system of units representing the above, and suffi-
ciently near to them to be adopted for the purpose of electrical
measurements and as a basis for legislation, the Conference re-
commends the adoption of the international ohm, the international
ampere, and the international volt defined according to the
following definitions: —

III. The ohm is the first primary unit.

IV. The international ohm is defined as the resistance of a
specified column of mercury.

y. The international ohm is the resistance offered to an un-
varying electric current by a column of mercury at the temperature
of melting ice, 14*4521 grammes in mass, of a constant cross-
sectional area and of a length of 106*300 centimetres.

To determine the resistance of a column of mercury in terms
of the international ohm, the procedure to be followed shall be
that set out in Specification I. attached to these Resolutions.

VI. The ampfere is the second primary unit.

VII. The international ampere is the unvarying electric
current which, when passed through a solution of nitrate of silver
in water, in accordance with Specification II. attached to these
Resolutions, deposits silver at the rate of 000111800 of a gramme
per second.

B, A. 48

^S4 PRACTICAL STANDARDS

YIII. The international volt is the electrical pressure, which,
when steadily applied to a conductor whose resistance is one
international ohm, will produce a current of one international
ampere.

IX. The international watt is the energy expended per second
by an unvarying electric current of one international ampere under
an electric pressure of one international volt.

Specification L

Specification relating to Mercury Standards of Resistance,

The glass tubes used for mercury standards of resistance must
be made of a glass such that the dimensions may remain as
constant as possible. The tubes must be well annealed and
straight. The bore must be as nearly as possible uniform and
circular, and the area of cross-section of the bore must be approxi-
mately one square millimetre. The mercury must have a resistance
of approximately one ohm.

Each of the tubes must be accurately calibrated. The cor-
rection to be applied to allow for the area of the cross-section of
the bore not being exactly the same at all parts of the tube must
not exceed 5 parts in 10,000.

The mercury filling the tube must be considered as bounded
by plane surfaces placed in contact with the ends of the tube.

The length of the axis of the tube, the mass of mercury the
tube contains, and the electrical resistance of the mercury are to
be determined at a temperature as near to 0° C. as possible. The
measurements are to be corrected to 0° C.

For the purpose of the electrical measurements, end vessels
carrying connections for the current and potential terminals are
to be fitted on to the tube. These end vessels are to be spherical
in shape (of a diameter of approximately four centimetres) and
should have cylindrical pieces attached to make connections with
the tubes. The outside edge of each end of the tube is to be
coincident with the inner surface of the corresponding spherical
end vessel. The leads which make contact with the mercury are
to be of thin platinum wire fused into glass. The point of entry
of the current lead and the end of the tube are to be at opposite

FOR ELECTBIOAL MKASUBEMENTS 755'

ends of a diameter of the bulb ; the potential lead is to be mid-
way between these two points. All the leads must be so thin
that no error in the resistance is introduced through conduction
of heat to the mercury. The filling of the tube with mercury for
the purpose of the resistance measurements must be carried out
under the same conditions as the filling for the determination of
the mass.

The resistance which has to be added to the resistance of the
tube to allow for the effect of the end vessels is to be calculated
by the formula —

A 0-80 /I

10637r

where r^ and r, are the radii in millimetres of the end sections of
the bore of the tube.

The mean of the calculated resistances of at least five tubes
shall be taken to determine the value of the unit of resistance.

For the purpose of the comparison of resistances with a
mercury tube the measurements shall be made with at least three
separate fillings of the tube.

Specification II.
Specification relating to the Deposition of Silver.

The electrolyte shall consist of a solution of firom 15 to 20
parts by weight of silver nitrate in 100 parts of distilled water.
The solution must only be used once, and only for so long that not
more than 30 per cent, of the silver in the solution is deposited.

The anode shall be of silver, and the kathode of platinum.
Th^ current density at the anode shall not exceed 1/5 ampere per
square centimetre and at the kathode 1/50 ampere per square
centimetre.

Not less than 100 cubic centimetres of electrolyte shall be
used in a voltameter.

Care must be taken that no particles which may become
mechanically detached from the anode shall reach the kathode.

Before weighing, any traces of solution adhering to the kathode
must be removed, and the kathode dried.

48—2

756 PRACTICAL STAHDABBS

SCHEDULE a

Weston Normal Cell.

The Weston normal cell may be conveniently employed as a
standard of electric pressure for the measurement both of e.m.f.
and of current, and, when set up in accordance with the following
specification, may be taken, provisionally*, as having, at a tem-
perature of 20° C., an E.M.F. of 1*0184 volt.

The Weston normal cell is a voltaic cell which has a saturated
aqueous solution of cadmium sulphate (CdSO« . 8/3 H3O) as its
electrolyte.

The electrolyte must be neutral to congo red-

The positive electrode of the cell is mercury.

The negative electrode of the cell is cadmium amalgam con-
sisting of 125 parts by weight of cadmium in 100 parts of
amalgam.

The depolariser, which is placed in contact with the positive
electrode, is a paste made by mixing mercurous sulphate with
powdered crystals of cadmium sulphate and a saturated aqueous
solution of cadmium sulphate.

The different methods of preparing the mercurous sulphate
paste are described in the notes f. One of the methods there
specified must be carried out.

For setting up the cell, the H form is the most suitable. The
leads passing through the glass to the electrodes must be of
platinum wire, which must not be allowed to come into contact
with the electrolyte. The amalgam is placed in one limb, the
mercury in the other.

The depolariser is placed above the mercury and a layer of
cadmium sulphate crystals is introduced into each limb. The

* See duties of the Scientific Committee, Schedule D.

t Notes on methods pursued at various standardising laboratories wiU be issued
by the Scientific Committee or the Permanent Commission, as an Appendix to ihia
Beport.

FOR ELIOTRIGAL IfXASDREHBNTS 757

entire cell is filled with a saturated solution of cadmium sulphate
and then hermetically sealed.

The following formula is recommended for the S.M.F. of the cell
in terms of the temperature between the limits O"" C. and 40° C. : —

Et^E^- 0-0000406 {t - 20*^) - 000000095 (f - 207

+ 000000001 {t - 20°)».

SCHEDULE D.

1. The Conference recommends that the various Qovemments
nterested establish a permanent International Commission for

Electrical Standards.

2. Pending the appointDient of the Permanent International
Commission, the Conference recommends* that the President,
Lord Bayleigh, nominate for appointment by the Conference a
Scientific Committee of fifteen to advise as to the organisation of
the Permanent Commission, to formulate a plan for and to direct
such work as may be necessary in connection with the maintenance
of standards, fixing of valuesf , inter-comparison of standards, and
to complete the work of the ConferenceJ. Vacancies on the
Committee to be filled by co-optation.

3. That laboratories equipped with facilities for precise elec-
trical measurements and investigations should be asked to co-
operate with this Committee and to carry out, if possible, such
work as it may desire.

* In accordance with the above, Lord Bayleigh has nominated the following
Committee, which has been approved bj the Conference, viz.: —

Dr Osuke Asano. Dr H. Haga. Dr £. B. Bosa.

M. B. Benoit. D. L. Kusminsky. Dr S. W. Stratton.

Dr N. Egoroff. Prof. St. Lindeck. Mr A. P. Trotter.

Prof. Eric Gerard. Prof. G. Lippmann. Prof. £. Warburg.

Dr B. T. Glazebrook. Prof. A. Bditi. Prof. Fr. Weber.

t This will include the reconsideration from time to time of the B.M.r. of the
Weston normal cell.

X With this object the Committee are authorised to issue as an Appendix to the
Beport of the Conference, Notes detailing the methods which have been adopted in
the Standardising Laboratories of the various countries to reaUse the International
Ohm and the International Amp^, and to set up the Weston Normal Cell.

758 PRACTICAL STANDARDS FOR ELECTRICAL SfEASXTREMENTS

4, The C!ommittee should take the proper steps forthwith for
establishing the Permanent Commission, and are empowered to
arrange for the meeting of the next Conference on Electrical
Units and Standards, and the time and place of such meeting
should this action appear to them to be desirable.

5. The Committee or the Permanent International Com-
mission shall consider the question of enlarging the functions of
the International Commission on Weights and Measures, with a
view to determining if it is possible or desirable to combine
future Conferences on Electrical Units and Standards with the
International Commission on Weights and Measures, in place of
holding in the future Conferences on Electrical Units and
Standards. At the same time it is the opinion of the Conference
that the Permanent Commission should be retained as a distinct
body, which should meet at different places in succession.

THIRTY-SEVENTH REPORT— SHEFFIELD, 1910.

Appendix. — Order in Council relating to Electrical Standards^

dated January 10, 1910.

The Report of the International Conference on Electrical Units
and Standards held in London in October 1908 was printed as an^
Appendix to last year's Report.

In January 1910 the Board of Trade took action in accordance
with the recommendations of the Report, and an Order in Council
relating to electrical units, dated January 10, which contains de-
finitions of the EiUglish standards of resistance, current, and electro-
motive force in conformity with the definitions adopted by the
Congress, has been issued. This is printed as an Appendix.

In accordance with a scheme approved by the International
Scientific Committee appointed by Lord Rayleigh at the London
Conference, international co-operative work on electric standards
has this year been carried out at the Bureau of Standards,
Washington. It was arranged that representatives of the Bureau
of Standards, the Laboratoire Central d'^lectricit^, Paris, the
Physikalisch-Technische Reichsanstalt, Berlin, and of the National
Physical Laboratory should take part in the work. The represen-
tatives of the Bureau of Standards were Professor K B. Rosa and
Dr F. A. Wolff, and the European delegates were Professor W.
Jaeger, Professor F. Laporte, and Mr F. E. Smith.

