Bulletin of the National Research Council

Google

This is a digital copy of a book that was preserved for generations on library 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 tliis 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 in forming 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/I

ti*t7 M-Gri. If 5-

I

ilarbarli CoOege library

FROM TRS

BRIGHT LEGACY

. w be upended for baski

iC.o<. '/cfi.ici-

BULLETIN

OF THE

NATIONAL RESEARCH

COUNCIL

VOLUMES
December, 1921, to August, 1922, inclusive

• »

Published by The National Research Council

OF

The National Academy of Sciences

Washington, D. C.

1921-1922

CONTENTS

Number i6

Research laboratories in industrial establishments of the
United States, including consulting research laboratories.
Pp. 135. Originally compiled by Alfred D. Flinn. En-
larged and revised by Ruth Cobb.

Number 17

Scientific papers presented before the American Geophysical
Union at its second annual meeting. Pp. 108.

Number 18

Theories of magnetism. Report of the National Research
Council Committee on Theories of Magnetism. Pp. 261.
By A. P. Wills, S. J. Barnett, L. R. IngersoU, J. Kunz,
S. L. Quimby, E. M. Terry, S. R. Williams.

o

Vol. 3. Part 1 DECEMBER, 1921 Number 16

Bulletin

OF THE

National Research

Council

RESEARCH LABORATORIES

IN INDUSTRIAL ESTABLISHMENTS

OF THE UNITED STATES

Including Consulting Research Laboratories

Originally compiled by
Alfbed D. Flinn, Secretary, Engineering Foundation

Revised and enlarged by
Ruth Cobb, Research Information Service

Published by The National Research Council

OF

The National Academy of Sciences
Washington, D. C.

1921

Announcement Concerning Publications

«

of the

National Research Council

The Proceedings of the National Academy of Sciences

has been designated as the official organ of the National
Research Council for the publication of accounts of research,
committee and other reports, and minutes.

Subscription rate for the "Proceedings" is $5 per yean
Business address: Home Secretary, National Academy of
Sciences, Smithsonian Institution, Washington, D. C.

The Bulletin of the National Research Council

presents contributions from the National Research Council^
other than proceedings, for which hitherto no appropriate
agencies of publication have existed.

The "Bulletin" is published at irregular intervals. The
subscription price, postpaid, is $5 per volume of approxi-
mately SCO pages. Numbers of the "Bulletin" are sold
separately at prices based upon the cost of manufacture (for
list of bulletins see third cover page).

The Reprint and Circular Series of the National

Research Council

renders available for purchase, at prices dependent upon the
cost of manufacture, papers published or printed by or for
the National Research Council (for list of reprints and circu-
lars see fourth cover page).

Orders for the "Bulletin" or the "Reprints and Circulars"
of the National Research Council, accompanied by remit-
tance, should be addressed: Publication Office, National
Research Council, 1701 Massachusetts Avenue, Washington,
D. C.

.^^^KRD coil
MAY lo 1923

«\^^L^.|4

BULLETIN

OF THE

NATIONAL RESEARCH COUNCIL

Vol. 3. Part 1 DECEMBER. 1921 Nombei 16

RESEARCH LABORATORIES IN INDUSTRIAL ESTABLISH-
MENTS OF THE UNITED STATES

Including Consulting Research Laboratories

Originally compiled by
Alfred D. Flinn, Secretary, Engineering Foundation

Revised and enlarged by
Ruth Cobb, Research Information Service

CONTENTS

Introduction 1

Alphabetical list of laboratories 4

Index to subject classification of laboratories 88

Subject classification of laboratories 94

Address list of directors of research 121

INTRODUCTION

The demand for information concerning industrial research laboratories
has indicated such a widespread interest in this subject that it seemed
desirable to issue an early revision of the list contained in Bulletin of the
National Research Council, number 2. The original publication was com-
piled early in 1920 by Mr. Alfred D. Flinn, Secretary of the Engineering
Foundation, with the assistance of Miss Ruth Cobb of the Research
Information Service. It contains the names of nearly 300 laboratories in
industrial establishments in the United States which had stated in direct
correspondence that they were engaged in research. The present publica-
tion has revised the original material as of August, 1921, and has added
about 250 new names.

As in the original list, all information here given has been obtained
directly by correspondence, and statements are based upon information
supplied by the laboratories. An endeavor has been made to follow the
phraseology of the laboratories wherever possible and to print each name

2 INDUSTRIAL RESEARCH LABORATORIES

exactly in the style used by the company, with regard for spelling and
abbreviations.

No investigation has been made to ascertain the character of any
laboratory listed nor the quality of work done. In order to avoid mistakes
through misinterpretation, the laboratories were given the opportunity to
approve or correct their material after it had been transcribed. During
August and September the majority of companies availed themselves of
this opportunity.

Three methods were used to collect information about laboratories not
originally listed: (a) Forms calling for corrections and additions were
distributed widely with the first edition of the list, (b) Special requests
for information were sent to over 200 companies believed to maintain
research laboratories. These names were obtained through the generous
cooperation of The Chemical Catalog Company, Inc., following its recent
survey of chemical firms in this country, (c) A press notice of the forth-
coming revision was sent to a selected group of technical and trade
journals. This notice requested information from directors of research
who had not already supplied it.

Of the 300 laboratories originally listed all except seven responded to
the appeal for a revision of the first statement. Fourteen other names
were dropped because the firms replied that they no longer maintained
research laboratories. Eleven had made new connections and appear here
under different names.

Laboratories connected with federal, state or municipal governments,
or with educational institutions, were from the outset excluded from the
inquiry, although frequently they are engaged upon investigations in
industrial research. The concerns which are not actually supporting
laboratories in their own works have not been included, nor have the
associations maintaining fellowships in certain educational institutions.
They are to be encouraged, but this compilation is limited to the labora-
tories themselves, rather than organizations supporting research.

The following information is given for each entry: name and address
of company and address of laboratory if different from that of company ;
name of director of research and number on his staff; chief lines of
research work; special equipment, if any, or equipment of unusual
character.

In addition to the alphabetical list of laboratories, which carries all the
information, there is given a subject classification. This classification
combines the headings used in the scientific and commercial classifications
of the first edition and thus eliminates some unnecessary duplication.
These classifications were devised by members of the National Research
Council, the Engineering Foundation and others interested in research
work, and are kised in part upon the classifications used in Chemical
Abstracts and Science Abstracts. They were revised and combined for
this edition with the help of Dr. C. J. West. An alphabetical index of
subjects and cross references provides a key to the classification.

A copy of the headings used in the publication was sent to each labora-

INDUSTRIAL RESEARCH LABORATORIES 3

tory listed with the request that it check the subjects under which its
name should appear. The suggestions of the laboratories were followed
wherever possible. In this way the research activities of the companies
by subject are more fairly represented than was possible when the material
was all classified by someone unfamiliar with the detailed work of these
laboratories.

The geographical classification of the original edition has been dropped
and in its place is given an alphabetical list of names and addresses of
directors of research in the laboratories included in the bulletin.

Corrections and additional information will be welcomed.

4 INDUSTRIAL RESEARCH LABORATORIES

ALPHABETICAL LIST OF LABORATORIES
z. Abb6 Engineering Company, 50 Church St., New York, N. Y.
(Designs pulverizing and grinding machinery.) Laboratory at 230
Java St., Brooklyn, N. Y.

Research sta^* H. F. Kleinfcldt and 3 men experienced in ma-
chinery.

Research work : Part time of 3 on the solution of problems which
involve crushing, grinding, pulverizing, mixing, and sifting machinery.
a. Abbott Laboratories, The, Chicago, 111.

Research staff : A. S. Burdick, 8 chemists and 4 biologists.

Research work: Three-fourths time of 12 on new anesthetics,
hypnotics, antiseptics, and other chemical research ; animal pathology
and bacteriology; pharmacology and investigations of new medicinal
preparations.

3. Abbott, William G., Jr., Wilton, N. H. (Research engineer.)
Research staff: W. G. Abbott, Jr., i engineer, i mechanical expert

and I chemist (part time).

Research work : Three-fourths time on waste recovery, special ma-
chinery and processes for mechanical, electrical, textile and chemical
trades.

4. Acheson Graphite Company, Niagara Falls, N. Y. (Graphite
products, including dry-cell filler, paint* pigment, stove polish, pencils,
electrodes, crucibles, tubes, muffles, graphite and grease lubricants.)

Research staff: A. M. Williamson and 8 assistants.
Research work: Three-fourths time of 9 on graphite, carbon and
lubricants.

5. Acme White Lead & Color Works, Detroit, Mich.
Research staff : Clifford D. Halley, 4 chemists and 2 engineers.
Research work : Full time of 7 on paints and varnishes.

Aetna Explosives Company, Inc. See Hercules Powder Co., Em-
porium Research Laboratory (p. 39).

6. Allen-Bradley Co., 286 Greenfield Ave., Milwaukee, Wis. (Elec-
tric controlling apparatus.)

Research staff: Lynde Bradley, 3 chemists and i mechanic.

Research work : Full time of 5 on resistance materials and insula-
tion.

Allied Dye & Chemical Corporation. See General Chemical Com-
pany (p. 35).

7. Aluminum Company of America, Oliver Building, Pittsburgh, Pa.
Central Laboratory at New Kensington, Pa. Branch of the Research
Bureau at Cleveland Plant of Aluminum Manufactures,- Inc.

Research staff : Francis C. Frary and others.
Research work : Aluminum production and utilization.

8. American Agricultural Chemical Company, The. Agricultural
Service Bureau, 92 State St., Boston, Mass. (Fertilizers.) Chemical
laboratory at Carteret, N. J.

Research staff: H. J. Wheeler, 9 agronomists and chemists, super-
intendent of experiment farm, i expert photographer.

Research work: Study of requirements of soils and crops where

INDUSTRIAL RESEARCH LABORATORIES 5

fertilizers are being introduced or have not been used ; study of citrus
fruits and other special crops in Florida in connection with various
types of soil ; experiments and demonstrations with fertilizers in Illi-
nois, Iowa, Minnesota, New Hampshire, Wisconsin and other states.
9. American Beet Sugar Company, Denver, Colo. Laboratory at
Rocky Ford, Colo.

Research staff: i chief, i director, i agricultural investigator, i
economic entomologist, 2 factory chemists, and i experiment station
assistant.

Researcfi work: Full time of 5 on all agricultural phases of sugar
beet improvement, including the analysis of irrigation waters and
soils, study of rotations, cultural methods, seed breeding, and the in-
vestigation of the life histories of economic insect pests.

Equipment : Complete plant pathological and entomological equip-
ment. Greenhouse for propagation and study of various economic
phases of plant breeding, control of diseases, and observations on
insect pest development and habits.
xo. American Blower Company, 6004 Russell St., Detroit, Mich.

Research staff: J. A. Watkins and 2 or more assistants.

Research work: Full time of 3 on air propelling mechanisms, air
conditioning apparatus, dehydrating or desiccating apparatus, con-
veying of dust and waste material, heating and ventilating, forced and
induced draft for combustion of all kinds of fuels and kindred lines
where air movement forms the basis for desired results.

Equipment: All kinds of instruments for measuring the pressure
and flow of air, electric dynamometers for determining power ex-
pended, electrical measuring instruments, instruments for deter-
mination of the purity, density, humidity, temperature and pressure
of the atmosphere, etc.

XX. American Brass Company, The, Waterbury, Conn. Chemical,
metallogn'aphic and metallurgical laboratory at Waterbury; physical
and electrical testing laboratory at Ansonia.

Research staff: William H. Bassett, 3 metallurgists, 2 chemists,
I physicist and metallographer, i metallographer, 2 metallurgical
engineers, i testing engineer and necessary assistants.

Research work: One-third time of 11 on nature and effect of im-
purities in copper and its alloys ; effects of mechanical working, heat
treatment, corrosion and conditions of exposure.

Equipment: Waterbury laboratory: metallographic equipment for
study of heat treatment of non-ferrous metals and alloys ; Adam Hilger
Quartz "D" spectroscope of high sensitiveness; facilities for produc-
tion of special alloys, corrosion and other special tests. Ansonia : 200,-
ooo-pound Olsen, 100,000-pound Riehle and smaller testing machines,
covering physical testing of all materials down to very fine wire;
fatigue and friction testing apparatus ; electrical apparatus for accurate
resistance and conductivity tests.

la. American Can Company, 120 Broadway, New York, N. Y.
Laboratoiy at nth Ave. and St. Charles Rfoad, May wood, 111.

Research staff: F. F. Fitzgerald, 2 assistant chemists, i metallur-
gist, 2 food technologists, 2 analysts and 4 laboratory assistants.

Research work: One-half time of 12 on cooperative work with

6 INDUSTRIAL RESEARCH LABORATORIES

packers of food products in investigating chemical changes taking
place in food products and their influence upon the preservation of
the food, its quality and its wholesomeness. Manufacturing opera-
tions, including study of fluxes, white metal alloys, coals, oils and
other materials.

Equipment: Special apparatus for analysis of tinplate and solder
for tin content ; apparatus for investigating tin cans, sealing them, etc.

13. American Chemical and Manufacturing Corporatioiit Cranford,
N.J.

Research staff : Harry P. Taber and i assistant.
Research work : Part time of 2 on animal and vegetable oils, resins,
varnish gums and cellulose esters.

14. American Chemical Paint Company^ 1126 S. nth St., Philadel-
phia, Pa.

Research staff: J. H. Gravell, 2 chemists, i engineer and i general
utility man.

Research work: Full time of 5 on rust-proof paints for iron and
steel; scale and rust removal; high temperature paint; methods of
preparing metals for painting, enameling and japanning ; water-proof
and acid-proof barrel linings.

25. American Cyanamid Omipany, 511 Fifth Ave., New York, N. Y.
Has three plants and a laboratory at each but research and develop-
ment work are being centralized at plant nearest New York.

Research istaff: W. S. Landis, 5 skilled chemists and several as-
sistants, as a minimum. Usually includes 10 or 15 skilled men being
trained for operating positions in new processes.