Professor S. W. Stratton kindly offered the facilities of the
Bureau of Standards for the investigation, and, in his capacity as
Treasurer of the International Committee, was able to secure the
funds to defray expenses. Towards this object the governing bodies
of the American Institute of Electrical Engineers, the National
Electric Light Association, the Association of Edison Illuminating
Companies, and the Illuminating Engineering Society, most gene-
rously subscribed £100 each. Some smaller contributions were
also received.

760 PRACTICAL STANDARDS

The primary object of the meeting was to determine the electro-
motive force of the Weston normal cell in terms of the international
units of resistance and current. At the same time it was necessary
to clear up certain outstanding problems on the standard cell and
the silver voltameter. Previous to the meeting a great deal of
experimental work had been done at each of the four institutions,
and the results obtained were compared before deciding on a
programme of experimental work.

The European delegates took with them from their own
laboratories a considerable quantity of apparatus and chemicals,
together with standards of electromotive force, resistance, and
mass. The results of the meeting are very valuable, and a full
report is in process of preparation.

Another careful research on the silver voltameter has been
made during the year by Professor F. Laporte at the Laboratoire
Central d'Electricit^. Professor Laporte shows that the result
obtained in 1908 by Professors Janet, de la Oorce, and himself,
is subject to an appreciable error, owing to the use of silver nitrate,
now known to be impure. With carefully prepared nitrate, and
using the Rayleigh form of voltameter, he obtains 1*11829 milligram
per coulomb for the electro-chemical equivalent of silver, the current
being measured in terms of the Weston cell as 1"01830 volts at
17"" C. and the international ohm as realised at the National
Physical Laboratory. The unit of current was, therefore, the same
as that used by Smith, Mather, and Lowry in 1908, and the value
for the electro-chemical equivalent found by Professor Laporte
is in very close agreement with the value 1'11827 obtained by
the British investigators.

The General Committee at Winnipeg accepted the recommen-
dation of the Council and the Committee of Section A in favour
of the republication of all the Reports of the Electrical Standards
Committee. Suitable arrangements for the work have, therefore,
been made, and the material is now with the printer, but in con-
sequence of the absence of Mr F. K Smith in America, and the
work of preparation required for this, progress has necessarily been
slow.

With regard to progress in electrical standardising work at the
National Physical Laboratory, the Lorenz apparatus is practically
complete, and some preliminaiy electrical measurements will, it is
hoped, be made in October of the present year.

FOR XLBCTRIGAL MEASUREMENTS 761

The Ayrton-Jones current balance continues to work most
satisfactorily ; small and gradual changes in E.M.F. of Weston cells,
amounting to less than three parts in 100,000 have been detected
by its aid.

The results of the investigation on cadmium amalgams at the
National Physical Laboratory were incorporated in a paper read
before the Phjrsical Society last February. It may be useful to
give here the limits of temperature between which various amalgams
may be most usefully employed in the Weston normal cell :

Peroenta(;e of
eadmiom in the

amal^m Lower limit Upper limit

6 Below 0" C. About 27-7" C.

7 „ „ 34-6

8 „ „ 41-0

9 „ „ 46-0

10 „ „ 51-0

11 About O'C. „ 560

12 „ sTC. „ eo-o

12J „ 121 Above 60O

13 „ 161 „ 60-0

14 „ 240 „ 60-0
16 „ 32-5 „ 60-0

The degree of reproducibility which is now obtainable with the
Weston cell far surpasses what it was five years ago. At the
National Physical Laboratory sixty-seven cells were tested in
1909, and of these sixty agreed with the Laboratory standards
within one part in ten thousand. What is not understood at
present is the occurrence of strange hysteresis effects in a few
cells. The e.m.f. of such cells may be normal at first, but changes
comparatively rapidly with time. Indeed, a large hysteresis effect
in a cell appears to be an indication that the E.M.F. will not remain
constant with time, whereas its absence is in general an indication
of constancy.

In view of the fact that the republication of the Reports is
not yet completed, the Committee recommended that they be
reappointed, that Lord Rayleigh be Chairman, and Dr R. T.
Glazebrook, Secretary.

762 PRACTICAL STANDAHDa

Appendix.

Order in Council Relating to Electrical Standards.

At the Court at Buckingham Palace, January 10, 1910.
Present, the King's Most Excellent Majesty in Council.

Whereas by the "Weights and Measures Act, 1889," it is,
among other things, enacted that the Board of Trade shall from
time to time cause such new denominations of standards for the
measurement of electricity as appear to them to be required for
use in trade to be made and duly verified.

And whereas by Order in Council dated the 23rd day of
August, 1894, Her late Majesty Queen Victoria, by virtue of the
power vested in H6r by the said Act, by and with the advice of
Her Privy Council, was pleased to approve the several denomina-
tions of standards set forth in the Schedule thereto as new
denominations of standards for electrical measurement.

And whereas in the said Schedule the limits of accuracy
attainable in the use of the said denominations of standards are
stated as follows:

For the Ohm within one hundredth part of one per cent.
For the Ampere within one tenth part of one per cent.
For the Volt within one tenth part of one per cent.

And whereas, at an International Conference on Electrical
Units and Standards held in London in the month of October,
1908, the International Electrical Units corresponding with the
said denominations of standards were defined as follows :

The International Ohm is the resistance offered to an un-
varying electric current by a column of mercury at the temperature
of melting ice 14*4521 grammes in mass of a constant cross sectional
area and of a length of 106'300 centimetres.

The International Ampfere is the unvarying electric current
which, when passed through a solution of nitrate of silver in water,
deposits silver at the rate of 0*00111800 of a gramme per second.

The International Volt is the electrical pressure which when
steadily applied to a conductor whose resistance is one International
Ohm will produce a current of one International Ampfere.

And whereas it has been made to appear to the Board of Trade

FOB ELBCTBICAL HEASUREUBNTS . 763

to be desirable that the denominations of standards for the measure-
ment of electricity should agree in value with the said International
Electrical Units within the said limits of accuracy 'attainable.

And whereas the denominations of standards made and duly
verified in 1894 and set forth in the Schedule to the said Order in
Council have been again verified.

And whereas the Board of Trade are advised that the said
denominations of standards agree in value with the said Inter-
national Electrical Units within the said limits of accuracy attain-
able, except that in the case of the Ohm the temperature should
be 16*4 C. in place of 15*4 C. as specified in the Schedule to the
said Order in Council.

And whereas it has been made to appear to the Board of Trade
that the said denominations of standards should be amended so
that the aforesaid exception may be remedied.

Now, therefore, His Majesty, by virtue of the power vested in
Him by the said Act, by and with the advice of His Privy Council,
is pleased to revoke the said Order in Council dated the 23rd day
of August, 1894, and is fiirther pleased to approve the several
denominations of standards set out in the Schedule hereto as
denominations of standards for the measurement of electricity.

Almeric Fitzroy.

* Schedule Above Referred To.

' I. Standard of Electrical Resistance,

*A standard of electrical resistance denominated one Ohm
agreeing in value within the limits of accuracy aforesaid with that
of the International Ohm and being the resistance between the
copper terminals of the instrument marked " Board of Trade Ohm
Standard Verified, 1894 and 1909," to the passage of an unvarying
electrical current when the coil of insulated wire forming part of
the aforesaid instrument and connected to the aforesaid terminals
is in all parts at a temperature of 16*4 C.

' II. Standard of Electrical Current.

'A standard of electrical current denominated one Ampere
agreeing in value within the limits of accuracy aforesaid with
that of the International Ampere and being the current which

764 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

is passmg in and through the coils of wire forming part of the
instrument marked ** Board of Trade Ampere Standard Verified,
1894 and 1909/' when on reversing the current in the fixed coils
the change in the forces acting upon the suspended coil in its
sighted position is exactly balanced by the force exerted by gravity
in Westminster upon the iridioplatinum weight marked A and
forming part of the said instrument.

' IIL Standard of Electricai Pressure,

'A standard of electrical pressure denominated one Volt agreeing
in value within the limits of accuracy aforesaid with that of the
International Volt and being one hundredth part of the pressure
which when applied between the terminals forming part of the
instrument marked "Board of Trade Volt Standard Verified, 1894
and 1909," causes that rotation of the suspended portion of the
instrument which is exactly measured by the coincidence of the
sighting wire with the image of the fiducial mark A before and
after application of the pressure and with that of the fiducial
mark B during the application of the pressure, these images being
produced by the suspended mirror and observed by means of the
eyepiece.

' In the use of the above standards the limits of accuracy attam-
able are as follows :

' For the Ohm, within one hundredth part of one per cent.
' For the Ampere, within one tenth part of one per cent.
' For the Volt, within one tenth part of one per cent.

'The coils and instruments referred to in this Schedule are
deposited at the Board of Trade Standardising Laboratory,
8, Richmond Terrace, Whitehall, London.'

THIRTY-EIGHTH REPORT— PORTSMOUTH, 1911.

The Committee have to regret the death since the last meeting
of the Association of Dr G. Johnstone Sfconey, F.RS. He had
been a member since 1861, and up to a few years since continued
his active interest in the work. In its earlier stages his skill in
definition and his admirable choice of nomenclature had proved
invaluable to the Committee. The collected Reports which are
to be issued shortly will indicate how large a share in the establish-
ment of the C.G.S. system of units is due to him.

Republication of Reports. — The republication of the Reports
is not yet completed, but this should be done within the present
year. The proofs of the Reports from 1862 to 1883 have been
finally revised and the remaining proofs will soon be ready.

Loreiiz Apparatus, — The progress made has been satisfactory.
Preliminary experiments have shown that the apparatus is un-
influenced by changes in the earth's magnetic field and that the
thermal E.M.F.S at the brushes on the two discs very nearly balance.
With the form of brush in use at present there are sudden changes
in the difference of the thermal E.M.F.S amounting to 2 x lO"' volt,
and it may be difficult entirely to eliminate these. With other
forms of brushes, e.g.y those made of gauze, the difference was often
1,000 times as great. It was this difficulty which led Lord Rayleigh
in 1883 to amalgamate the edge of the disc, and as a further im-
provement Professor Viriamu Jones and Professor Ayrton used
mercury jets instead of brushes. Since in the present apparatus
the changes are only 1 in 10,000 of the difference of potential
produced in one arrangement of the brushes and less for a second
arrangement, it is hoped that mercury contacts will not be necessary.
Further experiments will be made in order to obtain greater
perfection if such is possible.

Resistance Standards, — The construction of new mercury stan-
dards of resistance in accordance with the specification of the

766

PRACTICAL STANDARDS

London Conference is being proceeded with, and some of the
standards will be completed this year. Similar work is in progress
in France, in Germany, in Austria, and in the United State& In
the latter country four standards have had all of their constants
determined, and the resistance unit so obtained is in very close
agreement with that obtained from the old National Physical
Laboratory standards.

In the Committee's Report for 1908 it was shown that many
manganin resistance coils — some of which were purchased by the
Committee in 1895 — were very changeable in resistance, and in
consequence frequent comparison with mercury standards was
necessary. In 1908 it was shown at the Bureau of Standards,
and confirmed at the National Physical Laboratory and at the
Reichsanstalt, that these changes were largely due to the effect
of moisture on the shellac covering the wire. To eliminate this
source of trouble, many of the coils were hermetically sealed in
1909, and it is satisfactory to record that they are now much more
constant. The importance of this hermetical sealing is so great
when manganin resistances are to be sent to such places as cable
stations in the tropics that the attention of instrument manu-
facturers is drawn to the matter. Standard coils are readily sealed
and boxes of coils may be sealed in metal cases. The following
figures for standard coils of manganin show the advantage of
hermetical sealing.

Nominal value

100 ohms

1,000 ohms

10,000 ohms

No. 2460

No. 740

No. 2449

No. 2448
10,000-24

.Oct. 1903

99-9959

l,000-15s

l,000^0l2

1904

100-OOOa

•172

•244

2-4o

Open coils

1905
1906

•0048
•0092

•2l8

•248

•494
-668

357
d'8a

1907

■0132

•268

•8I4

3-74

V 1908

•0288

•30a

M3o

3-8«

Hermetically June 1909

■0369

•366

l-04«

5-5s

sealed in - 1910

0384

•357

1-076

5-55

paraffin oil [ 1911

•0399

•369

1-069

5-6,

It will be noted that the changes during the last three years
are very small.

FOR ELECTRICAL MEASUREMENTS 767

Silver Voltameter and Standard Cell, — Although the actions
which take place when a current passes through a solution of
silver nitrate in a voltameter are now well understood, the effects
of septa — such as silk, filter paper, and porous porcelain — are by
no means clear, and experiments have, therefore, been made to
decide whether any septum at all should be used in a voltameter.
Such experiments were suggested at the Washington Meeting in
1910. The results of the experiments made at the National Physical
Laboratory indicate that a septum of any kind is usually a source
of trouble, and may produce secondary reactions during the electro-
lysis which affect the weight of the silver deposit. Fortunately,
voltameters have been designed which render a septum unnecessary,
and these may be useful, not only in precise current measurements
with the silver voltameter, but for the deposition of metals other
than silver.

The reproducibility and constancy of the Weston normal cell
are still being carefully examined. The chief anomaly is the
hysteresis effect mentioned in last year's Report: for this effect
we have no explanation although one is much needed, as probably
it would enable cells to be made so as to remain even more constant
in E.M.F. than at present. It is necessary to point out that while
the effect is called a hysteresis one, the E.M.F. does not lag behind
the temperature. Briefly put, with ascending temperatures the
E.M.F. changes in close agreement with the temperature— E.M.F.
formula, but with descending temperatures the E.M.F. changes too
rapidly, corresponding to values at temperatures lower than the
temperature of the cell, by from 3° to 15"* C.

The Committee had hoped to have made this their last Report,
but in view of the fact that the republication is not complete they
ask for reappointment, with Lord Raylcigh as Chairman and
Dr R. T. Olazebrook as Secretary.

THIRTY-NINTH REPORT— DUNDEE, 1912.

It was understood at the last meeting of the Committee that
when the republication of the Reports was complete the Committee
would not ask for reappointment. The Reports from 1861 to 1911
inclusive have now passed through the press, and it is intended
that this, the 1912 and final Report of the Committee, should
conclude the reprints, which will be on sale in the autumn of the
present year.

It seems desirable, however, that the Committee should remain
in existence until all questions connected with the republication
are determined, and accordingly they ask for reappointment.

With regard to absolute measurements we have, as the direct
result of the work of members of the Committee, two pieces of
apparatus which should prove equal to any demand for precise
measurements in the absolute system for very many years.

A report of the British Association Ayrton-Jones current
balance appeared in 1908, and it was stated at that time that the
probable error associated with a determination of current in
absolute measure was about 2 parts in 100,000. Since then the
balance has been used on several occasions; it continues to give
satisfaction, and there appears to be no reason for doubt that so
far as the absolute measurement of current is concerned an
accuracy within at least 5 parts in 100,000 can still be guaranteed.
This conclusion is greatly strengthened by the results which were
communicated to the Association last year by Dr Dorsey of the
Bureau of Standards, Washington. At that institution Drs Rosa
and Dorsey have made experiments with a new current balance,
the coils of which are arranged in a manner similar to those used
by Joule and by Lord Rayleigh. They obtained results for the
electromotive force of the Weston normal cell which agree with
those obtained at the National Physical Laboratoiy within 4 parts
in 100,000. Whether this represents a real diflference in the
results given by the two balances, or is an actual difference in
the E.M.F.'s of the reference cells used has not yet been decided.

PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS 769

With regard to the absolute measurement of current else-
where, a current weigher has been built at the Laboratoire
Central d'l^lectricit^, Paris, and at the Beichsanstalt further
measurements are to be made in the near future. It will be
seen therefore that the absolute measurement of current is on
a very satisfactory basis. At the National Physical Laboratory
no e£forts will be spared to maintain the Ayrton-Jones balance
in good condition and to obtain results equal in precision to those
obtained at the present time.

Turning now to the absolute measurement of resistance. For
many years no measurements of this quantity have been carried
out, but at the present time the Lorenz apparatus at the National
Physical Laboratory and other apparatus now being constructed
at Berlin and Washington will place measurements of resistance
in a position equally satisfactory with those of current. The
Lorenz apparatus is now being employed for the measurement
of resistance, and it is believed that the probable error will not
exceed two parts in 100,000. This satisfactory state of affairs is
largely due to the design and size of the apparatus and the ease
with which the dimensions of the coils can be measured. Many
years ago Lord Rayleigh showed that it was not necessary to
measure accurately the diameter of the coils of a Joule balance ;
the ratio of the diameters was sufficient, and this ratio could at
any time be obtained by measuring the ratio of two currents. In
consequence, with a Joule balance an observer is not handicapped
in his measurements by the results of linear observations which
may have been made many years previously and which may be
incorrect owing to secular change. With the Lorenz apparatus
independence of previous linear measurements has been secured
by winding the coils with bare copper wire and leaving them in
this condition. This enables linear measurements to be made at
any time with ease and with precision.

Referring now to material standards, it is most gratifying to
record that measurements of resistance, of current, and of electro-
motive force are now made on the same basis in practically all
civilised countries This satisfactory state has been achieved
within the past four years and is a direct result of the labours
of the London Conference of 1908, in which this Committee was
so largely interested.

As is well known, the International standard of resistance is

B. A. 49

770

PRACTICAL STANDARDS

that of a specified column of mercury, and that of current depends
on measurements with the silver voltameter. The measurement
of electromotive force and of current may be conveniently made
by means of the Weston normal cell.

During the past two years comparisons of resistance coils and
of standard cells, and comparative experiments with the silver
voltameter, have been made by representatives of the National
Physical Laboratory and the standardising laboratories of America,
France, and Germany. The results obtained show better tban
any formal statement the remarkable agreement which now exists
between the electrical standards of the four countries named.

Table I. gives the results of measurements made at the Bureau
of Standards, the Reichsanstalt, and the National Physical La-
boratory, on four hermetically sealed resistance coils of manganin.
The values given are in international ohms at 25"" C.

Table I.

No. of

Besist-

ance

Goil

B.S.

March

1911

N.P.L.

April
1911

P.T.B.
June
1911

P.T.B.
Deo.
1911

N.P.L.
Dec.
1911

B.S.
Jan.
1912

B.S.
June
1912

Mazi-

mom

dif-

ferenoe

3939

3940

1-000053
1-000056
1-000099
1-000098

1-000052
1000053
l-OOOlOo
l-OOOlOo

1-000042
1-000043
1-000087
1-000083

1-000037

1-000038
1-000083
1 -000085

l-00005o
l-00005i

l-OOOlOo
l-OOOlOo

1-000063
1-000054
1-000098
1-000099

1-00004,
1W005,
l-OOOlOo
1-0001 Oi

0-OOOOUi
0-00001-
0-00001:
0-00001,

Table II. gives the results of measurements of the E.M.F. of the
Weston normal cell. The measurements were made at Washing-
ton by representatives of the Bureau of Standards, the Reichs-
anstalt, the Laboratoire Central d'Electricit^, and the National
Physical Laboratory. The current was measured by means of
silver voltameters of various types and capacities, and the electro-
lytes were from various sources. In the opinion of some of the
experimenters certain forms of the voltameters were untrust-
worthy and some of the electrolytes were known to be impure.
The agreement of the various means, while being very satisfactory,
is not therefore a true indication of the reproducibility of the
silver voltameter. To give an idea of this reproducibility the
results obtained at Washington with a non-septum form of

■

FOB JfiLECTBICAXi MEASUREMENTS

771

voltameter, designed at the National Physical Laboratory, are
given in Table III. The results of one experiment only have
been omitted and in that the current was unusually unsteady.