Research work: Full time of staff on fertilizers, nitrogen fixation,
cyanide phosphates, potash, nitrogen compounds and derivatives.
Much of the work done in the experimental plants and laboratories is
development, rather than true research.

Equipment: Apparatus is of commercial size; frequently a com-
plete small commercial plant is leased for experimental work.
x6. American Diamalt Company, 419 Plum St., Cincinnati, Ohio.
Laboratory at Riverdale, Cincinnati, Ohio.

Research staff: Joseph M. Humble and 5 chemists.

Research work: Half time of 6 on diastatic and malt sugar prod-
ucts in general.
17. American Hominy Company^ 1857 Gent Ave., Indianapolis, Ind.

Research staff: F. C. Atkinson and 10 to 12 assistants.

Research work: Approximately half time of 12 on corn products.
x8. American Institute of Baking, 1135 Fullerton Ave., Chicap^o, 111.

Research staff : Harry E. Barnard, 2 chemists and i technician.

Research work : Full time of 4 on special problems of baking and
their investigation from the standpoint of mdustrial development;
sanitation of bakeries.

Equipment : Complete baking equipment,
xg. American Optical Company, Southbridge, Mass.

Research staff : Charles Sheard, i physicist, i physicist and phj^sical
opticist, I physiological opticist, i astronomer, i general chemist, i
physical chemist, i metallurgist and 7 assistants, including mechan-
ician.

INDUSTRIAL RESEARCH LABORATORIES 7

Research work: Full time of 14 on metallurgical research in non-
ferrous metals, especially on ability of metals and alloys to stand re-
peated workings. Spectral transmission of glasses, for example,
glasses to reflect or absorb infra-red. Optical designing in general,
especially designing of scientifically correct ophthalmic lenses; also
optical instrument designing. Abrasive material for grinding and
polishing glass. Fusing together glasses of different types.
Adhesives. Glass strength investigations. Retinal currents due to
light stimulation. Relations between radiant energy and the eye.
Problems of ocular refraction. Limits of visibility in ultraviolet. Also
publishes American Journal of Physiological Optics.

Equipment: Optical measuring apparatus for transmission in the
ultraviolet, visible and infra red ; Zeiss metallographic outfit.

20. American Radiator Conu)any, BufiFalo, N. Y. Laboratory at 1807
Elmwood Ave., Buffalo, N. Y.

Research staff: Frank B. Howell, with an average of 11 technicians
and helpers.

Research work: Approximately full time of 12 on apparatus for air
warming and cooling, involving heating boilers for burning anthracite,
bituminous, and lignite coals, coke, gas, oil, etc., for Europe as well as
America. Radiators : induction, convection, radiation. Refrigeration.

Equipment : Innumerable brick and steel chimneys of various sizes
for determining accurately grate, fuel, ash, heating surface, flue sur-
face and total draft tensions.

21. American Radio and Research Corporation, Medford, Mass.
(Wireless telegraphs and telephones.)

Research staff: V. Bush, i engineer manager and 5 assistants.

Research work : Full time of 7 on phenomena at radio frequencies,
and other matters intimately connected with radio telegraphy and
telephony. Also investigation of power factor correcting equipment.

Equipment: Apparatus for measurements and research at high
frequency, such as arcs, oscillating bulbs, generators and bridges.

22. American Rolling Mill Co., The, Middletown, Ohio.
Research staff : Wesley J. Beck, 2 consulting chemical and metal-
lurgical engineers, i electrical engineer and assistants, i metallurgical
engineer and 2 assistants and i chemical engineer and 4 assistants with
routine chemists.

Research work : Nine-tenths time of staff on corrosion of iron and
steel ; alloys, paints, magnetic properties of iron and steel.

23. American Sheet and Tin Plate Company, 210 Semple St., Pitts-
burgh, Pa.

Research staff : R. E, Zimmerman, 7 chemical engineers, 2 chemists,
I physicist and i metallurgist.

Research work : Full time of 12 on chemical engineering problems
relating to the manufacture of sheet steel, tin plate, and galvanized
sheets ; metallurgy, metallography and pyrometry as applied to these
manufacturing processes.

24. American Sugar Refining Company, The, 117 Wall St., New York,
N. Y. Service Division.

Research staff : A. V. Fuller and i assistant.

Research work : One-half time of 2 on adaptability of various sugar

8 INDUSTRIAL RESEARCH LABORATORIES

cane products to special purposes ; causes of failure in manufacture of
sugar products and their remedies, and development of new sugar food
products.

Equipment : A trade candy kitchen in conjunction with the labora-
tory.

American Telephone and Telegraph Company. See Western
Electric Company, Incorporated (p. 84).
35. American Trona Corporation, Trona, Calif. (Borax, potash, etc.)

Research staff : R. W. Mumford, 2 chemical engineers and 5 chem-
ists.

Research work : Full time of 8 on study of the equilibrium between
the chlorides, sulphates, carbonates and borates of sodium and potas-
sium, development of proper evaporation methods for evaporating
Searles Lake brine and manufacture of boric acid and borates.

American Vanadium Co. See Vanadium Corporation of America
(p. 82).

26. American Window Glass Co.» Factory No. i , Arnold, Pa.
Research staff : L. P. Forman, 4 chemists and 2 ceramists.
Research work : One-third time of 7 on new developments in glass

industry, and ceramic work.

Equipment : Pyrometric apparatus ; high and low temperature elec-
tric furnaces.

27. American Writing Paper Co., Holyoke, Mass., Department of
Technical Control.

Research staff: F. C. Clark, director; Ross Campbell, assistant
director; L. E. Roberts, in charge of research section; 4 research
chemists, 5 chemical engineers, 3 analytical chemists and i laboratory
helper.

Research work: Full time of 9 on new fibres, new paper-making
processes, improvements in present processes ; mill experimental work
to improve present processes and effect economies in operation.

Equipment : 2 12-pound Noble and Wood beaters, 4 model digesters,
special machine for testing tub size. Complete experimental paper
mill with 66-inch combination Fourdrinier and cylinder paper ma-
chine ; small model paper machine producing a sheet of paper 4 inches
wide.
aS. Amoskeag Manufacturing Company, Manchester, N. H. (Textile

mills.)

Research staff: William K. Robbins, 3 chemist^ and i laboratory
helper.

Research work: Small part time of 4 on waste recovery, dye,
bleaching, sizing and testing, problems. Semi-commercial scale ex-
periments in plant.

Equipment: Exposure boards for light and weather tests, cloth
and yarn breaking machines.
29. Anaconda Copper Mining Co., Anaconda, Mont.

Research staff : F. F. Frick, 9 assistants and 10 to 20 non-technical
assistants.

Research work : Full time of 20 to 30 on problems connected with
the industry.

INDUSTRIAL RESEARCH LABORATORIES 9

30. Andrews, A. B., State Assayer, Lewiston, Me.
Research staff: A. B. Andrews, 2 chemists and i engineer.
Research work: Two-thirds time of 3 on paper, ceramics, naval

stores and electrical conductivity.

Equipment: Grinding equipment including i-ton ball mill and
digester, beater and calendar for paper.

31. Ansbacher, A. B., & Company, 527 Fifth Ave., New York, N. Y.
Laboratory at 310 N. 7th St., Brooklyn, N. Y.

. Research staff : D. N. Barad and 2 assistant chemists.

Research work : Dry colors and inorganic pigments.
3a. Ansco Company, Binghampton, N. Y. (Photographic equipment
and supplies.)

Research staff: Alfred B. Hitchins and 5 trained men.

Research work : Full time of 6 on photographic work.

Equipment: For photographic emulsions, spectroscopic work,
spectro-photogn'aphy, photometry and photo-micrography, testing of
dyes and color filters, polariscopic and refractometric work; high
temperature oveins. Experimental laboratory fbr motion picture
work.

33. Ansul Chemical Company, Marinette, Wis. (Liquified anhydrous
sulphurous acid.)

Research staff : H. V. Higley and i chemical engineer.

Research work : Three-fourths time of 2 on relation of anhydrous
sulphur dioxide to oils, metals and other materials; development of
allied products for manufacture ; study of customer's special problems
of the use of sulphur dioxide in bleaching, deodorizing, disinfecting,
mechanical refrigerating machines and chemical manufacturing.

Equipment : Special apparatus for sulphur dioxide analysis and for
plant control work.

34. Arlington Mills, Lawrence, Mass. (Worsted textiles.)
Research staff : Hugh Christison, 3 chemists and 3 assistants.
Research work : Problems connected with the manufacture of tex-
tiles in the application of dyestuffs.

35. Armour Fertilizer Works, 209 W. Jackson Blvd., Chicago, 111.
Research staff : M. Shoeld, 3 chemists and 2 engineers.
Research work : Full time of 6 men on general research relating to

fertilizer industry. (Research work at present interrupted.)

Equipment: Special type electric furnaces; special type fuel fired
furnaces.

36. Armour Glue Works, 3ist Place and Benson St., Chicago, 111.
Laboratory serves also Armour Soap Works, Armour Ammonia
Works, Armour Curled Hair Works, and Armour Sandpaper Works.

Research staff: J. R. Powell, 6 chemists, 6 laboratory assistants and
4 helpers.

Research work : Full time of i and part time of 2 on investigation
of some of the plant processes. Work is principally analytical, for
plant control.

37. Art in Buttons, Incorporated, Rochester, N. Y.

Research staff : F. W. Ross, chemical research ; Richard Stanforth,
industrial research, and assistants.

10 INDUSTRIAL RESEARCH LABORATORIES

Research work : Full time on problems incident to vegetable ivory
button manufacturing.

Associated Factory Mutual Fire Insurance Companies. See Fac-
tory Mutual Laboratories (p. 32).

38. Atlantic Dyestuff Company, 88 Ames Building, Boston, Mass.
Research staff : A. C. Burrage, Jr., and 3 assistants.

Research work : Part time of 4 on intermediates and dyes.

39. Atlantic Refining Company, The, 3144 Passyunk Avenue, Phila-
delphia, Pa. (Petroleum products.)

Research staff : T. G. Delbridge, 5 chemical engineers, 12 chemists,
I physicist and 18 assistants. Mechanical and electrical engineering
staffs collaborate with laboratory.

Research work: Three-fourths time of 37 on manufacturing
methods of petroleum refinery, including study of manufacturing
equipment and of equipment for testing.

Equipment: Laboratory-scale petroleum refinery, together with
complete equipment for study of petroleum products; large scale
manufacturing apparatus in the plant is at disposal of laboratory staff.
Atlas BaU Company. See S. K. F. Industries, Inc. (p. 72).

40. Atlas Powder Co., Wilmington, Del. (Explosives, leather cloth,
lacquers and heavy chemicals.) Maintains three laboratories for re-
search.

Research staff: R. L. Hill, Re)molds, Pa., G. C. Given, Stamford,
Conn., F. Bonnett, Jr., Landing, N. J., and 30 chemists.

Research work: Full time of 33 on explosives of all kinds, caps,
electric detonators, leather cloth, lacquers and miscellaneous chem-
icals.

Equipment : Designed for experimental work on explosives, leather
cloth and lacquers.

41. Ault & Wiborg Company, The, Cincinnati, Ohio. (Lithographic
and letter press inks, ink varnishes, dry colors and dryers ; varnishes,
lacquers and enamels ; typewriter ribbons and carbon paper ; writing
fluids, pastes and mucilages ; dealers in all lithographic supplies.)

Research staff : Robert W. Hilton and 3 research chemists.

Research work: Full time of 4 on pigments, varnishes, ribbons,
carbon paper, lacquers and enamels.

4a. Avri Drug & Chemical Company, Inc., 421 Johnston Ave., Jersey
City. N. J.

Research staff : L. M. Avstreih and i assistant.

Research work: Pharmaceutical, technical and analytical chem-
istry.

43. Babcock & Wilcox Co., The, Bayonne, N. J. (Steam engine
boilers.)

Research staff: J. B. Romer and 7 assistants.

Research work : Full time of 2 and part time of 4 on development
of refractory materials, embrittlement of steel, aluminum coating on
steel and betterment of boiler practice.

Equipment: Furnaces and apparatus for pyrometer and thermom-
eter calibration; 150,000-pound Riehle testing machine; Upton-Lewis
torsional and alternate bending machine; Brinell machine; sclero-

INDUSTRIAL RESEARCH LABORATORIES 11

scope ; special equipment for refractories research ; special equipment
for investigation of hydrogen embrittlement in steel.

44. Babcock Testing Laboratory, 803 Ridge Road, Lackawanna, N. Y.
Research staff: S. C. Babcock, Bartlett Ramsdell, i chemical en-
gineer, I chemist and i helper.

Research work : One-half time of S on driers for paint, varnish and
printer's ink trade ; by-products utilization, soap, gums, oils and waxes.

Equipment: Small scale unit (200-pound) for production of soap,
driers, etc. Destructive distillation equipment.

45. Baker & Co., Inc., Newark, N. J. (Refiners and workers of plati-
num, gold and silver.)

Research staff: F. Zimmerman, chemical department, and 11 as-
sistants. F. E. Carter, physical department, and 4 assistants.

Research work : Large part of time of 17 on chemical research, and
on production and application of precious metal and other alloys.

Equipment: Ajax-Northrup induction furnace, Arsem furnace,
metallographic outfit, Brinell hardness machine, Erichsen testing ma-
chine, Kelvin bridge, precision potentiometer.

46. Baker, J. T., Chemical Co., Phillipsburg, N. J.
Research staff : Wm. P. Fitzgerald and 3 assistants.

Research work: Full time of i on methods of testing reagents,
methods of manufacture, etc.

47. Baldwin Locomotive Works, The, Philadelphia, Pa.
Research staff: H. V. Wille, 2 chemists and 7 assistants.
Research work : Small part time of 10 on problems connected with

the plant.

Equipment : 4 Olsen testing machines up to 600,000 pounds
capacity; Brinell machines and scleroscope.

48. Banks & Craig, 51 East 42nd St., New York, N. Y. (Consulting
Engineers and Chemists.)

Research staff : Henry W. Banks, 3rd, and assistants.