Table II.

Number of

Caioolated e.m.f.

Difference

■■

Date, 1910

VoUameien

of Weston Nonnal

from Mean.

in Oiroiiit

Cell at 20° G.

lxlO-»

April 14

1-01826

-6

„ 16

-2

„ 18

-4

„ 20

„ 22

-2

„ 26

+ 6

„ 28

+ 1

„ 30

May 2

„ 3

+ 5

„ 5

+ 4

„ 7

-3

» 12

-1

„ 19

-6

Mean =

= 1-01831.

Table III.
Results at Washington with N,P,L. Non-septum Voltameter.

Calculated e.m.f.

Difference

Dftte, 1910

of Weston Normal

from Mean.

Cell at 20° C.

lxlO-«

April 15

1-01831

„ 15

-1

„ 20

,, 20

-1

„ 30

+ 2

May 2

-2

,, &

-1

« 12

-4

Mean « 101829.

Table IV. gives the results of measurements on a number of
Weston normal cells. The values given are the diflFerences in
microvolts between the e.m.f.'s of the cells and the reference
standards of the various laboratories.

49—2

772 PRACTICAL STANDARDS FOR ELECTRICAL MEASUREMENTS

Table IV.
Differences in Microvolts,

Stand.

GeU No.

B.S.

N.P.L.

P.T.R.

N.P.L.

L.G.E.

N.P.L.

B.S.

Jane and

Aug.

Sept. and

Oot.

Oct.

' Not. and

Jan.

July I9I1

1911

Oct. 1911

1911

Dec. 1911

1912

262

- 6

- 70

- 80

-60

267

— -

—

268

—

- 15

-58

—

- 70

—

- 30

•—

-69

—

-115

-130

—

301

-24

- 5

- 30

^—

- 15

—^

-40

304

— >

—

309

-36

-27

- 45

- 20

—

-66

310

- 4

- 25

^—

- 10

-44

-13

-12

- 15

- 10

—

-22

- 30

—

- 15

- 7

—

- 1

-27

- 45

-30

—

-28

-31

- 40

-29

—^

-30

238

- 2

—

-10

350

-24

- 20

—

— -

-24

352

-31

—

- 45

-30

—

-30

133

142

—

1-3

—

—.

- 6

1-33

—

-16

- 5

- 8

The Committee feel that these results are sufficient to show
that the primary objects for which they were appointed have been
achieved, and that the present position of electrical standards — as
outlined in this Report — is very satisfactory.

With a view to completing the business arrangements con-
nected with the republication, the Committee recommend that
they be reappointed, that Lord Rayleigh be Chcurman and
Dr R. T. Glazebrook Secretary.

INDEX OF NAMES.

AdamB, Prof. W. G. 825, 889
Ames, Prof. J. S. 540, 542, 550
Arrhenius, Prof. STante 750, 751
Asano, Dr Osake 752, 757
Ayrton, Prof. W. E. 428, 409, 486, 488,
509, 560, 567, 576, 577, 581, 589, 592,
608, 614, 627, 656, 662, 668, 674, 697,
718, 744, 765

Babcook, H. D. 717

Baily, F. G. 582, 588, 586

Bandin 608

Beoqnerel, E. 24, 277

Bedford, T. G. 592, 600

BenoK, Prof. J. Bend 880, 422, 608,

751, 757
Blavier 194
Blomefield, Sir Thos. W. P. 469, 475,

509
BoUxmanu, Prof. Dr 547
Boflsoha 44, 45
Bottomley, J. T. 888
Bowley, A. S. 460

Boyle, Goartenay, C.B. 428, 469, 509
Bright, Sir Chas. 1, 8, 7, 12, 13, 46,

194, 280
Bryce, Rt Hon. Jas. 509
Budde, Dr Emil 486
Buff, H. L. 279

Gallendar, Prof. H. L. 416, 484, 542,

550, 592, 595, 597, 598, 599, 600, 602,

608
Campbell, A. 618
Oardew, Major P. 428, 434, 469, 509,

572, 573
Garhart, Prof. Henry S. 434, 438, 439,

454, 467, 485, 488, 489, 648, 649, 651,

656, 658, 667, 672, 750, 751
Carpenter, Dr H. C. H. 612, 654, 655
CasteUo, Dr A. 750, 752
Chaney, H. J. 425
Chappuis, Dr Pierre 542, 548, 592, 597,

625, 639, 645
Chavez, Augnstin W. 487
Chree, Dr Chas. 573, 624, 640, 641, 645
Chrystal, Prof. G. 319, 348, 348, 849,

850, 851, 353, 358, 369, 368, 621, 719,

720
Churchill, Bt. Hon. Winston S. 748
Clark, G. M. 413

Clark, Latimer 1, 8, 7, 18, 46, 116,

201, 214, 219, 284, 295
Clement, Dr P. 750, 751
Collins, M. J. 751, 752
Cooke, L B. 279
Crawley, C. W. 8. 751, 752
Crompton, B. E. 425, 481
Crossley, A. W. 608, 611
Crova, 361
Camming, 277

Darle^, C. W. 750, 752

Darwin, Horace 605

Dayy, Sir Humphry 277

Day, W. S. 561

de la Gorce 760

De la Rive 281

De la Tooanne 486

Deniman, W. H. 591

De Santy 885

Dewar, Prof. Sir Jas. 504, 507

Diesselhorst, 677

Dieterid, Prof. Dr C. 544

Divers, Prof. 648, 653

Dom, Prof. Dr 542, 544

Dorsey, Dr 768

Douglas, Col. 159

Dnddell, W. 751, 752

Edlund, Prof. 47
Egoroff, Dr N. 752, 757
Elder, H. M. 825, 434
Esselbaoh, Dr 13, 14, 16, 44, 166
Everett, Prof. Dr J. D. 221, 225, 520,
538, 617, 647

Faraday, Dr Michael, 62, 67, 118, 129,

180, 220, 283, 282
Feohner, Prof. Th. 47, 278
Ferraris, Comm. Galileo 486
Feussner, Dr 440, 454
Fitzgerald, Prof. G. F. 608
Fitzpatrick, Bev. T. C. 333, 341, 842,

848, 860, 864, 365, 367, 868, 397, 423,

4H3, 484, 502, 623, 719
Fitsroy, Sir Almeric 763
Fleming. Dr J. A. 819, 320, 340, 343,

314, 345, 347, 351, 352, 853, 858, 868,

379, 425, 604, 507, 621, 680, 719, 720,

723
Forbes, Geo. 485

774

INDEX OF NAMES

Forde, H. G. 274, 356

Foster, Prof. G. Carey 199, 293, 296,
820, 342, 355, 863, 367, 423, 485, 466,
469, 484, 509, 520, 521, 588, 548, 560,
564, 576, 730

Gannon, W. 561, 562, 563

Gamett, Prof. Dr Wm. 324, 423

Gassiot, J. P. 274

Gaagain 210

Gaass, G. F. 281, 282

G6rard, Prof. Erie 751, 757

Glazebrook, Dr B. T. 319, 821, 825,
380, 832, 883, 841, 342, 343, 360, 363,
365, 367, 878, 408, 417, 419, 422, 428,
484, 435, 438, 453, 455, 458, 460, 461,
465, 466, 469, 475, 476, 488, 484, 490,
494, 497, 500, 509, 520, 521, 543, 572,
576, 577, 581, 584, 586, 588, 598, 602,
604, 608, 609, 610, 613, 615, 621, 628,
627, 630, 636, 649, 650, 656, 660, 664,
665, 667, 674, 675, 676, 699, 700, 715,
716, 719, 720, 722, 724, 727, 728, 747,
761, 767, 772

Grant, W. 303, 804

Graves, E. 423

Gray, Prof. A. 677, 680

Gray, Dr Elisha 487

Gray, McFarlane 557

Gray, B. E. 745

Gray, Thos. 340, 437, 677

Green, Geo. 96, 234

Green, 607

Gresham, Hon. W. Q. 485

Griffiths, PruiMpal £. H. 866, 402, 411,
484, 466, 476, 539, 541, 542, 544,
561, 562, 568, 597, 598, 599, 600, 607,
615, 617

Guillaume, Dr M. G. 417, 484, 487,
488. 467, 542, 548, 549

Haga, Dr H. 752, 757

Hankel, 279

Barker, Dr J. A. 561, 592, 597, 613,
614, 620, 625, 638, 639, 645, 649, 719

Harris, Sir Wm. Snow 283, 284, 277

Harrison, Hugh Erat 424

Harrison 480, 675

Harsanyi, Dr Desir^ 750, 752

Heaviside, O. 532, 587, 677

Henry, Dr J. 47

Hioks-Beaoh, Sir Michael, Bart. 424,
467

Higman, Ormond 487, 750, 752

Himstedt, Prof. 333, 343

Hookiu, G. 164, 166, 169, 170, 190, 195,
199, 211, 213, 274, 294, 804, 305, 306,
812, 818, 319, 348, 344, 848, 349, 861,
358, 355, 856, 859, 621, 714, 717, 719,
722, 728, 725

HopkinsoD, Dr Jno. 428, 469, 509

Hopwood, Sir Francis J. S. 509

Horsford, 279

Hospitalier, Prof. Edooard 410, 486

Honsman, B. H. 683

Halett, G. A. 651, 656, 666, 667

Jacobi, Prof. 43, 45, 47. 74, 84, 279, 280

Jaeger, Prof. Dr W. 449, 677, 717, 738,
760, 751, 759

Janet, Piof. P. 760

JenkiB, Fleeming 10, 14, 16, 32, 37, 46,
50, 59, 72, 86, 140, 161, 166, 199, 201,
202, 209, 212, 240, 274, 277, 856, 409,
712

Jones, Prof. J. Viriamn 363, 465, 478,
488, 484, 489, 497, 543, 560, 567, 671,
576, 577, 581, 689, 690, 592, 693, 594,
595, 603, 608, 614, 627, 663, 744, 766

Joale, Dr J. P. 6, 44, 63, 66, 74, 111,
130, 165, 195, 196, 229, 256, 288, 641,
662, 568, 666, 566, 768, 769

Eahle, Dr 484, 438, 450, 462, 463, 464,
618, 660

Eahlenberg 609

Karmarsoh 19

Kelvin, Loid 2. 3, 6,6, 9, 10,11, 14,21, 96,
83, 34, 46, 54, 69, 64, 66, 67, 69, 70,
72, 74, 76, 77, 98, 101, 111, 112, 114,
120, 122, 128, 125, 126, 128, 133, 137,
141, 147, 168, 166, 169, 190, 196, 202,
208, 204, 219, 266, 258, 271, 273, 274,
276, 280, 281, 282, 284, 285, 296, 997,
299, 389, 840, 860, 386, 423, 40», 500,
564, 565, 566, 712, 716, 746

Eenelley, Dr 650

King, W. F. 271

Kirohboff, Pn>f. G. 12, 16, 87, 39, 45,
47, 281

Kohlrausoh, Prof. F. 118, 276, 358, 437,
563, 603

Kopp, Hermann 166, 281

Kreichgaaer, Dr 449

Kusminsky, Dr L. 751, 757

Landry, Dr Jean 750, 752

Langsdorf 280

Laporte, Prof. F. 759, 760

Le ChateUer, Prof. 542, 649

Ledoo, Dr S. 486

Lenz, H. F. E. 278

lievser 45

Lindeok, Dr Si 484, 435, 440, 461, 462,

630. 677, 717, 738, 747, 767
Lippmann, Prof. G. 751, 767
Lookyer, Sir J. Norman 578
Lodge, Sir Oliver J. 520, 682, 533, 684,

537, 538
Lorenz 419, 478
Lowry, Dr T. M. 718, 760
Lummer, Dr Otto 486

Madariaga, Prof. J. Ma. de 750, 762
MaUory, F. 661
Marie-Davy, 280, 281

INDEX OF NAMES

775

Mascart, Prof. E. 820, 380, 437, 466,
486, 589

Mather, Prof. T. 670, 672. 674, 576, 689,
662, 668, 676, 697, 699, 700, 702, 718,
760

Mattencci, Prof. C. 47

Matthey, G. 693, 602, 607, 608

Matthiessen, Dr A. 8, 10, 11, 12, 14, 16,
21, 24, 28, 38, 42, 48, 46, 47, 49, 75,
76, 79, 82, 162, 168, 164. 165, 166,
167. 170, 190, 191, 192, 198, 196, 199,
211, 284, 286, 287, 294, 806, 813, 848,
844, 349. 361, 864, 355, 366, 859, 864.
866, 897, 898, 899, 408. 406, 406. 408,
409, 410. 411, 501, 502, 508, 506, 507.
508, 620, 621, 626. 629, 714, 717, 719

MaxweU. Prof. J. Clerk 59, 73, 86, 140,
161, 166, 256, 271, 274, 283. 284, 806.
373

Mendenhall. Dr T. C. 486. 489

Miculesou 641, 563

Miller, Prof. W. H. 46

Montenegro. Prof. A. 750, 752

Moorby 563. 617

Mairbead, Dr Alex. 294, 318, 824, 841.
860, 866. 873. 874. 879, 886, 887, 389,
890, 425, 484, 466, 484, 501, 620, 618.
668

Mandella. Bt. Hon. A. J. 467

Myers 609

Natanson. Prof. 547

NerviUe. Dr T. de 760, 761

Neiuoann, Prof. 89. 47

Nichols, Prof. Dr E. L. 485, 489, 542, 658

Nicholson 208

Noble, Sir Andrew 592, 608

Ohm, G. S. 277, 278
Olszewski, Prof. Dr E. 542, 547
Ostwald, Prof. Dr W. 642, 644

Palaz, Prof. A. 487
Parker. Thos. 424
Parker 480

Pasohen. Prof. Dr F. 645
Peel, G. L. 510
Perry, Prof. J. 649, 664
Planck. Prof. Dr. W. 542, 645
Poggendorfif, Prof. J. G. 47
Potier et Pellat. Profs. 487
PouiUet. G. S. M. 47. 278
Preece, Sir Wm. H. 840, 342, 860, 428,
465, 469, 486, 609. 614

Bamsay. Sir W. 549

Bayleigh. Bt. Hon. Lord 298, 296, 818,
819. 821, 322. 324, 826, 880, 845. 354,
880. 419, 428, 484. 487, 488, 489. 452,
468. 462. 469. 491. 509, 618. 563. 564,
576. 577. 589. 592, 598. 604, 608. 615,
627. 686. 660, 661. 665, 672, 676, 677.
690, 691. 700, 714, 716. 720, 727, 728,

744, 747, 748, 757, 759, 761, 766, 767,

768 769 772
Begnault. H. Y. 166, 281, 589, 648,

565, 566, 567, 558. 659. 645
Bennie, J. 488, 484, 499
Beynolds, 663, 617
Bhodes, W. G. 670. 571
Bhodin 608, 504, 506
Boberts-Aasten. Sir Wm. 598
Bditi, Prof. A. 762. 767
Bosa, Prof. Dr Ed. B. 716, 717, 757,

769. 768
Bose-Innes, Jno. 666
Bonz. 410
Bowland, Prof. H. A. 485, 487, 489,

689, 641. 542, 648, 646. 668, 566. 657,

658, 559, 661, 568, 671, 608

Sabine 76. 170

Saholka, Dr Johann 487

SalYioni 420, 421

Saander. 819, 848, 719, 720

Schrader, A. 486

Sohnster, Prof. Dr 298, 296, 484, ^87,

488, 468, 641, 661, 662, 568, 608, 611,

677, 679
Searle. Dr G. F. C. 878, 879. 880
Shaw, W. N. 826, 881, 417, 589, 548,

656
Shimidza 648, 658
Sidgwick, Mrs 714, 727, 728
Siemens. Alex. 486
Siemens, C. W. 199, 208, 206, 209, 210
Siemens, Dr Werner 2, 7. 8. 9, 12, 18, 16,

28, 30, 32. 89. 44, 46, 47. 74, 76, 162,

164, 166, 170, 192, 198. 280, 281, 286,

419
Siemens, Lieut. Werner 280
Skinner, S. 866, 411, 428, 484, 488,

607. 608. 609, 610
Smith, F. £. 612, 620, 621, 622, 628,

624. 627, 686. 647, 648, 661, 668. 666,

666, 674, 676, 697, 699. 700, 712, 718,

714, 716, 759, 760
Smith, Willonghby 201. 274
Smyth, Prof. Piazzi 256
Solomon, M. 676, 581
Stewart, Prof. Balfoor 69. 72, 140, 166,

174, 175, 281
Stoney, Dr G. Johnstone 428, 520, 688,

766
Stratton, Prof. Dr S. W. 760, 761, 757,

769
Streoker 419. 421
Swan, 502. 608. 504. 606
Swentorzetzky, Dr L. 760, 762
Swinburne, J. 648, 661, 666, 667

Tait, Prof. B. G. 225. 273

Taylor. Herbert A. 294. 806, 817, 318,

355, 866, 368, 869, 868, 871. 406. 626,

780
Teddersen, Dr 46

776

INDEX OF NAMES

Thompeon, Prof. Dr SilTanus P. 428,

424, 486
Thomson, Prof. EUhn 485, 489
Thomson, Prof. Sir Jos. J. 381, 873,

879, 880, 680, 750, 751
Thomson, Prof. Sir Wm. — §ee Lord

Kelvin
Thory, Bend 487
Tinsley, H. 702
Tomlinson, H. 824, 428
Trotter. A. P. 622, 647, 661, 700, 718,

760, 752, 767
Trouton, Prof. 729, 780
Twyman 571

Tan der Eolk, Schroder 81

Yarley, C. F. 162, 194, 201, 203, 204,

280
V4ter, Dr J. 750, 752
Violle, Prof. T. 486
Vogt, Dr C. 17, 18, 28, 82, 86, 169
Volt, Prof. Dr Ernst 486
Von Bose 24, 28
Von Hehnholtz, Hermann L. F. 59, 64,

69, 114, 210, 484, 485, 488, 466, 467,

476, 486, 487, 488

Von Lang, Prof. Viktor 750, 751
Von Tonzehnann, G. W. 304

Waidner, C. W. 661

Walker, E. O. 483, 484, 501, 520

Warhorg, Prof. B. 761. 757

Weber, Prof. Fr. H. 752, 767

Weber, Prof. W. 1, 2, 3, 4, 5. 6, 9, 10,
18, 38, 39, 43, 44, 45, 46, 59, 64, 65,
69, 70, 74. 112, 114, 118, 125, 140,
141, 208, 276, 280, 281, 282, 283, 284

Wennman, 487

Werner, 609

Weston, Dr 444, 446

Wenilleumeier 487

Wheatstone, Prof. Chas. 2, 46, 47, 203,
278

Whipple, Bobt S. 602, 604

Wilberforce, L. B. 428, 434

Williamson, Prof. A. W. 8, 10, 14, 16,
24 44 46

Wil^n, H. 330

WUson, W. N. 691

Witkowski, Prof. 547

Wolff, Dr F. A. 667. 702, 759

Wtillner, Prof. Dr 542, 546

INDEX OF SUBJECTS.