Research work : Food dehydration, food products and processes of
food manufacture ; organic colloids and engineering problems in con-
nection with water supply, sewage disposal, sanitation, etc.

49. Barber Asphalt Paving Company, The, Philadelphia, Pa.
Research staff: Charles N. Forrest and 15 assistants.

Research work: Part time of 16 on application of asphalt and
petroleum to commercial purposes.

Equipment : Miniature oil refinery and complete laboratory equip-
ment for chemical and physical testing of bituminous materials,

50. Barber-Colman Company, Rockford, 111. (Small tools, machine
tools and textile machinery.)

Research staff : 2 chemists and i engineer.

Research work : Approximately one-half time of 3 on improvements
on cutting tools, alloy steels and special steels.
Equipment: Complete metallographic equipment.

51. Barrett Company, The, 4o Rector St., New York, N. Y. (Coal
tar products.) Research Department at the New York office. J. M.
Weiss, Manager of Research Department. Research laboratories at
Edgewater, N. J. Chemical Department for the manufacture of re-
fined coal tar products at Frankford, Philadelphia. Research on

12 INDUSTRIAL RESEARCH LABORATORIES

operating processes also carried on at Frankford.' A works laboratory
at each of the 35 tar plants.

Research staif (Edgcwater laboratory) : C. R. Downs, chief
chemist, C. G. Stupp, assistant chief chemist, 40 chemists and chemical
assistants, 5 engineers and 20 other men. Special products depart-
ment, under direct control of research department, employs 20 process
men and mechanics.

Research work: Full time of 65 on problems in connection with
improvement of products or processes, and development of new uses
for normal products. General manufacturing department undertakes
many experimental engineering problems, for which research depart-
ment acts in consulting capacity.

Research laboratories occupy 18,000 square feet; adjoining is a
40- X 50-foot building for experimental plant operations. Special prod-
ucts department buildings occupy about 10,000 square feet additional
space.

5a. Bausch & Lomb Optical Co., Rochester, N. Y. (Lenses and
optical instruments.)

Scientific Bureau

Research staff: Hermann Kellner, 15 optical engineers and physi-
cists and 6 laboratory assistants.

Research work: Three-fourths time on development of optical
apparatus: ophthalmic optics, microscope optics, photographic and
projection apparatus, photometers, spectrometers, glass making prob-
lems, etc. One-fourth time on development of manufacturing
methods, testing apparatus, etc.

Equipment: Complete equipment of optical instruments.

Chemical Laboratory

Research staff : Frank P. Kolb, 3 chemists and 2 assistants.

Research work: One-half time on emery and rouge washing and
grading, grinding and polishing experiments, cements, fillers, glass
washing, glass silvering, metal plating.

53. Beaver Board Companies, The, Beaver Road, Buffalo, N. Y.
(Beayerboard and other wallboards for buildings.)

Reisearch staff: H. F. Gardner, 20 chemists, 2 engineers and 40
inspectors, testers and laboratory assistants.

Research work: One-fourth time of 63 on pulp and board mill,
wallboard, asphalt roofing and gypsum products.

54. Beaver Falls Art Tile Company, Beaver Falls, Pa.

Research staff : George E. Sladek, i ceramic chemist and 2 labora-
tory assistants.

Research work: Full time of 3 on factory control work, raw ma-
terial testing, equipment testing and research for the betterment of
the product.

Equipment: Miniature plant, microscopic equipment for petro-
graphic work.

Beaver Valley Glass Co. See Fry, H. C, Glass Company (p. 34).

55. Beckman and Linden Engineering Corporation, Balboa Buildmg,
San Francisco, Calif.

Research staff: J. W. Beckman, H. E. Linden, and a varying num-
ber of chemists, physicists and assistants.

INDUSTRIAL RESEARCH LABORATORIES 13

Research work : Full time of staff on chemical, electrochemical and
organic problems; salts occurring in natural brines; chemistry of
barium and strontium salts ; electrolytic manufacture of metallic mag-
nesium directly from its oxides ; cracking of oils by high-tension dis-
charges.

Equipment: Large .motor-generator set for direct-current elec-
trolysis and transformers for high-tension work ; lOO kw. electric fur-
nace suitable for experimental purposes.

56. Beebe Laboratories, Inc., 161 3rd St., St. Paul, Minn.
Research staff: W. E. King, i expert biological chemist, i phar-
maceutical chemist, 6 bacteriologists and a number of technical
workers and assistants.

Research work : Approximately one-third time of 9 on development
of new biological products, therapeutic agents, and modification and
direction of processes of manufacture. Studies in bacteriology, im-
munology, serology, biological and pharmaceutical chemistry.

Equipment: Specially equipped for chemical, bacteriological and
serological work, and animal experimentation.

57. Belden Manufacturing Company, 23rd St. and Western Ave.,
Chicago, 111. (Rubber insulated wires and cables, coil-winding ma-
chines, electromagnets and similar products.)

Research staff: J. V. Van Buskirk.

Research work : Problems relating to own industry.

58. Bennetts' Chemical Laboratory, 1 142 Market St., Tacoma, Wash.
(Analytical and consulting chemists, assayers and metallurgists.)

Research staff: B. H. Bennetts, 4 chemists and i metallurgist.

Research work : Part time of 6 on concentration of manganese ores
of Pacific Coast. Atomizing of copper, zinc and aluminum. Agricul-
tural chemistry.

Equipment: Metal atomizing plant for copper, zinc and aluminum.

59. Berry Brothers, Inc., Detroit, Mich. (Varnishes and paint spe-
cialties.)

Research staff : John F. Thomas and 3 chemists.
Research work: One-third time of 4 on paint vehicles, varnishes
and shellacs.

60. Bethlehem Shipbuilding Corporation, Ltd., Union Plant, San
Francisco, Calif.

Research staff : S. R. Thurston and i assistant chemist.

Research work : On improvement of strength and homogeneity of
non-ferrous alloys.

Equipment: Low voltage generator, Olsen automatic and auto-
graphic universal testing machine of 200,000 pounds capacity ; Shore
scleroscope, Brinell hardness apparatus.

61. Betz, Frank S., Company, Henry and Hoffman Sts., Hammond,
Ind. (Electric X-ray apparatus, hospital, surgical and dental sup-
plies.)

Research staff : P. M. Phillips and i assistant chemist.

Research work : One-fourth time of 2 chemists on improvement of
existing formulas of pharmaceutical products; development of new
ideas in drug and toilet preparations; research to improve plant
methods. Field covered: Syrups, elixirs, tinctures, fluid extracts,

14 INDUSTRIAL RESEARCH LABORATORIES

solid extracts, mixtures, ointments, suppositories, compressed coated

and hypodermic tablets and lozenges.

6a. Bloede, Victor G., Co., Station D, Baltimore, Md. (Chemicals.)

Research staff: Victor G. Bloede, 4 chemists and assistant.

Research work: Full time of 6 on developing adhesive products
including dextrines, vegetable glues and special hydrolyzed starch
products, and developing special sizings and gums for textile, carpet,
wall paper manufacturing purposes, etc.

Equipment: Special dextrinizing and hydrolyzinpf apparatus and
facilities for testing and developing vegetable adhesives and sizings ;
also machinery for testing out these products on a commercial scale.

63. Bond Manufacturing Corporation, Monroe and Fifth Sts., Wil-
mington, Del. (Bottle seals.)

Research stail: William G. Bond, i engineering chemist and 2
engineers.

Research work: Full time of 4 on industrial problems connected
with manufacture of composition cork, collapsible tubes and bottle
crowns.

64. Boonton Rubber Manufacturing Company, Boonton, N. J. (Elec-
trical insulation and molded products.)

Research staff : R. W. Seabury, i chemist, i electrical engineer and
I mechanical engineer.

Research work: One-third time of 4 on such problems as non-
carbonizing molded insulation for high-tension automobile ignition
apparatus; synthetic resins. Development of satisfactory insulation
for high frequency, and commercial tests for same.

Equipment : 100,000-volt testing transformer ; special apparatus for
coating paper and fabric with resins in solution. High frequency
phase displacement testing apparatus.

65. Borromite Co. of America, The, 105 W. Monroe St., Chicago, 111.
(Water softening systems.) Laboratory at 54 E. i8th St., Chicago, 111.

Research staff: John A. Montgomery, 3 chemists and 3 engineers.

Research work : Approximately one-half time of 7 on water soften-
ing and equipment. A natural zeolite is employed as the water
softening medium.

66. Borrowman, GecK'ge, 130 N. Wells St., Chicago, 111. (Chemist.)
Research staff : George Borrowman and i chemist.

Research work: One-half time of 2 on materials of engfineering,
such as waters, fuels, metals, cements, paints and clays.

Boston Biochemical Laboratory, Inc., The. See Skinner, Sher-
man & Esselin, Incorporated (p. 72).

67. Bowker Insecticide Company, 49 Chambers St., New York, N. Y.
(Insecticides and fungicides.) Laboratory at Everett, Mass.

Research staff : Firman Thompson and i chemist.
Research work : Full time of 2 on insecticides, fungicides and dis-
infectant^.

68. Boyer Chemical Laboratory Company, 940 N. Clark St., Chicago,
111. (Private label chemical specialties and manufacturing chemists
for the wholesale trade.)

Research staff : A. D. Boyer, i chemist and i laboratory assistant.

INDUSTRIAL RESEARCH LABORATORIES 15

Research work: One-third time of 3 on varnishes, oils, waxes, pol-
ishing materials, gums, disinfectants, etc.

69. Brachy E. J., and Sons, 215 W. Ohio St., Chicago, 111. (Candies.)
Has a laboratory for control and research and a manufacturing labora-
tory.

Research staff : C. O. Dicken and 3 chemists.

Research work : One-third time of 4 on improvement of analytical
methods and problems in manufacture of candy.

70. Bridgeman-Russell Company, iioo W. Superior St., Duluth, Minn.
Research staff: Benjamin F. Eichinger and 2 or more assistants.
Research work : One-half time of 2 on chemical and bacteriological

problems of manufacture of all dairy products, including sanitation,
standardization and testing new methods of manufacture.

Equipment: Highly perfected equipment for complete chemical
and bacteriological analysis of all dairy products, water and food.

71. Bridgeport Brass Company, Bridgeport, Conn.

Research staff: W. R. Webster, i metallurgist, i mechanical en-
gineer, 3 chemists, and 5 assistants.

Research work: Full time of 2 and one-half time of 3 on general
problems incidental to manufacture and fabrication of a large variety
of alloys. A large amount of research work is done with the coopera-
tion of the operating departments in the factories.

Equipment : Apparatus for testing free cutting qualities of metals.
7a. Brooklyn Union Gas Company, The, 176 Remsen St., Brooklyn,
N.Y.

Research staff: E. C. Uhlig, 2 assistant chemists, i chemical en-
gineer, 9 analysts, 3 photometric inspectors, 21 gas testers and i
photographer.

Research work: Part time of 38 on problems of manufacture and
distribution of gas.

Unusual equipment: Apparatus for experimental gasification of
oils ; photometer for spherical candle power of lamps.
73. Brown Company, Portland, Me. Formerly Berlin Mills Com-
pany. Mills: Berlin, N. H., and La Tuque, P. Q.; laboratory, Berlin,
N. H. (Paper, sulphate and sulphite fiber, chemicals and lumber.)

Research staff: Hugh K. Moore, 28 graduate chemists, i mechani-
cal engineer, 2 technical photographers, and 10 chemical assistants.

Research work : Full time of 27 on work including following sub-
jects : improvements in the various mill processes of sulphite and sul-
phate pulp making ; study of commercial electrolytic cells ; plant im-
provements in the hydrogenation of oils and laboratory study of the
process ; production of liquid chlorine, bleach powder by a continuous
process, and acetylene tetrachloride; drying and impregnation of
fiber tubes; production of alcohol by fermentation of hydrolyzed
wood waste and sulphite waste liquor; properties of SO, solutions;
study of physical and chemical properties of wood pulp, the beating
process, testing of pulp and paper; performance of paper machines;
study of lubrication problems; recovery and utilization of para-
cymene; performance of steam boiler equipment; preservation of
wood and pulp; purification of sulphate turpentine; study of color

16 INDUSTRIAL RESEARCH LABORATORIES

measurement; determination of characteristics of gas-absorption
towers ; new uses for evaporated sulphite waste liquor.

Equipment : 100,000 K. v. a. transformers and switchboard for elec-
tric furnace work ; Audiff ren-Singrun refrigerating machine ; constant
temperature and humidity apparatus for pulp and paper testing ; high
pressure gas compressor.

74. Brown & Sharpe Mfg. Co., Providence, R. I. (Machinery and
tools.)

Research work : On gray iron.

75. Brunswick-Balke-CoUender Co., The, Muskegon, Mich. (Me-
chanical and hard rubber products.)

Research staff: A. Brill and 3 men.

Research work: One-fourth time of 4 on rubber, glue and wood-
working.

76. Buchanan, C. G., Chemical Company, Station H, Cincinnati, Ohio.
(Case hardening and carbonizing compounds.) Laboratory at Baker
Ave., Norwood, Ohio.

Research staff: R. F. Catherman, i electrical engineer and i
chemist.

Research work: Variable amount time of staff on metallic salts,
pigments, industrial chemicals and their application in the various
industries.

77. Buckeye Clay Pot Co., Bassett and Ontario Sts., Toledo, Ohio.
(Fire-clay products.)

Research staff: W. K. Brownlee.

Research work: Two-thirds time of i on tests of clay including
determinations of dry transverse strength, water of plasticity, linear
drying shrinkage, screen analysis for fineness, etc., also melting point,
ability to withstand load at high temperatures, porosity, linear burn-
ing shrinkages, burned strength, and other properties of burned clay.

Equipment : For making both routine and special tests of clays.

78. Buffalo Foundry and Machine Co., 1543 Fillmore Ave., Buffalo,
N. Y. (Vacuum dryers, evaporators and industrial chemical appa-
ratus.)