••Abs" (prefix) 660
Absolute condenser 121, 122
electrometer 128, 203, 233, 234, 239,

249, 253, 271, 272, 273
measurement of resistance 140, 296,

489, 567, 700, 769
measurement of current, $ee Ampere

balance and current weigher
permeability or inductivity 525
resistance of meroury 419, 420, 421,
422, 436, 437, 489, 494, 574, 576,
577
standard of light 361, 362
system 5, 59, 61, 65, 288, 551, 768
«*Abstat" 650
Accumulator 361, 363
Air condenser 200, 294, 324, 340, 360,
363, 365, 373, 382, 385, 396, 422,
458, 607, 613
Air, specific heat of 556
Alloy, Wood's cadmium 180

dental 312
Amalgam, cadmium 659, 667, 672, 707,
710, 742, 745, 766, 761
zinc 450, 455, 659, 660
Ampere, definition of 425, 426, 429, 435,
464, 468, 469, 470, 487, 510, 512,
590, 698, 701, 703, 704, 713, 741.
753, 756, 762
balance (Ayrton - Jones Current
Weigher) 576, 589, 603, 608, 614,
627, 647, 661, 666. 675, 697, 700,
713, 744, 761, 768
Board of Trade 713, 714, 764
International 488, 489, 675, 697, 701,
702, 703, 712, 716, 738, 739, 741,
749, 753, 754, 757, 762, 763
**Ance'' (termination) 621
Apparatus, Lorenz 478, 482, 567, 614,
698, 699, 716. 744, 760, 765, 769
Atmospheric electricity 231, 242, 254,

255
Ayrton-Matber galvanometer 572

<*B,'* induction density 522, 523, 524,
525, 530, 537

B.A. unit fee Ohm

Bain's electrochemical telegraph instru-
ment 480

Balance, Ayrton - Jones, $ee Ampere
balance
EeWin 590

Wheatstone's 73, 144, 171, 190, 205,
258, 285, see also Bridge, Wheat-
stone's
Battery, Daniell's, $ee Cell, Daniell's
Grove's 275
Leyden 252, 273
Bennet's doubler 228

gold leaf electroscope 254, 376
Board of Trade Ampdre, iee Ampere,
Board of Trade
Ohm 510, 573, 574, 575, 591, 714,

763
Unit 620, 622, 623, 624, 634, 635 ,

636
Volt 714
Bohnenberger's electroscope 254
Bridge, Carey-Foster 682, 583, 687
Fleming 320, 323, 344, 400, 402, 434,

676
Kelvin double 678, 680, 682, 683, 685,

688, 692, 694
Wheatstone's, $ee also Balance,
Wheatstone's 50, 314, 323, 399,
440, 449, 460, 476, 497, 680, 688,
719
British Association scale of temperature

596
British Association unit, iee Ohm
Bunsen's cell 171

Cable, submarine 5. 117, 201, 202, 279
Cadmium alloy, see Alloy, cadmium

amalgam, see Amalgam, cadmium

cell, see Cell, Weston normal
Calor 547

Calorie 543, 547, 560. 616, 617, 618, 619
Candle as unit of light 361, 362
Capacity 109, 115, 212, 341

specific inductive 129, 130, 202, 247,
525

standard of 293, 294, 295, 318, 324,
331, 340, 607, 615

unit of 199, 219. 291, 292. 488
Carey Foster bridge, see Bridge*, Carey

Foster
Cell, Bunsen's, see Bunsen's cell

778

INDEX OF SUBJECTS

Cell, cadmium, iee Cell, Weston normal
Cell. Clark's:
definition of 430

depolariser of 480, 450, 514, 651
E.if.F. of 116, 423, 427, 438, 453,
469, 476, 488, 552, 563, 661, 699
preparation of 480, 513, 516
temp.-<soef. of 439, 452
type of 431, 439, 450, 454, 462, 515,
518, 651, 660
Cell, Daniell's 5, 18, 14, 44, 46, 62, 116,
122, 187, 147, 200, 202, 206, 218,
219, 231, 246, 251, 259, 262, 265,
294
Cell, Hellesen's 501
Cell, Leclanoh^ 332, 400, 498
Cell, Weston normal:
amalgam of 667, 707, 761
definition of 707
depolariser of 648, 651, 670, 708
E.M.F. of 698, 743, 771, 772
hysteresis of 761, 767
international experiments with 771
preparation of 666, 707, 742, 756
temp.-coef. of 711, 743, 757
type of 672, 743
Cell, Weston cadmium, $ee Cell, Weston

normal
Cheval-vapeur 551
Coaxial helices, mutual induction of

592, 593
Coefficient of induction 861, 362, 489,
521, 526, 527, 528, 529, 530, 534,
535, 536, 537, 570, 571, 590, 593
Coils, resistance:
of German silver, see German silver

resistances
of gold, iee Gold resistances
of manganin, tee Manganin resist-
ances
of platinum, see Platinum resistances
of platinum-silver, see Platinum-silver

resistances
of platinum-iridium, see Platinum
iridium resistances
Condenser, absolute 121, 122
air, tee Air, condenser
mica 201, 212, 214, 219, 365, 874, 877,
384, 385, 386, 388, 391, 392, 893,
394, 396, 459
paraffin 385, 387
standard 294, 365, 458, 466, 624
Conductance 521

Conducting power, specific 129, 805
Constantan alloy 444
Constant *'t;'' 202, 204, 276, 379
Copper-nickel manganese fJloy, tee

Aianganin
Copper, specific resistance of 360, 364,
865, 397, 483, 484, 502, 626
standard 74, 287, 626, 646
temp.-coef. of resistance of 365, 899,
405, 408, 504, 646

Coulomb 839, 521, 524, 525, 706, 700

International 488
Current balance, tee Amp^ balance

weigher. Joule's 256

effective 861, 362, 363

International unit of, tee Ampere
International

mean 361, 862

weigher, tee Ampdre balance

Daniell's battery, tee Cell, Daniell's
Declination, magnetic 152
Declinometer 157
Density, electric 128, 242, 247
Dental alloy, tee Alloy, Dental
Derived magnetic units 133

mechanical units 132
Dielectric 127, 129, 194, 200, 390
Differential galvanometer 205, 277, 449,
681, 682, 683, 685, 690, 694, 695, 696

inductance 536, 587

resistance measurer 285

permeability 537

permeance 538
Dimensions, table of electrical 133
Divided ring electrometer 221, 222
Doubler, Bennet's, tee Bonnet's doabler

Nioholson*s revolving 228
Dynamical equivalent of heat 256

Earth's magnetism 67, 70, 73, 77, 109,
147, 152, 154, 160, 212, 273, 297,298,
301
Earth-quadrant 526, 528
Earth-reading with electrometer 245,

246, 247, 248
Effective current, tee Current, effective

electromotive force 861, 862
Electric force, resultant 128

resistance in electrostatic units 120
Electrical dimensions, table of 183
potential, definition of 126, 127
pressure, standard of 469, 511, 764
pressure, unit of 509, 762
quantity 97, 98, 99, 100, 108, 117
Electricity, atmospheric, tee Atmo-
spheric electricity
statical 117, 122, 203, 234
Electro-chemical equivalents ISO, 131,
132, 137, 139
-chemical equivalent of silver 487,
464, 564, 589, 591, 697, 713, 760
dynamometer, tee also Ampi^ balance

78, 115, 165, 194, 204, 271, 292
dynamometer, Weber's 104, 105
-magnetic measure 112, 115, 116, 118,
119, 120, 121, 124, 125, 131, 141,281
-magnetic system 4, 5, 65, 70, 118,
188, 184, 282, 425, 468, 650, 753
Electrometer, absolute 123, 203, 233,
234, 239, 249, 258, 271, 272, 278
attracted disc 221
classification 221

INDEX OF SUBJECTS

779

Electrometer, divided ring, iee Divided
ring electrometer
gauge a08, 228
heterostatic 238
idiostatio 238, 252
long range 252, 253
portable 239, 246, 248, 249, 251, 252,

253, 273
quadrant 239, 251, 252, 255, 272, 427»

469
reflecting 123
repulsion '221
standard 249, 252, 272
stationary 273
symmetrical 221
Electromotive Force (e. m. f.) :
chemical affinity of 131, 132, 166, 294
dimensions of 133, 134
efifective 361, 362
electrostatic measure of 119
meaning of 100, 101
measurement and comparison of 76,
110, 111, 114, 119, 122, 123, 235
standards of, see also Cell, Clark,
and Cell, Weston normal 115,
116, 293, 294, 296, 318, 331, 434,
461, 488, 563, 749, 754, 756, 759
unit of 3, 4, 14, 46, 62, 64, 65, 69,
109, 114, 121, 135, 139, 219, 282,
291, 296, 435, 476, 488, 753
Electiophorus 203, 244, 376
Electroscope, Sennet's gold leaf, $ee
Sennet's gold leaf electroscope
Sohnenberger's, »u Sohnenberger's
electroscope
Electrostatic force 127, 219
measure of capacity of a conductor 121
measure of electromotive force 119
measure of resistance 120
system 4, 64, 65, 118, 133, 134, 650
system of units 117, 134
unite 64, 65, 78, 118, 119, 120, 123,
200, 246, 271, 274
Electrotonio state 114, 524
Energy, unit of 543, 545, 551, 615
Equipotential surfaces 95, 96, 97, 106,

106. 127
Equivalent, dynamical, of heat, nee
Dynamical equivalent of heat
electro-chemical, iee Electro-chemical
equivalent
**Etalion"466
Etalon, Jacobi*s 45
Farad 219, 339

International, 488
Fleming bridge, see Bridge, Fleming
Flux, magnetic 524, 604, 650
Footpound 60, 61

Force, magnetic 103, 267, 522, 585,
604
lines of 96, 97, 113, 594
magneto-motive 521, 522, 527, 529
unit of 60, 64, 90, 91, 282, 283

Frequency of alternations 361, 362
Fundamental standards 89, 295
Fundamental units 91, 132, 283, 296,
465, 521, 675, 700, 712, 753

"p" 664, 698

Qalvanometer, Ayrton - Mather, see
Ayrton-Mather Galvanometer
ballistic 388, 524

dififerential, see Differential galvano-
meter
sine 103, 109
standard, for absolute measurements

78, 165, 194
tangent 4, 65, 66, 103, 109, 110, 206,
298, 300, 301
Qas thermometer 595, 696, 597, 599,

625
Gauge, electrometer, see Electrometer

gauge
Gauss 521, 522, 523, 525, 526, 527, 529,
531, 532, 533, 534, 535, 536, 538,
603
Gaussage 522, 523, 526, 527, 528, 529,

533, 534, 538
German-silver alloy:
changes in molecular condition of 28,

163, 169, 441, 442
resistance coils of 10, 18, 22, 37, 41,
73, 74, 142, 145, 156, 158, 175, 190,
196, 213, 303, 359, 419, 436
temperature coefficient of res. of 305,

317, 442
standards of resistance 435, 633, 634,
738, 746, 770
Gold:
conducting power of 24, 25, 30, 31,

187
constancy of res. of 75, 191
resistance of metre gramme of 182,
186
Gold-leaf electroscope (Sennet's), see

Sennet's gold-leaf electroscope
Gold-silver alloy:
conducting power of 17, 33, 34, 36,

170, 398
constancy of resistance of 34, 36, 75,

162, 169, 353, 359, 717. 719, 726
resistances of 8, 15, 18, 22, 32, 42,
47, 164, 181, 190, 196, 285. 344,
621, 717
resistance of metre gramme of 182,

186
temperature coef. of resistance of 17,

36, 864, 722
silver copper alloy 19
Grove's battery 275

'*H" (horizontal intensity of magnetic
force) 522, 623, 625, 526, 529, 530,
537

Heat, mechanical equivalent of 5, 130,
165, 198, 342, 466, 561, 566

780

INDEX OF SUBJECTS

Heat, of water, total 539, 543, 555, 556,
558,559
specific, of air, $ee Air, specific heat of
specific, of water 541, 543, 545, 548,
549, 552, 553, 555, 557, 559, 560,
561, 562, 608, 616
unit of 78, 137, 139, 842, 539, 541,
542, 544, 545, 546, 551, 552, 553,
596, 612, 615, 617
Helices, coaxial, 9ee Coaxial helices
Hellesen's diy-cell 501
Henry, unit of induction 466, 489, 525,
526, 528, 529, 531, 532, 533, 534,
535, 536, 537
Heterostatic electric system 238, 239,

252, 253, 254, 255
Horse power 61, 187, 138, 361, 862, 551
Hysteresis in manganin 445
in Weston cells 761, 767

**I," intensity of magnetisation 532, 533
Ice unit of heat 542, 554
Idiostatic electrometer 238, 252, 258
**Ility'' (termination) 521
Impedance 363

Inductance 361, 362, 489,521, 526, 527,
528, 529, 530, 534, 535, 536, 537,
570, 571, 590, 593
Inductive-capacity, specific 129, 180,

202, 247, 525
Inductivity 521, 525, 538
Inductors 226, 227
Insulation resistance of standard coils

332, 333, 386, 371, 458, 483
Intensity of magnetic field 69, 94, 95,

96, 287, 288, 292, 535
International amp^ 488, 489, 675, 697,
701, 702. 703, 712, 716, 738, 739,
741, 749, 753, 754, 757, 762, 763
coulomb 488
farad 488
International ohm:
definition of 487, 620, 623, 675, 740,

749, 763, 754, 762
determination of, in absolute measure

489, 494, 543

mercury resistances representing 624,

636, 700, 701, 714, 738, 739, 754, 770

International thermometric 8tandards598

unit of current, see International

Ampere
volt, definition of 488, 675, 749, 753,

754, 762, 764
watt, 489, 754
Iridio-platinum alloy, ue Platinum-iri-

dium alloy
«'Ivity'» (termination) 521

Jacobi's standard 45, 74, 279, 283
Joule (unit of work), definition of 342,

861, 362, 489, 615
relation with unit of heat 543, 545,

547, 550, 551, 552, 554, 555, 556,

557, 558, 559, 596, 615, 616, 617,
618, 619
Joule and Thomson's law 282

** Kelvin," as unit of heat 545

balance 590

double bridge 678, 680, 682, 683, 685,
688, 692, 694
Eilogausses 534
Kilowatt, definition of 361, 362

•hour 551

<*L'* (mductance) 530, 531, 532, 536
Latent heat of steam 556
Latimer Clark*s cell, see Cell, Clark's
Lead, resistances of 164, 180, 181, 186,

187, 188
Leclanchd cells 382, 400, 498
Legal ohm, definition of 825
ohm, ratio to B.A. Unit 326, 329, 885,

422
standards 330, 336, 839, 354, 864, 378,

433,456
Lines of magnetic force 67, 68, 94, 96,

97, 105, 106, 107, 113, 114, 127
Long range electrometer 252, 253
Lorenz apparatus 478, 482, 567, 614,

698, 699, 716, 744, 760, 765, 769

Magnetic declination 152

field, definition of 67, 93, 526, 537,
603

field, due to current 105, 106

field, equipotential surfaces in 95

flux 522, 524, 604, 650

induction, $ee Inductance

intensity, tee Intensity of magnetio
field

moment 66, 94, 133

pole 65, 67. 93, 133, 153, 535

potential 95, 96, 107, 126, 521, 522,
523, 526, 529, 531, 538, 650

reluctance 650

units, names of 520, 521, 650

units, derived 133
Miagnetism, earth's, eee Earth's magne-
tism
Magnetization, intensity of 532, 533
Magneto-motive force, eee Force, mag-
neto-motive
Manganin resistances:

annealing of 443, 446

B.A. coils 435, 634, 635, 687, 730, 784,
770

construction of 447

humidity effect on 737, 738, 766

Beichsanstalt standards 444, 788,
746, 770

secular changes of 449, 717, 780, 731,
732, 733, 736, 766

temperature coefficient of 444, 445,
446, 490, 613, 635
Maxwell, unit of magnetic flax 604, 650

INDEX OF SUBJECTS

781

Mean current 861, 362
Measurement, absolute, of resistance, <ee
Absolute measurement of resistance
absolute, of current, iee Ampere
balance and Current weigher
Measurer, differential, resistance 285

Siemens* resistance 205
Mechanical equivalent of heat, iee Heat,
mechanical equivalent of
units 61, 65, 89, 90, 132, 282
Mega-erg 545

Mercury, absolute resistance of 419, 420,
421, 422, 486, 437, 489, 494, 574
resistance of metre gramme of 186
Mercury Standards of Besistance :
choice for material standard 1, 6, 10,

14, 15, 29, 85, 36, 87, 88, 41
comparison with other standards 185,

687, 682, 683, 722, 724
definition and specification of 43, 165,
826, 342, 422, 426, 434, 468, 475,
487, 494, 510, 739, 740, 753, 754,
755, 762
French 438
glass tubes for 40, 48, 49, 182, 184,

607, 612, 648, 765
Beichsanstalt 444, 449, 461, 462, 628,

747
reproduction and constancy of 76,
168, 164, 170, 183, 192, 199, 286,

622, 636, 701, 714, 718, 726, 727,
728, 735

Siemens 2, 7, 8, 12, 18, 89, 44, 45,

46, 74, 76, 281, 437
temp. coef. of resistance of 36, 488
Metrical system, relation to British

system 138
** Meyer" as unit of heat 545
Bfica condenser, tee Condenser, mica
Moment, magnetic 66, 94, 133
Mutual Induction of coaxial helices

592, 593

Nicholson's revolving doubler 228
Nickel, patent, resistances of 443,444, 445
Nickelin aUoy 440, 443
Nickel manganese copper alloy, ue
Manganin