Research staff: Willard Rother, metallurgical and physical test-
ing department; D. J. Van Marie, chemical department; Charles
Lavett, vacuum laboratory and testing departments ; 2 assistant chem-
ists and 5 assistant engineers and operators.

Research work : Small part time of 10 on practical experiments on
materials furnished by customers to determine in advance what can
be done by means of vacuum apparatus.

Equipment: Completely equipped metallurgical, chemical and
testing laboratories.

79. Burdett Manufacturing Company, St. Johns Court at Fulton St.,
Chicago, 111. (Oxygen and hydrogen gas generating apparatus.)

Research staff: J. B. Burdett and i chemist.

Research work: Full time of 2 on rates of diffusion of gases, ex-
plosive limits of gases, effect of electrolytic action incident to decom-
position of water on various materials used in construction (steel,
rubber and asbestos) development of special compounds for per-

INDUSTRIAL RESEARCH LABORATORIES 17

manent resistance to such action and to action of comparatively strong
alkaline solutions.

Burke Tannery. See International Shoe Co. (p. 44).

80. Butterworth-Judson Corporation, Newark, N. J. (Chemicals,
intermediates, dyes.)

Research staff : A. Riker, Jr., 6 chemists and 2 helpers.

Research work : Full time of 9 on problems relating to dye manufac-
ture.

Equipment: Particularly adapted for work on intermediates, dyes,
acids and heavy chemicals, including semi-commercial scale apparatus.

81. Byers, A. M., Company, Pittsburgh, Pa. (Wrought iron pipe,
oil well tubing and casing.)

Research staff: James Aston, several metallurgists and chemists,
and assistants.

Research work : One-half time of staff on corrosion and protective
coatings of iron ; development of wrought iron.

Equipment: Apparatus for corrosion tests and for determining
physical characteristics. Electric furnace and auxiliary equipment for
experimental heats of iron.
8a. Cabot, Samuel, Inc., 141 Milk St., Boston, Mass.

Research staff: Samuel Cabot and i assistant.

Research work: One-third time of 2 on coal tar distillates, dis-
infectants, paints, stains, varnishes.

83. Calco Chemical Company, The, Bound Brook, N. J.
Research staff : M. L. Crossley and 22 chemists.

Research work : Full time of 23 and part time of plant engineers on
intermediates, dyes and pharmaceuticals, including fundamental prob-
lems of the reactions involved, development of new processes and
plant improvement.

84. California Fruit Growers Exchange, Box 518, Corona, Calif.
Research staff : C. P. Wilson and 2 chemists.

Research work: Five-sixths time of 3 on by-products from citrus
fruits ; chemical problems connected with production, preparation and
sale ol citrus fruits.

85. California Ink Company, Inc., Camelia and 4th Sts., Berkeley,
Calif. (Printing and lithographic inks, varnishes and rollers.)

Research staff: E. T. Frickstad and 3 chemists.
Research work : One-half time of 4 on oil, varnish, dry color, dyes,
intermediates and inks.

86. Calumet & Hecla Mining Company, Calumet, Mich. Laboratory
at Lake Linden, Mich.

Research staff: C. H. Benedict with an average of 4 assistants.
Research work : Hydrometallurgy of copper.
Equipment : Large scale operation in leaching, flotation, etc.
Carbide and Carbon Chemical Corporation. See Union Carbide
and Carbon Research Laboratories, Inc. (p. 78).

87. Carborundum Company, The, Niagara Falls, N. Y. (Abrasive
and refractory materials.)

Research staff: M. L. Hartmann, 15 chemical and electrochemical
engineers, 7 assistants with technical experience and 4 non-technical
helpers.

18 INDUSTRIAL RESEARCH LABORATORIES

Research work : Full time of 27 on problems relating to abrasives
and refractories. Development of new products and improvement of
present processes. Semi-commercial scale equipment is available in
electric furnace laboratory. Especially interested in study and de-
velopment of the specialized refractory materials. Problems also in-
clude those relating to adhesives, rubber, shellac, paper and cloth.

88. Cam^e Steel Company, 1054 Frick Annex Building, Pittsburgh,
Pa. Central Research Bureau for United States Steel Corporation.

Research staff: J. S. Unger; chemists, physicists, engineers and
assistants selected from works staffs as needed.

Research work: At steel plants, covering problems of steel manu-
facture, properties of refractories and other materials used in steel
manufacture, by-products and the testing of finished products, par-
ticularly service tests.

89. Canu Chemical Company, La Salle, 111. (Permanganates, man-
ganese salts, titanium salts, saccharine, toluene sulphonamides and
their chlorine derivatives, benzoates.)

Research staff : Karl Kleimenhagen and 7 men.
Research work: Three-fourths time of 8 on development of pro-
cess for producing chemicals manufactured by the company.

90. Case Research Lab<Mratory, Auburn, N. Y.

Research suff: Theodore W. Case, 3 technical men and several
assistants.

Research work: Full time of at least 4 on problems in light and
photo-electricity.

Equipment: Apparatus for photo-electric work.
92. Caulk, L. D., Company, The, Milford, Del. (Dental materials.)

Research staff : Arthur W Gray, director physical research ; Paul
Poetschke, director department of chemistry; D. Anton Zurbrigg,
director clinical department.

Research work: Properties and application of materials used in
dentistry.

Equipment : Chemical and physical apparatus for determining the
properties of dental products of all kinds. Equipment for bacteriolog-
ical, biological and chemical investigation of dental problems.

92. Celite Products Company, Van Nuys Building, Los Angeles,
Calif. (Manufacturers and distributors of heat insulating materials,
filtering materials and mineral fillers.) Laboratory at Lompoc, Calif.

Research staff: P. A. Boeck, 2 chemical engineers, i chemist and
4 assistants.

Research work: Approximately three-fourths time of 10 on filtra-
tion of industrial liquids, measurement of thermal insulation and ca-
pacity of heat insulating materials and microscopical analysis.

Equipment: Complete equipment for pressure and gravity filtra-
tion of industrial liquids, apparatus for the determination of thermal
conductivity of insulators and furnace equipment for refractory
testing.

93. (Antral Dyestuff and Chemical Co., Plum Point Lane, Newark,
N. J. (Coal tar colors and intermediates.)

Research staff : John Prochazka, 14 chemists and assistants.

INDUSTRIAL RESEARCH LABORATORIES 19

Research work: Nine-tenths time of 15 with assistants on dye-
stnffSy pharmaceuticals, and coal tar intermediates.

Equipment: Separate experimental factory 60x40 with adequate
machinery, such as suitable stills, filter presses and autoclaves for
small scale manufacture.

94. Central Scientific Company, 460 East Ohio St:, Chicago, 111.
(Physical, chemical, agricultural and biological apparatus.)

Research staff : Paul E. Klopsteg and 2 assistants.
Research work: Full time of 2 on development of new apparatus
and instruments, and improvement in devices already manufactured.

95. Champion Ignition Company, Flint, Mich.

Research staff : T. G. McDougal and 2 ceramic engineers.

Research work : Three-fourths time of 3 on perfection of high tem-
perature insulation (electrical) ; super-refractory furnace linings for
own use ; continuous high temperature factory processes.

Equipment: Laboratory and factory facilities for operations up to
1800 C. Equipment for measuring electrical leakage up to 900 C.

96. Champion Porcelain Company, Detroit, Mich. Formerly Jeffery-
Dewitt Co. (Porcelain products.)

Research staff: Frank H. Riddle and 12 assistants.

Research work: One-third time of 13 on ceramic investigations
necessary in ignition and high tension porcelain manufacture includ-
ing development of bodies, methods of testing, manufacturing, etc.
Also development of furnaces, special refractories and similar equip-
ment.

Equipment: Electrical equipment for tests of porcelains, for igni-
tion and high tension work, special furnaces for tests of refractories.

97. Charlotte Chemical Laboratories, Inc., 606 Trust Building, Char-
lotte, N. C.

Research staff: FJ. Bartholomew, i chemist, 2 chemical engineers.

Research work: Two-thirds time of 6 on development of plant
processes.

Equipment: Electric vacuum furnaces. Large capacity grinding
units.

98. Chase Metal Works, Waterbury, Conn. (Brass, bronze, copper
and nickel, silver, rod, wire, sheet and tubing.)

Research staff: Harry George, 3 chemists, i electrochemist, 3
metallurgists and 8 assistants.

Research work : One-fifth time of 16 on improvement of properties
and methods of manufacture of copper-zinc alloys ; also investigation
of steels, lacquers, fuels and oils.

Equipment: 100,000-pound Olsen testing machine, 50,000-pound
Riehle testing machine, 10,000-1,000-pound Olsen wire testing ma-
chine, Brinell machine, Spring tester, scleroscopes ; metallographic
equipment, electric annealing muffles with electrically controlled
thermostats.

99. Chemical Economy Company, 1640 N. Spring St., Los Angeles,
Calif. (Photographers' chemicals.)

Research staff : C. W. Judd and 4 chemists.

Research work: One-tenth time of 5 on celluloid and by-products
and photographic chemicals.

20 INDUSTRIAL RESEARCH LABORATORIES

loo. Chemical Products Company, 44 K St., South Boston, Mass.
(Manufacturing chemists.)

Research staff : H. S. Mork, a chemists and i assistant.

Research work : One-fifth time of 4 on cellulose chemistry,
loz. Chemical Service Laboratories, Inc., The, W. Conshohocken,
Pa. (Analytical, consulting and engineering chemists.)

Research staff: J. Ed. Brewer and 4 assistants.

Research work: One-fourth time of 5 on coal tar, coal tar distil-
lates, fuels, gasworks ; raw materials and products.

Equipment : For plant scale experiment,
zos. Chicago Mill and Lumber Company, Conway Bldg., Chicago,
111.

Research staff: Don L. Quinn, 2 engineers in forest products, i
mechanical engineer and i chemical engineer.

Research work: Study of designs and mechanical properties of
packing boxes, crates and methods of packing; also chemical studies
on fibre board construction.

Equipment : 16 ft. revolving drum testing machine which subjects
packages to most of the hazards of transportation.
Z03. Childs, Charles M., & Co., Inc., 41 Summit St., Brooklyn, N. Y.
(Paints.)

Research staff : F. D. Heim, 2 chemists and i assistant chemist.

Research work : Full time of 4 on production of new color lakes.

Equipment : Special equipment for producing color lakes and ma-
chines for coating and polishing paper.

Z04. Cleveland Testing Laboratory Co., The, 511 Superior Building,
Cleveland, Ohio.

Research staff : C. A. Black, 2 chemists and assistants as required.

Research work: One-third time of 3 on problems in connection
with industrial plants.

Z05. Cochrane, H. S. B. W., Corporation, 17th St. and Allegheny
Ave., Philadelphia, Pa., and Earnest, Pa. Formerly Harrison Safety
Boiler Works.

Research staff: P. S. Lyon and 5 engineers; J. D. Yoder and 2
chemists.

Research work : Full time of 6 on treatment of boiler feed water ;
experiments on V-notch weirs and other flow meters; water soften-
ing; problems in the development of traps, valves, steam and oil sepa-
rators, etc.

106. Coleman & Bell Company, The, Norwood, Ohio. Successors to
National Stain and Reagent Co. (Biological stains and indicators.)

Research staff: A. B. Coleman, W. H. Bell and i assistant.

Research work : Approximately full time of 3 on syntheses of chem-
ically pure organic dyestuffs and compounds for use in biology, path-
ology, botany, and medicine in general; preparation and testing of
all kinds of indicators for use in chemistry, biology, etc. ; preparation
of chemically pure organic compounds and reagents and research upon
practical industrial problems in organic chemistry.

Equipment: Complete semi-commercial equipment for the prepa-
ration of dyestuffs and facilities for testing chemicals and dyes for use
as biological stains and indicators.

INDUSTRIAL RESEARCH LABORATORIES 21

107. Columbia Graphophone Manufacturing Company, Bridgeport,
Conn.

Research staff ; W. R. Palmer, general superintendent of engineer-
ing.

Research work: General development work in semi-plastics, ac-
coustics, electroplating, material testing and specifications, machine
developments, cabinet design and manufacturing methods.

108. Commercial Testing and Engineering Co., 1785 Old Colony
Bldg., Chicago, 111. (Coal analysis and boiler room economies.)

Research staff : Jerome F. Kohout, 3 chemists and i engineer.

Research work: Three-tenths time of 5 men on coal problems, —
particularly coking low grade coal at high and low temperature ; mix-
ing of coals to produce either high coke yield or large recovery of by-
products or both. Examination of coal with special reference to
proper time, temperature, and pressure conditions in coke oven. De-
terioration of coal in storage with reference to its coking properties.
Design of furnaces and boilers to meet special conditions of fuel or
other requirements.

109. Commonwealth Edison Company, 72 West Adams St., Chicago,
111. (Operator of large electric light and power generating and dis-
tributing systems.)

Research staff : Louis A. Ferguson and 6 trained men.

Research work: Part time of 7 on insulation deterioration, poten-
tial rises due to switching operations, heat dissipation, electric fur-
nace investigations and storage battery problems.

Equipment : Primary and secondary standardizing instruments,
especially for heavy currents ; oscillograph and high potential instru-
ments; special generators and transformers; apparatus for dielectric
and insulation tests.

zzo. Condensite Company of America, Bloomfield, N. J. (Phenolic
condensation products, chlorine substitution products, hydrochloric
acid.)

Research staflF: W. T. Hutchinson and i assistant.

Research work: Three-fourths time of 2 on improvement of prod-
ucts.

zzi. Consolidated Gas Company of New York, 130 E. 15th St., New
York, N. Y. Consolidated laboratories at Lawrence Point, Astoria,
N. Y.

Research staff : Charles A. Lunn, 5 chemists, 5 chemical engineers,
15 assistant chemists and 6 laboratory assistants.

Research work : Part time of staflF on problems consequent to the
manufacture and distribution of illuminating gas (coal gas and car-
buretted water gas).

zza. Consolidated Gas Electric Light and Power Company of Balti-
more, Lexington and Liberty Sts., Baltimore, Md. Laboratory at
Spring Gardens Plant, Baltimore, Md.

Research staflF: Minor C. K. Jones, 2 chemists and 5 laboratory
assistants.

Research work: One-tenth time of 8 on gas purification and gen-
eral gas manufacture.

Equipment: Complete experimental purifier equipment.

22 INDUSTRIAL RESEARCH LABORATORIES

1x3. Conwell, B. L*, ft Co., Inc., 2024 Arch St, Philadelphia, Pa.
(Engineera, chemiats, inapcctors.)

Reaearch staff: E. L. Conwell and 3-15 assistants.

Research work : Variable amount of time on cement manufacture ;
lime products manufacture; uses of cements, limes, etc; various in-
dustries, involving calcination, grinding, etc., and recovery and util-
ization of waste products.

1x4. Cooper Hewitt Electric Company, 730 Grand St., Hoboken, N.
J. (Lamps and rectifiers.)

Research staff : R. D. Mailey and 2 assistants.

Research work: Vapor electric apparatus and applications.

Equipment : Facilities for fabricating clear fused quartz apparatus
and methods for fusing (hermetic) clear quartz to all vitreous mate-
rials, including metallic leads.
1x5. Coming Glass Works, Coming, N. Y. (Technical glass.)

Research staff: E. C. Sullivan, 3 chemists, 5 physicists and 4 en-
gineers.

Research work: Two-thirds time of 11 on physical properties of
g:lass as related to chemical composition; lens design; furnace de-
sign ; refractories ; manufacturing problems ; and new uses for glass.

Equipment : Facilities for high temperature work.
Z16. Com Products Refining Cknnpany, Edge water, N. J.

Research staff : Christian E. G. Porst, 3 chemical engineers, 4 chem-
ists and 13 helpers and laborers.

Research work: Full time of 21 on problems confined to the in-
dustry.

Corona Chemical Co. See Pittsburgh Plate Glass Co. (p. 65).
117. Cosden & Company, Tulsa, Okla. (Producers and refiners of
petroleum.)

Research staff : Charles K. Francis and about 50 chemists, physi-
cists, engineers and assistants.

Research work : One-third time of about 50 on petroleum and pe-
troleum products, including gas.

Equipment: General chemical and physical equipment for petro-
leum work.
iz8. Cosmos Chemical Co., Inc., 709 Berckman St., Plainfield, N. J.

Research staff : Charles Blanc and 3 assistants.

Research work : Organic synthetic compounds for commercial util-
ization and factory problems.

zi9-iao. Cramp, William & Sons Ship & Engine Building Co., The,
Philadelphia, Pa.

119. /. P. Morris Hydraulic Laboratory

Research staff: F. H. Rogers, 2 engineers, 2 observers and i ma-
chinist.

Research work : Three-fourths time of 6 in the field of hydraulics
and h}rdrodynamics.

Ec[uipment: Hydraulic testing laboratory designed specially for
testing models of hydraulic turbines, centrifugal pumps, spiral pumps,
current meters, Pitot tubes, etc. Contains headrace flume, tailrace
flume, motor driven pumps, tank for rating current meters and other
necessary instruments.

INDUSTRIAL RESEARCH LABORATORIES 23

lao. Cramp Chemical Laboratory

Research staff: N. H. Schwenk and i chemist

Research work : Half-time of 2 on research work along metallurgi-
cal lines.
X9Z. Crane ft Co., Dalton, Mass. (Paper makers.)

Research staff: C. Frank Sammet.

Research work : Full time of i on development of new procedures,
novelties and mill problems.

Equipment: Well equipped for research relative to paper manu-
facture.

X99. Crane Co. (Metallurgical Department), South Avenue, Bridge-
port, Conn., and 836 South Michigan Ave., Chicago, 111. (Valves,
pipes, fittings and other supplies from iron, steel, brass and bronze,
for water, gas, and steam work.)

122a. Bndgeport laboratory

Research staff : Allen P. Ford, 2 metallurgists, i chemist, 3 assist-
ant chemists and 2 helpers.

Research work : Small part time of 9 on problems connected with
the industry.

Equipment: Entirely equipped for routine metallurgical work.
100,000-pound tensile testing machine; transverse, torsion and hard-
ness testing machines.

122b. Chicago laboratory

Research staff: L. W. Spring, i assistant and 12 men, 2 of whom
are doing physical and metalloppraphic testing.

Research work : One-tenth time of 14 on problems connected with
the industry.
193. Crompton ft Knowles Loom Works, Worcester, Mass.

Researjch staff: V. E. Hillman, 2 metallurgists, i chemist, i libra-
rian and 2 non-technical assistants.

Research work : Full time of 7 on heat treatment of steel ; case car-
burizing and cyanide hardening ; quenching mediums ; core oils ; mold-
ing sands; molding methods; blow holes and shrinkage cavities in
cast iron ; illumination ; copper plating metal parts ; and work on non-
ferrous alloys — ^aluminum, brass, bronze and bearing metals.
Z94. Crucible Steel Company of America, Pittsburgh, Pa.

Research staff: Charles Morris Johnson and 39 chemists and
physicists.

Research work: Chemical department, one-fifth time of 8 men.
Physical division, four-fifths time of 3 men.

Equipment : i Olsen 100,000-pound tensile testing machine, i Olsen
impact machine, i Olsen torsion machine, i Olsen new ductility ma-
chine for testing the ductility of plates up to one-fourth inch thick,
2 Pittsburgh Instrument Company Brinell- testing machines, 2 Shore
scleroscopes, i O-Z cutmeter tachometer, i Brown instrument (criti-
cal point machine), Leitz microphotographic outfit and i Olsen ex-
tensometer.

Z95. Cudahy Packing Co., The, South Side Station, Omaha, Nebr.
(Meat packers, etc.) General and research laboratory, Omaha, Nebr.
Laboratories also in Chicago, III., and Kansas City, Kans.

24 INDUSTRIAL RESEARCH LABORATORIES

Research staff: Millard Langfeld, superintendent of laboratories,
5 chemists and 2 workers.

Research work: Gland products, oils and greases, glues, curing
meats, etc.

126. Cumberland Mills, Cumberland Mills, Me. S. D. Warren Co.,
Boston, Mass., proprietors. (Pulp and paper.)

Research staff: E. Sutermeister, 2 to 4 chemists and 2 or 3 as-
sistants.

Research work : One-third time of 6 on problems relating to pulp
and paper industry. Tests of various woods and fibrous materials;
studies on soda and sulphite pulp processes and on solubility, adhe-
sive strength and viscosities of caseins and their solutions and coating
mixtures; studies of black ash waste and its possible utilization;
studies of rate of absorption of moisture by paper; investigations of
the storage conditions for pulp wood ; studies on the frothing of coat-
ing mixtures ; tests of new sizing agents and further studies on rosin
sizing. Bleaching studies on sulphite and soda fiber to show effects
of variable factors ; further applications of a beating test to show rela-
tive strength of fibers ; investigations relating to manufacture of satin
white; studies of defects in papers and of means to overcome them.

Equipment : Apparatus for the manufacture of paper on laboratory
scale ; complete testing apparatus. Available in mill ; 400-pound ver-
tical soda digester; 350-pound beater, and small Fourdrinier paper
machine. Apparatus to study foaming of coating mixtures.

127. Curtiss Aeroplane & Motor Corporation, Garden City, L. I.,
N. Y.

Research staff : H. T. Booth, 2 engineers, i mechanic and i model
maker.

Research work : One-half time of 4 on wind tunnel tests of wing^,
bodies, propellers, etc. Load tests of complete airplanes, perform-
ance tests of complete airplanes and miscellaneous investigations
along different aeronautical lines.

Equipment : One four-foot wind tunnel in which wind velocities of
75 m. p. h. are obtained. One seven-foot wind tunnel in which wind
velocities of 100 m. p. h. are reached.

Curtiss Engineering Corporation, The. See Curtiss Aeroplane &
Motor Corporation.

128. Cutler-Hammer Mfg. Co., The, Milwaukee, Wis. (Electric con-
trolling devices.)

Research staff: Arthur Simon, i physicist, i glassblower and me-
chanical helpers as needed. Has help of Experimental Department
with its staff of developing engineers and mechanics.

Research work: Full time of 2 in connection with electrical dis-
charge in gas, particularly evacuated tubes and bearing on control of
electric currents.

129. Davis-Boumonville Company, Jersey City, N. J. (Welding and
cutting apparatus.)

Research staff: Frank J. Napolitan and i assistant.

Research work: Large part time of 2 on metallography of oxy-
acetylene welding, design of new apparatus and development of scope
of process.

INDUSTRIAL RESEARCH LABORATORIES 25

Equipment: Gas laboratory equipped for measuring flow of gas
under high pressures; micro-manometers for measurement of high
pressures.

130. Davis Chemical Products, Inc., Springfield, N. J.
Research staff: E. J. Fry, i engineer and i chemist.

Research work: One-half time of 3 on cellulose esters, nitrocellu-
lose, nitrocellulose solvents and solutions, artificial and imitation
leather, coatings, lacquers and films ; explosives, commercial and mili-
tary.

Equipment : Apparatus for testing the physical and chemical prop-
erties of films and coatings based on cellulose esters, including viscos-
ity, stability, aging, accelerated life tests and strength; facilities for
large scale experiments and demonstrations.

131. Davison Chemical Company, The, Baltimore, Md. (Sulphuric
acid.)

Research staff: A. E. Marshall and trained research men as re-
quired.

Research work : Full time of staff on improvement of manufactur-
ing processes for sulphuric acid and utilization of waste materials.

Equipment: Semi-commercial equipment for development of proc-
esses evolved in laboratory.

Dayton Engineering Laboratories Company. See General
Motors Research Corporation (p. 35).

133. Dean Laboratories, Inc., ^th St. and Walton Ave., Philadelphia,
Pa.

Research staff: J. Atlee Dean, 3 chemists, 3 bacteriologists and i
technician and clerical helper.

Research work : One-half time of 8 on physiological, pharmaceuti-
cal and clinical chemistry; hypodermic preparations, especially the
endocrine glands ; laboratory reagents such as colloidal gold and mi-
croscopic stains.

Equipment : Facilities for rapid and accurate examinations of body
fluids.

133. Dearborn Chemical Company, McCormick Building, Chicago,
111. (Scientific boiler feed water treatment.) Laboratories at 1029
W. 35th St., Chicago, 111.

Research staff : D. K. French, 5 chemists and 5 assistants.

Research work : Small part time of 1 1 on scientific boiler feed water
treatment and chemical control of corrosion.

Equipment: Hess-Ives tintometer and Thurston friction machine;
all types of viscosimeters.

134. Dehls & Stein, 237 South St., Newark, N. J. (Manufacturing
chemists.)

Research staff : L. Stein and i chemist.

Research work: One-half time of 2 along lines of fermentology,
synthetic essential oils, caramel.

135. Deister Concentrator Company, The, 611 High St., Ft. Wayne,
Ind. (Concentrating tables for every purpose.)

Research staff : Regular force consists of i metallurgical engineer,
together with occasional assistance in advisory capacity from other
members of the company.

26 INDUSTRIAL RESEARCH LABORATORIES

Research work : On gravity or table concentration of various ores
sent us for this purpose from all parts of the world; extensive work
in the washing of the finer sizes of coal (both anthracite and bitumi-
nous) below Uiat usually handled on jigs, etc. This work is done in
both small lots and in carload quantities.

Equipment: One i6 by i8-inch Pennsylvania roll crusher, I pair
lo-inch corrugated rolls, i pair 5>^-inch smooth rolls for regrinding,
I Mitchell vibrating screen, i No. 7 Deister-Overstrom diagonal dedc
coal-washing table, i No. 6 Deister-Overstrom diagonal deck table for
ore treatment, i No. 14 Deister-Overstrom diagonal deck, jr., table,
I 12-foot Dorr thickener, i size 4-1 American vacuum filter, i Inger-
soll-Rand vacuum pump.

136. DeLaval Separator Co.» That 165 Broadway, New York, N. Y.
(Centrifugal machinery.)

Research staff: A. F. Meston and i assistant.

Research work : Full time of 2 on purifying used oils, clarification
and separation of commercial products, making of emulsions, clari-
fication of extracts, purifying of crude and fuel oils, application of
centrifugal machines to industrial processes, etc.

Equipment : Centrifugal apparatus of all classes.

137. Dennis, Martin, Company, The, 859 Summer Avenue, Newark,
N. J. {Chrome tannage.)

Research staff : Harold Dennis, i chemical engineer and 2 chemists.
Research work: Three-fourths time of 4 on tanning and tanning
materials.

138. Detroit Edison Company, The, Detroit, Mich. (^Operating elec-
tric light and power generating stations and distributin|^ systems;
central heating stations and distributing systems and illummating gas
plants and distributing systems.)

Research staff : C. r . Hirshfield, i engineer, 2 to 8 trained men, and
4 or more assistants.

Research work: Problems in better generation, distribution and
utilization of electricity, steam for heating and artificial gas.

139. Detroit Testing Laboratory, The, 3726 Woodward Ave., Detroit,
Mich. (Analytical consulting and research chemists.)

Research staff : W. P. Putnam, 6 chemists, i bacteriologist, i chem-
ical engineer and i electrical and mechanical engineer, i pharmaceuti-
cal engineer, i foundry engineer, i steam engineer and i automobile
engineer.

Research work : Full time of i chemist and 2 engineers on special
problems in shale oil development, fertilizer manufacture, metallurgi-
cal problems, heat treatment of metals, fuel problems, water purifica-
tion and ore dressing.

Equipment: 100,000-pound Reihle testing machine, 10,000-pound
Olsen testing machine, Weston precision laboratory type instruments,
shunts and multipliers for instrument calibration and precision test-
ing, Leeds and Northrup precision type potentiometer and large ca-
pacity storage batteries.

140. Dewey & Akny Chemical Companyt Harvey St., Cambridge,
Mass.

INDUSTRIAL RESEARCH LABORATORIES 27

Research staff: Bradley Dewey, i chemical engineer and 2 chem-
ists.