Ohm:
absolute 140, 166, 284, 292, 296, 478,

489, 567, 614, 700
Board of Trade 426, 433, 468, 510,

573, 574, 575, 714, 768
de 1898 466
determination in co.s. measure 140,

166, 284, 292, 296. 478, 489, 567,

614, 700
legal, definition of 825
International, $ee International Ohm
(B.A. unit) relation to ohm in cm. of

mercury 866, 419, 420, 421, 436,

623, 727, 728

Ohm: .
(B.A. unit) relation to absolute ohm

804, 385, 433, 436, 487, 465
(B.A. unit) relation to B.O.T. ohm

500, 622, 628
(B.A. unit) relation to legal ohm 326,

829, 385, 422
(B.A. unit) relation to Siemens* unit

487
(B.O.T.) relation to absolute ohm 574,

575
(B.O.T.) relation to International Ohm

624, 634, 763
Ohm standards :
of Board of Trade 426, 433, 468, 510,

573, 574, 714, 763
of Bureau of Standards 746, 747, 770
of Laboratoire Central d 'Electricity

770
of National Physical Laboratory 627,

677, 716, 746. 747, 770
of Beichsanstalt 440, 634, 746, 747,

770
Ohmad 284

Paraffin condensers 385, 387
Patent nickel alloy 443, 444
Period, definition of 361, 362
PermeabUity 521. 525, 528, 532, 535, 587
Platinum resistance coils:
BritUb Association 196, 354, 714,

716, 719, 726
secular changes of 75, 191, 353, 354,

714, 715, 716, 719, 726
temperature coef. of 354, 605, 722,
726
Platinum-Iridium alloy :
B.A. resistances of 18, 190, 196, 285,

348, 353, 719
secular changes of 859, 621, 717, 722,

726
temperature coef. of 171, 354, 722,
726
Platinum-silver alloy:
B.A. resistances of 19, 76, 163, 190,
285, 844, 864, 627, 715, 716, 726
secular changes of 344, 847, 353, 371,

620, 682, 715, 716, 726. 729
temperature coef. of 11, 198, 294,

805, 605

Platinum resistance thermometry 208,
366, 402, 411, 592, 595, 602, 604,
607, 613, 614, 615, 624, 625, 627,
638, 649

Platinum standard of temp. 595, 650

Pole, magnetic, ue Magnetic pole

Portable electrometer, $ee Electrometer,
portable

Potential, electric 126, 127
magnetic, eee Magnetic potential

Potentiometer 680, 681, 682, 688, 685,
692, 694

Power, unit of 840, 861, 362, 489, 758

782

INDEX OF SUBJECTS

Precise measurements of eleotrio resist-
ance 676
F^dine voltameter 608, 611
Pyrometer, Siemens' 199

"Quadrant" (unit of induction) 361,
362
electrometer see Electrometer, quad-
rant
Quantity, electrical 97, 98, 99, 100, 108,
117
electrical unit of 3, 4, 13, 62, 64, 108,
135, 219, 282, 291, 292, 488

Badiation experiments (Joule) 260, 262,

264, 268, 270
Bayleigh voltameter 437, 564, 609, 701,

708, 760
Beichsanstalt standards of resistance

440, 634, 746. 747, 770
Beluctance, magnetic 650
Resistance, absolute measurement of,

see Absolute measurement of resist-
ance
absolute, of mercury, see Absolute

resistance of mercury
coils, see German silver, Gold, Manga-
* nin. Platinum, Platinum-silver,

and Platinum-iridium resistances
electrical, precise measurements of

676
electrostatic, measurement of 120
-measurer 205, 285
specific, see Specific resistance
standards, see Ohm standards
thermometry, see Platinum resistance

thermomet^
tubes, mercury, see Mercury standards

f\f T^Bi Rtfl>rice

Beplenisher 203, 222, 223, 226, 227,

228, 251, 273
Besultant electric force 128
*'BowIand" (as unit of heat) 545

Scale of Temperature, British Association

596
Selenium resistances 199
Sensitiveness of resistance measure-
ments 677
Siemens' mercury unit, see Mercury
standards of resistance
resistance-measurer 209
Silver, Resistance of metre gramme of
186, 505
resistances, constancy of 75, 162, 164,

177, 191
specific resistance of 423, 483, 502
electro-chemical equivalent of 437,
464, 564, 589, 591, 697, 713,
760
-gold alloy, see Gold-silver alloy
-platinum alloy, see Platinum-silver
alloy

Silver Voltameter:
effect of impurities in 453, 767
effect of pressure on 437, 453, 706
effect of temperature on 707
International experiments with 771
specifications 428, 470, 488, 512, 703,

749, 755
with pyridine 607
Sine galvanometer 103, 109
Specific conducting power 129, 805
heat of air 556
heat of water 541, 545, 548, 549, 552,

555, 557, 559, 560, 561, 562, 608,

616
inductive capacity 129, ISO, 202, 247,

525
resistance, definition of 129
resistance of copper 360, 364, 365,

397, 483, 484, 502, 626
resistance of mercury 437, 489, 636,

722, 724, 727, 728
resistance of silver 423, 483, 502
Standard cell, see Cell, Clark's, and cell,

Weston normal
coils, see German-silver, gold, manga*

nin, platmum-sUver, platinum,

and platinum-iridium resistances
condenser 294, 365, 458. 466, 624
electrometer 249, 252, 272
galvanometer, see Galvanometer

standard
Jacobi's 45, 74, 279, 283
legal, see Legal ohm standard
of capacity, see Capacity, standard of
of electrical current, see Ampere

balance and Silver voltameter
of electrical pressure 469, 511, 764
of electrical resistance, see Ohm

standards
of electromotive force, see Electro-
motive force, standard of
thermal unit 539
Statical electricity 117, 122, 203, 234
Stationary electrometer, see Electro-
meter, stationary
Steam, latent heat of 556
Stroboscopic method 480
Submarme cable 5, 117, 201, 202,

279
System, absolute 5, 59, 61, 65, 288, 551,

768
electromagnetic, see Electro-magnetic

system
electrostatic, see Electrostatic system
electrostatic, of units 117, 134
metrical 138

Table of electrical dimensions 133

Tangent galvanometer, see Galvano-
meter tangent

Tellurium resistances 27

Temperature, British Association Scale
of 596

INDEX OF SUBJECTS

783

Temperature ooeflficient of resistanoe
of:
constftntan 444

copper 365, 399, 405, 406, 504, 646
Oerman-silver 805, 317, 442
gold-silver 854, 722
manganin 444, 445, 446, 490, 685
mercuiy 488
patent nickel 448, 444
platinum 854, 605, 722, 726
platinum-iridium 854, 722
platinum-silver 805, 354, 457, 490,
578, 579, 586, 588, 604, 632, 721
silver 507
Temperature coefficient of:
Clark's ceU 439, 452
Weston normal cell 711, 748, 757
Specific heat of water 552, 554, 559
Temperature, platinum standard of 595,

650
Terrestrial magnetism, 9ee Earth's

magnetism
Therm. 842, 547, 549, 619
Thermal experiments (Joule), 259, 261,
263, 265, 266, 267, 270
unit 589, 540, 544, 547, 548, 549, 550,
551, 554, 555, 616
Thermometer, gas, tee Qas thermo-
meter
platinum resistance, $€e Platinum
resistance thermometry
Thermometrio standards, international

598
Thomson's reflecting electrometer, 9ee

Electrometer, reflecting
Total heat of water 539, 548, 555, 556,

558, 559
Tubes for mercury standards, 9ee Mercury
standards of resistance

Unit, absolute, tee Absolute system
electro-magnetic, tee Electro-magnetic

system
Board of Trade, tee Board of Trade

unit
British Association, tee Ohm, B.A.

unit
fundamental, tee Fundamental unite
G^uss, ue G^uss unit
Jacobi, tee Jacobi's standard
magnetic pole, tee Magnetic pole
of magnetic potential, tee Magnetic

potential
of capacity, tee Capacity, unit of
of current, see Ampere
of current, absolute, ue Ampere

balance and Current weigher

Unit of current, international, tee
Ampere, international
of electrical pressure, tee Electrical

pressure, unit of
of electricity, tee Quantity, electrical
of electromotive force, tee Electro-
motive force, unit of
of energy 548, 545, 551, 615
of force 60, 64, 90, 91, 282, 288
of heat, tee Heat, unit of
of inductance, tu Inductance
of magnetic flux, tee Magnetic flux
of magnetomotive force 521, 522, 527,

529
of 1862 7

of permeability, tee Permeability
of power 340, 361, 862, 489, 753
of quantity, tu Quantity, electrical
of resistance, tee Ohm
of work, definition of 361, 862, 489
pole, tee Magnetic pole
quantity of electricity, tee Quantity,

electrical
Siemens' mercury, tee Mercuiy

standards of resistance, Siemens'
thermal, tee Thermal unit

*<i7" constant 202, 204, 274, 276, 879
Volt, definition of 296, 427, 469, 510,

675, 758, 754, 762, 764
Voltameter, pyridine 607
silver, tee Silver voltameter

Watt, definition of 840, 842, 861, 862,
489, 753, 754
International 489, 754
Water, specific heat of, tee Heat, specific,
of water
total heat of, tee Heat of water, total
Weber (unit of current) 296

(unit of magnetic field) 520, 521,
524, 525, 526, 527, 528, 529, 580,
531, 532, 533, 534, 585, 586, 588
Weber's eleotrodynamometer 104
Weston cadmium cell, tee Cell, Weston
normal
normal cell, tee Cell, Weston normal
Wheatstone's balance, tee Balance,
Wheatstone's
bridge, tee Bridge, Wheatstone's
Wippe 66
Worl, unit of 861, 362, 489

Zinc, amalgam 450, 455, 659, 660
influence of, on resistance alloys 17,
441, 443

Candinlige :

PRINTED BT JOHN OLAT, M.A.
AT THE UNIVBB8ITT PRESS.

UNIVERSnY OF hRCHKUN

3 9016064383601

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Reports of the Committee on Electrical Standards appointed by the British Association for the Advancement of Science"

See other formats