Research work : One-half time of 4 on adhesives, fluxes, and seal-
ing compomids.

141. Dartre Products, Inc., 25 Illinois St., Buffalo, N. Y. (Soluble
starch and dextrin products.)

Research staff: A. D. Fuller and 2 assistants.

Research work : One-fourth time of 3 on hydrolysis of starch, tor-
rification of starch, colloids as related to adhesives and dextrin.
243. Diamond Chain ft Manufacturing Company, 502 Kentucky Ave.,
Indianapolis, Ind. (Steel roller and block chains, sprockets, etc.)

Research staff: H. B. Northrup, i chief metallurgist and i assistant
metallurgist.

Research work: Approximately one-half time of 3 on carburizing
compounds and carburizing, hardening and drawing of alloy vs. plain
carbon steels for chain parts.

143. Diamond Match Co., The, Oswego, N. Y.

Research staff : Frederick VanDyke Cruser, 7 chemists and chemi-
cal engineers, i mechanical en^neer and 3 assistants.

Research work: One-half time of 12 on problems connected with
match manufacture and its allied branches.

144. Dicks David Company, Incorporated, Varick and N. Moore Sts.,
New York, N. Y. (Dyestuffs and chemicals.) Laboratory at 22d
St. and Stewart Ave., Chicago Heights, 111.

Research staff: H. Philipp, P. H. Condit, W. G. Brunjes, 8 chem-
ists and 4 engineers.

Research work: Small part time of 15 chiefly on triphenylmethane
dyestuffs.
Z45. Digestive Ferments Co., Detroit, Mich.

Research staff : Howard T. Graber, director of the chemical labora-
tory; Henry G. Dunham, director of the bacteriological laboratory,
and assistants.

Research work: Two-thirds time of assistants devoted to physio-
logical and proteid chemistry and commercial classification of bac-
teriology.

Equipment: Apparatus for the electrometric estimation of hydro-
gen ion concentration. Vitreosil mufile furnace with thermocouple
and Brown recording pyrometer for the accurate estimation of ash at
definite temperatures. Experimental laboratory vacuum drier, ther-
mocouple and recording thermometer for moisture determinations.
Schmidt and Haensch saccharimeter with bichromate cell.

146. Dill ft Collins Co., Richmond and Tioga Sts., Philadelphia, Pa.
(Paper makers.)

Research staff: Frank H. Mitchell, 2 chemists, 2 chemical engi-
neers and 3 assistants.

Research work : One-half time of i chemist to full time of 2 chem-
ists on problems of the paper industry.

147. Dodge Brothers, Detroit, Mich. (Automobiles and accessories.)
Research staff: F. E. McCleary, 17 chemists, 25 engineers, physical

testers and trouble men.
Research work : Approximately one-tenth time of staff on automo-

28 INDUSTRIAL RESEARCH LABORATORIES

bile materials, treatment, application, etc. This covers cast iron, steel,
brass and bronze, babbitt, aluminum, wood, rubber, etc.; lubrication,
paints and varnishes, baking japans and fuel.

148-150. Doehler Die-Casting Co., Court, Ninth and Huntington Sts.,
Brooklyn, N. Y. Laboratories also at Smead and Prospect Aves., To-
ledo, Ohio, and at Chicago, 111.

148. Brooklyn Laboratory

Research staff: Charles Pack, 5 chemists, 6 junior chemists, i fuel
engineer, i steel metallurgist.

Research work: One-fifth time of 14 on problems pertaining di-
rectly or indirectly to casting of metals, particularly non-ferrous
metals.

149. Toledo Laboratory

Research staff: Charles Pack, i metallurgist, i chemist and 5 junior
chemists.

Research work : One-fifth time of 8 on problems pertaining to cast-
ing of metals.

150. Chicago Laboratory ,

Research staff: J. C. Fox and 2 chemists.

Research work: One-tenth time of 3 on non-ferrous alloys.
151. Doherty Research Company, Empire Division, Bartlesville,
Okla.

Research staff: J. P. Fisher, i superintendent and 10 engineers.

Research work : Full time of 12 on research problems dealing with
production, transportation and refining of petroleum; transportation
and distribution of natural gas ; conservation of fuel.
15a. Dorite Manufacturing Company, The, 116 Utah St., San Fran-
cisco, Calif. (Stucco, flooring, magnesite.)

Research staff: E. H. Faile and i assistant.

Research work: One-half time of 2 on investigation of the best
methods for the manufacture of various magnesite products, including
stucco and flooring and particularly of the most practical methods in
their application and use.

153. Dorr Companv, The, loi Park Ave., New York, N. Y. (Engi-
neers.) Testing plant and laboratory at Westport Mill, Westport,
Conn.

Research staff : H. A. Linch, i analytical chemist, i chemical engi-
neer, I sanitary engineer, i mechanical engineer, 4 assistants. Chemi-
cal, metallurgical, sanitary and mechanical engineers from the New
York ofHce are available for advice and work as needed.

Research work: Major problems in connection with the produc-
tion of water-floated materials for pigments, fillers, etc. Concentra-
tion and sulphating. Roasting of ores. Washing and classification
of abrasives. Studies dealing with the development of mechanical set-
tling and dewatering, classification, continuous agitation and counter-
current washing. Trade waste and sewage treatments.

Equipment: Bins, crushers, grinding mills, classifiers and washers
of various types, thickeners, filterers, concentrating tables, flotation
machines, mechanical multiple-hearth furnace, electric roasting fur-
nace, etc. Plant fully equipped to work out hydrometallurgical and
wet chemical and industrial problems.

INDUSTRIAL RESEARCH LABORATORIES 29

154. Drackett» P. W., & Sons Co., The, Cincinnati, Ohio. (Manu-
factures heavy chemicals; distributes Solvay Process Co. alkalis and
other heavy chemicals.)

Research stafiF: K. S. Kersey and i assistant.

Research work : Development of products and their uses.

155. Dunham, H. V., 50 E. 41st St., New York, N. Y.
Research staff : H. V. Dunham with from 2 to 6 assistants.
Research work : Full time of staff on food products, oils, including

mineral oils and especially developments and improvements in the
making and use of milk casein and milk products.

Equipment: Mixing machines, dryers and other semi-industrial
equipment.

156-160. du Pont, E. I., de Nemours & Company, Wilmington, Del.
Chemical Department operates 5 research laboratories in addition to
organization at its main office. (Information concerning the entire
department is followed by separate accounts of the 5 laboratories.)

Research staff: Charles L. Reese, 200 graduate chemists and en-
gineers, 122 other salaried employees and 200 payroll employees.

Research work: Practically full time of 522 on manufacturing
operations of the du Pont Company, including miscellaneous chem-
icals, dyes and intemediates, explosives, artificial leather, rubber goods,
plastics, pyroxylin solutions, lacquers, paint and varnish, including
the production of miscellaneous raw materials as mineral acids and
nitrate of soda.

156. Pyralin Laboratory, Arlington, N. /.

Research staff : E. A. Wilson, 22 graduate chemists and engineers,
13 other salaried employees and 24 payroll employees.

Research work: Practically full time of 59 on pyralin, pyroxylin
solutions, and raw materials therefor.

Equipment: Fairly complete line of semi-manufacturing scale
equipment for the experimental manufacture of paper, nitrocellulose
and pyralin.

157. Eastern Laboratory, Box 424, Chester, Pa,

Research staff: C. A. Woodbury, 23 graduate chemists and en-
gineers, 13 other salaried employees and 33 payroll employees.

Research work : Practically full time of 69 on high explosives and raw
materials therefor, processes of manufacture, and methods of testing.

Equipment: Very complete facilities for testing properties of ex-
plosives.

158. Experimental Station, Henry Clay, Del.

Research staff : A. P. Tanberg, 28 graduate chemists and engineers,
30 other salaried employees and 63 payroll employees.

Research work: Practically full time of 121 on smokeless powder,
black powder, nitrocellulose, heavy chemicals, paint and varnish, and
raw materials therefor. Also miscellaneous organic, inorganic, and
biochemical research.

Equipment: For experimental manufacture of propellant powders,
constant temperature magazines for stability tests, and storage of
smokeless powder, experimental equipment for the manufacture of
coated fabrics, ranges for testing small arms powders for velocity,
pressure and accuracy.

30 INDUSTRIAL RESEARCH LABORATORIES

1 59. Jackson Laboratory, Box $25, Wilmington, Del.

Research staff: Fletcher B. Holmes, 80 graduate chemists and
ennneers, a8 other salaried employees and 71 payroll employees.

Research work: Practically full time of 179 on dyes and inter-
mediates.

Equipment: Extensive equipment for semi-works operation and
investigation of a variety of chemical processes.

160. Redpath Laboratory, Parlin, N. J.

Research staff: E. B. 6enger» 14 graduate chemists and engineers,
8 other salaried em]>loyees and 7 payroll employees.

Research work : Practically full time of 29 on film work.

Equipment: Small scale apparatus for coating films, and equip-
ment for physical and chemical testing of film and photo-chemist^.
z6i. Durfee, Winthrop C, 516 Atlantic Ave., Boston, Mass. (Con-
sulting and manufacturing chemist.)

Research staff : Winthrop C. Durfee, 5 chemists, i physicist and 3
assistants.

Research work: One-half time of 10 on application of dyes and
chromium compounds in wool dyeing; chrome tanning.
z69. Duriron Company, Inc., The, N. Findlay St., Dayton, Ohio.
(Acid-proof alloy castings.)

Research staff : P. D. Schenck, i metallurgist, i chemist, i assistant
chemist, i engineer and i laboratory assistant

Research work: One-fourth time of 6 on chemical corrosion of
metals, metallurgical problems, physical properties, etc.; problems
relating to the handling of corrosives.

Equipment: Experimental foundry.
263. Dye Products & Chemical Company, Inc., aoo 5th Ave., New
York, N. Y.

Research staff : C. K. Simon, i chemist and 2 assistants.

Research work : Full time of i chemist and part time of 2 assistants
on problems connected with the manufacture of dyes and intermedi-
ates and the improvement of present processes.

164. Eagle-Picher Lead Company, The, 208 S. LaSalle St., Chicago,
111. (Manufacturers, miners and smelters of lead products.) Labora-
tory at Joplin, Mo.

Research staff: J. H. Calbeck and 4 chemists.

Research work : Full time of 5 on physical and chemical properties
of paints and white pigments ; storage battery oxides and chemical and
metallurgical problems pertaining to the manufacture and uses of the
oxides of lead and zinc.

Equipment : Pfund's colorimeter, spectrometer, photometer, micro-
photographic equipment.

165. Eastern Finishing Works, Inc., Kenyon, R. I.
Research staff: William H. Adams and 2 assistants.

Research work : Part time of 3 on test valuation and general study
of waterproofing, dyeing, sizing and mildew resistance in connection
with finishing cotton goods.

166. Eastern Malleable Iron Company, Naugatuck, Conn. (Cast-
ings.)

INDUSTRIAL RESEARCH LABORATORIES 31

Research staff: W. R. Bean and 6 assistants.

Research work: Full time of 3 and one-half time of 4 on metal-
lurgical research as applied to composition, annealing and production
of malleable iron.

Equipment: Special laboratory muffle annealing furnace, elec-
trically heated, with automatic electric temperature control bath for
maintaining indefinitely temperatures up to 2000^ F. and also con-
trolling rate of heating and cooling at several rates between 4^ F. per
hour and 20** F. per hour.
Z67. Ea3tem Manufacturing Company, Bangor, Me. (Paper.)

Research staff : H. H. Hanson, 5 chemical engineers, 2 chemists, 3
routine chemists, i electrical engineer and i assistant.

Research work: Full time of 14 on standardization of processes,
increasing production, development of by-products and development
of improved processes.

Equipment: Small paper beater, apparatus for determining slow-
ness of beater stock, strength of stock in beaters and on finished-paper.
z68. Eastman Kodak Company, Rochester, N. Y.

Research staff: C. E. K. Mees, 45 chemists, physicists and photo-
graphic experts and 60 assistants.

Research work: Full time of 105 on theory of photography, de-
velopment of new photographic materials and methods, and the study
of the theory of manufacturing processes, and the production of syn-
thetic organic chemicals.

Equipment : Sensitometric and lens testing apparatus, physical and
colloidal chemical apparatus for use in the study of photographic
theory.

169. Eavenson ft Levering Co., 3rd and Jackson Sts., Camden, N. J.
(Wool scouring and carbonizing.)

Research staff: Chas. E. Mullin, 2 or 3 chemists and assistants.

Research work: Approximately one-half time of staff on textiles,
wool particularly ; wool scouring, carbonizing and dyeing ; utilization
of wool waste and refuse such as scouring liquors ; wool grease and
detergents.

170. Edison, Thomas A., Laboratory, Orange, N. J.

Research staff: Thos. A. Edison and about 250 machinists, chem-
ists, physicists, experimenters, designers and draughtsmen.

Research work : Nearly full time of 250 on almost every branch of
scientific research.

Equipment: Large scrap heap from which to rob to build other
apparatus, and accumulations of every kind of material and chemical
so as not to wait.

171. Eimer ft Amend, Third Ave., i8th to 19th St., New York, N. Y.
(Industrial and educational laboratory apparatus, assayers' materials,
chemicals and drugs.)

Research staff: O. P. Amend, C. G. Amend, 2 chemists, 4 expert
glass blowers and i mechanic.

Research work : Organic chemicals and special glass and metal ap-
paratus for scientific investigations.

Z72. Electrical Testing Laboratories, 80th St. and East End Ave.,
New York, N. Y.

32 INDUSTRIAL RESEARCH LABORATORIES

Research staff : Clayton H. Sharp, i chief engineer and 7 research
men.

Research work : One-tenth time of 9 on dielectric losses ; thermal
conductivity of heat insulators at high and low temperatures; radia-
tion efficiency of gas heaters; special cases of electrolysis by stray
currents ; breakdown voltage of sheet insulation.

Equipment: Very complete for electrical standardizing and re-
search, photometry, mechanical measurements, fuel testing, paper
and textile testing, thermometer and pyrometer standardization.

173. Electro Chemical Company, The, Dayton, Ohio. (Electrolytic
cells for producing sodium hypochlorite.)

Research staff: John Gerstle and i chemical engineer.

Research work : Two-thirds time of 2 in connection with producing
sodium hypochlorite from a sodium chloride solution, principally in-
creasing efficiency of electrolytic cells.

174. Electrolabs Company, The, 2635 Penn Ave., Pittsburgh, Pa.
(Electrolytic gas specialists.)

Research staff: I. H. Levin, i chemist, i engineer and i physicist.
Research work : Full time of 4 on electrolytic dissociation of water,
application of hydrogen to vegetable oil refinement, etc.

Electro Metallurgical Company. See Union Carbide and Car-
bon Research Laboratories, Inc. (p. 78).

Ellis, Carleton, Laboratories. See Ellis-Foster Company.

175. Ellis-Foster Company, 92 Greenwood Ave., Montclair, N. J.
(Chemical products and processes.)

Research staff : Carleton Ellis and a variable number of assistants.
Research work : Approximately full time of staff on organic chem-
istry and ceramics.

176. Emerson Laboratory, 145 Chestnut St., Springfield, Mass.
Research staff: H. C. Emerson and 5 chemists.

Research work: One-fourth time of 6 on paper and textile prob-
lems.

Empire Gasoline Co. See Doherty Research Company, Empire
Division (p. 28).

Empire Tannery. See Gallun, A. F., & Sons Co. (p. 34).

177. Eppley Laboratory, The, 12 Sheffield Ave., Newport, R. L
(Physical-chemical laboratory.)

Research staff : Warren C. Vosburgh, 2 chemists and i instrument
maker.

Research work: One-half time of 4 on cadmium standard cells,
physico-chemical apparatus, standards of electromotive force, spec-
troscopy, theory of solutions from electrical standpoint and thermo-
couples for precise measurements.

Equipment : Spectroscopes and potentiometers.

178. Eustis, F. A., 131 State St., Boston, Mass. (Metallurgical en-
gineer.)

Research staff: F. A. Eustis.

Research work: Part time of i on metallurgical problems con-
nected with copper, sulfur and iron and the purification of smelter
smoke.

179. Factory Mutual Laboratories under the supervision of Asso-

INDUSTRIAL RESEARCH LABORATORIES 33

ciated Factory Mutual Fire Insurance Companies, Inspection Depart-
ment, 31 Milk St., Boston, Mass.

Research staff: C. W. Mowry, 2 chemists and 8 engineers.

Research work: One-sixth to one-fourth time of 11 on fire-protec-
tion engineering problems.

Equipment: Apparatus for chemical, hydraulic and mechanical
tests and investigations of fire-protection devices.
i8o. Fahy» Proc^ P., 50 Church St., New York, N. Y.

Research staff: Frank P. Fahy.

Research work: Full time of i on magnetic-mechanical analysis of
iron and steel products.

Equipment : Special magnetic testing devices.
i8i. Falls Rubber Company, The, Cuyahoga Falls, Ohio.

Research staff : G. D. Kratz, 4 chemists and 2 eng^ineers.

Research work : One-half time of 5 and one-fourth time of 2 on the
investigation of raw rubbers and the process of vulcanization; new
machines and mechanical methods.

Equipment : For the study of problems in the vulcanization of rub-
ber.

i8a. Fansteel Products Company, Inc., North Chicago, 111. (Elec-
trical, steel and chemical products.)

Research staff: Clarence W. Balke, 2 chemists, i engineer, and i
assistant.

Research work: One-half time of 5 on rare metals, tungsten,
molybdenum, cerium, tantalum and columbium.

183. Feculose Co. of America, Ayer, Mass. (Pastes, adhesives, size,
etc.)

Research staff: John T. Gibbons and 3 chemists.

Research work : Full time of 4 on starches and starch products.

184. Federal Phosphorus Company, Anniston, Ala.
Research staff: J. N. Carothers and 3 chemists.

Research work : Full time of 2 men on plant process for production
of phosphoric acid by electric smelting of phosphate rock ; production
of phosphoric acid salts.

185. Federal Products Company, The, 7818 Lockland Ave., Cincin-
nati, Ohio. (Cologne spirits and denatured alcohol.)

Research staff: J. F. Kraeger and i assistant chemist.

Research wotk : One-half time of 2 on production of ethyl alcohol
from materials containing fermentable substances and recovery of
valuable by-products from distillery waste.
x86. Firestone Tire & Rubber Company, Akron, Ohio.

Research staff: E. W. Oldham, director of general laboratory; N.
A. Shepard, director of organic research ; E. C. Zimmerman, director
of physical chemical research and 20 chemists and engineers; J. E.
Hale, director of development department, and 12 engineers.

Research work : Full time of o on study of vulcanization, physical
and chemical properties of vulcanized rubber in conjunction with
various accelerators and compounding materials, and problems aris-
ing in connection with the manufacture of rubber products.
187. FitzGerald Laboratories, Inc., The, Niagara Falls, N. Y.

Research staff: F. A. J. FitzGerald and 3 assistants.

34 INDUSTRIAL RESEARCH LABORATORIES

Research work: One-half time of 4 on electric furnaces, refrac-
tories and electrometallurgy.

Equipment: For electro-thermal laboratory.
z88. Florida Wood Products Co., Jacksonville, Fla. (Phosgene gas.)

Research staff: E. B. Smith and i chemist

Research work : Part time of 2 on development of products of phos-
gene gas; pharmaceuticals derived from wood products.

Equipment: Special facilities for handling destructive distillation
problems, being equipped with iron retorts capacity of 50 pounds to
1500 cubic feet.

289. Fort Worth Laboratories, Box 1008, Fort Worth, Texas. (Con-
sulting, analytical chemists and chemical engineers.)

Research staff: F. B. Porter, R. H. Fash, and assistants, 6 chem-
ists and about 8 helpers.

Research work: Small part time on industrial problems as pre-
sented, cotton oil refining and boiler water problems.

190. Foster-Heaton Company, 27 Badger Ave., Newark, N. J.
Research staff : Edward W. Rhael, i chemist and i engineer.
Research work : Approximately one-third time of 3 on development

of coal tar dyestuffs soluble in oils, fats and waxes.

191. Frees, H. E., Co., The, 2528 W. 48th Place, Chicago, 111. (Brew-
ers and distillers laboratory.)

Research staff: Herman E. Frees, i chemist and i fermentologist.
Research work : Approximately one-half time of 3 on foods, yeasts,
fermentation and beverages.

292. Fry, H. C, Glass Company, and Beaver Valley Glass Co.,
Rochester, Pa.

Research staff : R. F. Brenner and 2 assistants.

Research work : More than one-half time of 3 on new varieties and
compositions of glass. This work is carried out first in small crucible
meltings and then in regular factory pots.

Equipment: High-temperature gas-fired furnace.

293. Gallun, A. F., & Sons Co., Milwaukee, Wis. (Proprietor, Em-
pire Tannery.)

Research staff: John Arthur Wilson and 7 chemists.

Research work: Approximately four-fifths time of 8 on experi-
mental tanning, pure and applied colloid chemistry, physical chemis-
try, photomicrography, ultramicroscopy, histology of skin, and special
applications of concentration cells.

Equipment: Experimental tannery.
194. Garfield Aniline Works, Inc., Box 196, Passaic, N. J. Labora-
tory at Garfield, N. J.

Research staff : Arthur F. F. Mothwurf and 6 chemists.

Research work : Full time of 6 on coal tar intermediates, coal tar
dyes (azo-colors and triphenylmethane derivatives) and sample dye-
ing.

295. General Bakelite Company, Perth Amboy, N. J. Supplementary
laboratory in Yonkers, N. Y.

Research staff : L. H. Baekeland, 2 engineers and 5 chemists.

Research work: Full time of 8, confined almost exclusively to

INDUSTRIAL RESEARCH LABORATORIES 35

phenol-formaldehyde condensation products, both development and
commercial applications.

Equipment: In form of electric ovens, stills, vulcanizers, pebble
mills and rubber machinery.

196. General Chemical Company, Research Department, 25 Broad
St., New York, N. Y.

Research staff : G. P. Adamson and approximately 45 chemists.

Research work : Full time of 46 on improving existing processes of
the company, and devising new processes.

General Chemical Company has recently become a part of the Allied
Dye & Chemical Corporation and reorganization of its research de-
partment is now in progress.

297. General Electric Company, Schenectady, N. Y. Laboratories
also at Lynn and Pittsfield, Mass., Harrison, N. J. and Cleveland,
Ohio.

Research staff: Willis R. Whitney, 2 assistant directors, 50 chem-
ists, 12 physicists, 13 engineers, 50 research assistants, and machinists,
glass-blowers, electricians and clerks.

Research work: Full time of staff devising new forms of electric
lights and improving existing forms. Development of Coolidge X-ray
tube. Invention of new and development of existing forms of electric
equipment and apparatus. Study of metals and alloys for electrical
uses. Wireless transmission development. Study of insulation.
Many fundamental physical and chemical scientific researches also
are carried on.

See National Lamp Works of General Electric Company (p. 56).
198. General Engineering Company, Incorporated, The, 159 Pier-
pont St., Salt Lake City, Utah. (Consulting engineers, ore testing.)

Research staff: J. M. Callow, i chemist, 2 metallurgical engineers
and 2 helpers.

Research work: Full time of 6 on metallurgical and engineering
problems, specializing on ore treatment problems.
X99. General Motors Research Corporation, Box 745, Moraine City,
Dayton, Ohio.

Research staff: C. F. Kettering, president and active directing
engineer, F. O. Clements, director of research, and 251 employees,
divided into specialized groups or departments, made up of chemists,
metallurgists, electrical engineers, mechanical and other research
engineers, assistants and helpers. (Control division made up of 147
additional employees and manufacturing division, having at the pres-
ent time 18 members, bring the total number of employees up to. 416.)

Research work: Full time of staff on strictly automotive research
of interest to General Motors Corporation.

Equipment : Laboratories capable of conversion, upon short notice,
into mechanical, chemical or electrical laboratories. Complete shop,
foundry and heat treat departments.
200. General Tire & Rubber Co., Akron, Ohio.

Research staff: H. B. Pushee and 2 men.

Research work : One-tenth time of 3 on development of better rub*
ber compounds ; rubber accelerators ; coefficient of vulcanization.
20X. Gibbs Preserving Company, 2303 Boston St., Baltimore, Md.

36 INDUSTRIAL RESEARCH LABORATORIES

Research staff : David R. Dotterer and i assistant.

Research work : Canned goods and jellies.
3oa. Gillette Safety Razor Co.. 47 W. ist St., Boston, Mass.

Research staff : Henry E. K. Ruppel, 4 chemists, i special engineer
and technicians.

Research\ work : Part time of 6 or more on development and im-
provement of analytical methods; precision measurements; heat
treatment of steel: (a) metallographic investigations, (b) practical
applications; electro-deposition of metals; abrasives; study of edges
with special reference to shaving.

303. Glass Container Association of America, 3344 Michigan Ave.,
Chicago, 111.

Research staff : A. W. Bitting and 4 assistants.

Research work: Full time of 5 on standardization of glass con-
tainers, improved methods of packing glassware for shipment, foods
and beverages in glass and improvement in containers and closures.

Equipment : Complete equipment for the preparation and packing
of foods in glass and testing bottles, jars and packing materials.
203a. Glidden Company, The, Cleveland, Ohio. (Paints, varnishes,
enamels, stains, dry colors, insecticides, vegetable oils.)

Research staff: F. M. Beegle, chief chemist, 6 chemists, 2 chemical
engineers and a number of physicists. Research committee of 7 mem-
bers, comprised of the general superintendent and the head of each
manufacturing department.

Research work: The greater part of the time of the members of
the research committee, as well as that of all the chemists, is spent on
research or development work on synthetic gums, treated oils, var-
nishes, paints, enamels, stains, dry colors, and insecticides.

Equipment: Stacks for oil boiling and varnish making; an elec-
trically heated humidor, the humidity and temperature of which can
be controlled and regulated to duplicate the conditions of various
manufacturing plants; an oil treating plant and spraying apparatus.

204. Globe Soap Company, The, St. Bernard, Ohio.

Research staff: C. P. Long, chemical director, 3 chemists and 2
chemical engineers.

Research work : One-tenth time of 6 on investigation of problems
connected with the industry.

205. Glysyn Corporation, The, New York, N. Y. Laboratory at
Bound Brook, N. J.

Research staff : Harold F. Saunders and 3 chemists.
Research work : Full time of 4 on chlorination processes.

206. 'Goodrich, B. F., Company, The, Akron, Ohio. (Rubber goods
of every description.)

Research staff: W^ C. Geer, vice-president, in charge of develop-
ment. Research physical laboratory: 4 physicists and 4 assist-
ants. Engineering and testing laboratory: 3 engineers and 2 engi-
neering assistants. Chemical laboratories: 8 chemists and 3 assist-
ants. Development laboratories: 18 chemical engineers and 26
assistants.

Research work: The entire time of the staff is spent on research
and factory control work, although in rubber the factory control is

INDUSTRIAL RESEARCH LABORATORIES 37

never quite distinguishable from research. The fundamental lines of
research are those of compounding ingredients, including the chemical
and physical properties of crude rubber, reclaimed rubber, mineral
ingredients, and organic chemical individuals, the study of vulcaniza-
tion, and in particular the main efforts have to do with the physical
and chemical design of compositions and articles for particular lines
of industrial service.

Equipment: Development laboratory equipped with mills, vul-
canizing apparatus, etc.
307. Goodyear Tire & Rubber Company, The, Akron, Ohio.

Research staff: Wm. S. Wolfe, development manager, K. B. Kil-
bom, experimental engineer in charge of machine design, tire desi|^
and highways transportation divisions; R. C. Hartong, chief chemist
in charge of development service and chemical and physical research ;
W. E. Shively, chief tire designer, H. E. Morse, manager mechanical
goods development and service division; 4 chemical engineers, 3 as-
sistant chemical engineers, 8 research chemists, 5 research physical
chemists and physicists, 9 research engineers, 25 technical service,
chemical and mechanical engineers, 8 chemical laboratory chemists
and assistant chemists, 18 physical laboratory assistants, 8 mechanical
goods design engineers, 11 tire design engineers, 6 assistant tire de-
sign engineers, 12 compound development chemists, 6 machine design
engineers, 12 machine designers, 5 machine design detailers and
tracers, 37 machine design workshop machinists, 10 machine design
expert template makers, 2 highway transportation engineers. Total
employees of department approximately 360.

Research work: Full time of research and development men on
mechanism of vulcanization, compounds which affect the rate of vul-
canization, development of organic compounds especially adapted to
rubber work ; application of physical chemistry to study of rubber and
compounding materials; physical properties of rubber, and methods
of testing and studying them ; chemistry of fibrous materials, particu-
larly cotton, and properties of materials used as films or protective
agents; industrial processes, such as reclaiming and coagulation of
rubber.

ao8. Grasselli Chemical Company, 1300 Guardian Bldg., Cleveland,
Ohio. Laboratory at Cleveland mainly for inorganic work. Labora-
tory also at Grasselli, N. J., for organic work strictly.

Research staff: Henry Howard and a large number of chemists
and assistants.

Research work: Full time of staff on problems connected with
possible improvements in products at present being manufactured as
well as in connection with chemicals, dyes, intermediates, etc., the
manufacture of which is being contemplated.

209. Gray Industrial Laboratories, The, 961 Frelinghuysen Ave.,
Newark, N. J.

Research staff: Thomas T. Gray, David Drogin, G. C. Hargrove,
E. V. Espenhahn and assistants.

Research work : Full time of 2 on petroleum and its products.

Equipment: Complete semi-commercial oil refining equipment.

38 INDUSTRIAL RESEARCH LABORATORIES

2Z0. Great Western Electro-Chemical Company, 9 Main St., San
Francisco, Calif. (Chlorine products.)

Research staff : Ludwig Rosenstein, 2 chemists and 2 assistants.

Research work: Utilization of chlorine, manufacture of chlorine
products, manufacture of caustic and electrolysis of brine,
azz. Great Western Sugar Company, The, Sugar Building, Denver,
Colo.

Research staff: H. W. Dahlberg, i chief chemist, 4 chemical engi-
neers, 4 research chemists, 2 mechanics, i experimental process man,
3 analysts.

Research work: Full time of 16 on investigations of fundamental
principles of processes and practices now in use, examination of pro-
posed new processes and apparatus and study of utilization of by-
products and waste products; production of crude potash, sodium
cyanide, ammonium sulphate and certain rare organic chemicals from
the Steffen's waste water; refining of crude potash leading to pro-
duction of carbonate, hydrate, etc.; recovery of organic acids from
waste waters.

Equipment: Complete equipment for manufacture of sugar on a
small scale under such conditions that special attention may be paid
to any stage of the process.

2xa. Grosvenor, Wm. M., 50 E. 41st St., New York, N. Y. (Con-
sulting chemist and factory engineer.)

Research staff: From 2 to o.

Research work: Flotation of ores, non-ferrous metallurgy, paper,
starch, glues and adhesives, textiles, paper and their finishing,
methods of manufacture of organic intermediates, utilization of by-
and waste products.

Equipment: Viscosimeters, high speed moving picture equipment,
autoclaves up to 1000 lbs. per sq. in.

2x3. Gulf Pipe Line Company, Houston, Tex. (Producers and trans-
porters of petroleum.)

Research staff: F. M. Seibert and 2 trained research men.

Research work: Full time of 3 on methods for production and
transportation of oil ; special problems on treatment of crude oil emul-
sions, conservation of oil, gas, etc.

2x4. Gurley, W. & L. E., 514 Fulton St., Troy, N. Y. (Instruments
for civil, mining and hydraulic engineers, and land surveyors.)

Research staff: E. W. Arms, 3 engineers, 3 mechanicians and as-
sistants as needed.

Research work : Practically full time of 7 on investigations for de-
sign and manufacture of instruments for civil, mining and hydraulic
engrineers, such as automatic water stage registers, current meters,
hook gages, transits and levels.

Equipment: For testing and calibrating standard precision meas-
ures of weight, capacity and length ; for investigation of water meas-
urements and for design of instruments for this purpose; automatic
water stage registers, current meters and hook gages; special divid-
ing engfines for accurate angular and linear graduation ; for drawing
platinum wire from o.ooi- to 0.00002-inch diameter for cross-wire reti-
cles and in research experiments.

INDUSTRIAL RESEARCH LABORATORIES 39

2x5. Habirshaw Electric Cable Company, Inc., Yonkers, N. Y.

Research staff: William A. Del Mar, 3 to 6 engineers, 2 to 6 chem-
ists and o to 7 assistants.

Research work: Seven-tenths time of staff on insulating materials
and electric cable manufacture.

Equipment : Miniature manufacturing plant for making rubber in-
sulated wire in the laboratory.
2x6. Hamersley MTg Co., The, Garfield, N. J. (Waxed papers.)

Research staff: i chemical engineer and 5 chemists.

Research work : One-third time of 6 on pulp, paper, and paper mill
chemicals.

Equipment: Well equipped for paper mill experiments on semi-
commercial scale.

2x7. Harbison-Walker Refractories Company, Farmers Bank Build-
ing, Pittsburgh, Pa. (Fire-clay, silica, magnesite and chrome bricks
and other refractory products.)

Research staff: R. H. Youngman, i to 2 special technical men, i
chief chemist and i or 2 chemists.

Research work: One-half time of staff on problems in connection
with refractories.

Equipment : i coal and i gas-fired test kiln, i small ore crusher, 2
Braun planetary pulverizers and'i hydraulic press of 104 tons capacity.
218. Harrison Mfg. Co., The, 55 Union St., Rahway, N. J. (General
chemicals and chemical products ; thorium nitrate and other rare earth
salts and oxides; writing inks.)

Research staff: C. W. Squier.

Research work: Full time of i on general lines of research.

Harrison Safety Boiler Works. See Cochrane, H. S. B. W.,
Corporation (p. 20).

2xg. Hayes, Hammond V., 84 State St., Boston, Mass. (Consulting
engfineer.)

Research staff: Hammond V. Hayes, 5 electrical engineers and
physicists.

Research work : Full time of 6 on electro-dynamic problems.

Hajmes Stellite Co. See Union Carbide and Carbon Research
Laboratories, Inc. (p. 78).

220. Heap, William, & Sons, Grand Haven, Mich. (Celluloid and
china.)

Research staff: H. Stirling Snell and i chemist.

Research work : Three-fourths time of 2 on thermoplastics.

221. Heinrich Laboratories of Applied Chemistry, looi Oxford St.,
Berkeley, Calif, (formerly Tacoma, Wash.).

Research staff : E. O. Heinrich and i chemist.

Research work : Full time of 2 on chemical and photomicrographi-
cal problems as applied to criminal investigation.
222-224. Hercules Powder Co., Wilmington, Del. (Explosives.)
Laboratories at Kenvil, N. J., Brunswick, Ga., and Emporium, Pa.
Executive staff, consisting of G. M. Norman and 6 assistants, super-
vises work on problems on explosives, mineral acids, nitrogen fixa-
tion, pyroxylin solutions, plastics, smokeless powder, and naval stores
at three research laboratories.

222. Experimental station, Kenvil, N. J,

40 INDUSTRIAL RESEARCH LABORATORIES

Research staff: C. 1^. Bierbauer, i6 graduate chemists and engi-
neers, 7 other salaried employees and 24 payroll employees.

Research work : Approximately full time of 48 on research on high
explosives, smokeless powders, plastics, pyrox;y^lin solutions, and naval
stores. Some time also devoted to investigations of analytical meth-
ods in reference to above.

Equipment: Complete equipment for testing properties of dyna-
mite. Equipment for manufacturing propellant powders, and ranges
for testing same for velocity and pressure, semi-works equipment for
the manufacture of organic chemicals and plastics.

223. Naval Stores Laboratory, Brunswick, Ga.

Research staff: C. M. Sherwood and 3 graduate chemists.

Research work : Seven-tenths time of 4 on problems connected with
the manufacture of turpentine, rosins and pine oil, by the steam sol-
vent process.

Equipment: Semi- works scale apparatus duplicates plant process.

224. Emporium Research Laboratory, Emporium, Pa.

Research staff: R. B. Smith and i assistant chief chemist, 5 as-
sistant chemists and 2 laboratory assistants.

Research work : Full time of 10 on general research and on meth-
ods of manufacture of mineral acids and domestic explosives.

Equipment : Semi-commercial scale apparatus for nitration ; special
equipment for analysis of explosives and for explosive testing.

Hes8-Bright Manufacturing Co. See S. K. F. Industries, Inc.

(p. 72).

335. Heyden Chemical Company of America, Inc., Garfield, N. J.

Research staff: Robert O. Bengis, 7 chemists, i engineer and i
laboratory assistant.

Research work: Two-fifths time of 10 on medicinal and pharma-
ceutical chemistry ; salicylates and metallic colloids.

Equipment : Equipped semi-commercial plant adjacent to research
laboratory.

226. Hirsch Laboratories, Inc., The, 50 E. 41st St., New York, N. Y.
Laboratory at 593 Irving Ave., Brooklyn, N. Y.

Research staff: Alcan Hirsch and 5 chemists.

Research work : One-half time of 6 on organic chemicals, interme-
diates, dyestuffs and pharmaceuticals. Metal products; cerium,
thorium and molybdenum products.

Equipment: Fully equipped for semi-plant operations. Facilities
for duplicating and testing on commercial scale any proposed plant in-
stallation or process.

Hirsch, Stein & Company. See United Chemical and Organic
Products Co. (p. 79).

227. Hochstadter Laboratories, 227 Front St., New York, N. Y.
(General chemical analyses and investigations; consultants and tech-
nical experts.)

Research staff: Irving Hochstadter, W. B. Stoddard and 2 chemists.

Research work: One-half time of 4 on manufacture and prepara-
tion of food and pharmaceutical products with special emphasis on
problems relating to pure food regulations and on problems relating
to the rare metals, especially "Tungsten" compounds.

INDUSTRIAL RESEARCH LABORATORIES 41

aa8. Holt Manufacturing Company, The, Peoria, 111. (Tractors.)

Research staff: R. M. Hudson, research engineer, and 2 mechani-
cal engineers; F. W. Grotts, inspection and metallurgical engineer,
and 2 chemists; i expert in microphotography.

Research work : Full time of supervisors and staff on technical, in-
dustrial and commercial problems. Industrial research on wage sur-
veys, costs of living, industrial relations and organization problems
and principles.

Equipment: Special microphotographic apparatus with grinding
and polishing machines; oil distillation apparatus and viscosimeter ;
electric furnace for experimental heat treating; dynamometer for
motor research.
aag. Hood Rubber Company, Watertown, Mass.

Research staff: Warren E. Clancy, 2 chemists and several routine
assistants.

Research work: Small part time of staff on new methods of ex-
amination of materials ; study of various organic derivatives.

Equipment : Devices and machines for testing rubber, cloth, yarns ;
large experimental mill room equipped with heavier machinery and
heavier testing machines for testing tires (solid, pneumatic, etc.).

330. Hooker Electrochemical Company, Niagara Falls, N. Y.
Research staff: T. L. B. Lyster, director of development, A. H.

Hooker, technical director, W. J. Marsh, research chemist, i research
chemist and 4 assistants.

Research work: Full time of 5 on development of new processes
and betterment of present processes.

Equipment: Furnace room, annex and industrial laboratory
equipped for intermediate scale or development work.

331. Hoskins Manufacturing Company, Lawton Ave., at Buchanan,
Detroit, Mich. (Electric-furnaces, pyrometers and heating ap-
pliances.)

Research staff : W. A. Gatward and 4 engineers.
Research work : Almost full time of 5 on the improvement and pro-
duction of alloys and allied products.

332. Houghton, E. F. & Co., 240 W. Somerset St., Philadelphia, Pa.
(Oils, mechanical leathers and steel heat treating materials.)

Research staff: George W. Pressell, 9 chemists and engineers.
Consulting engineers and chemists sometimes employed.

Research work : Research staff is working constantly in producing
oils for the industries, mechanical leathers and steel treating materials ;
also improving methods in the manufacturing industry.
233. Howard Wheat and Flour Testing Laboratory, The, Old Colony
Building, Minneapolis, Minn. (Comparative baking tests, records and
reports, milling tests, chemical and microscopical analyses.)

Research staff : C. H. Briggs and 3 chemists.

Research work : Small part time of 4 on problems connected with
causes of peculiar variations of wheats and other cereals when baked
into bread or used for other food purposes ; efforts to improve methods
of separation of wheat proteins; improved methods of quantitative
analysis; chemical causes of loaf expansion and effects of various

42 INDUSTRIAL RESEARCH LABORATORIES

activating materials in bread making, carried out by cooperation of
baking and chemical departments. Some work on distinguishing
cereal flours one from another.

Equipment: Moisture testers for grain, haemocytometer, yeast
testing apparatus of special design, wheat and grain cleaning and
milling department and a baking test department, equipped for hand-
ling more than loo individual tests daily with automatic control of
kneading machines, bread raising cabinets, etc.

234« Hunt, Robert W.» and Co.» 175 W. tackson Blvd., Chicago, 111.
(Engfineers.) Laboratories also at 251 Kearney St., San Francisco,
Calif. ; 90 West St., New York, N. Y. ; Monongahela Bank Bldg., Pitts-
burgh, Pa. ; 905 McGill Bldg., Montreal, Canada ; and Syndicate Trust
Bldg., St. Louis, Mo.

Research staff: J. H. Campbell and assistant, 12 chemists and 8
engineers.

Research work : Part time of 22 on materials of construction, iron,
steel, stone, cement and bitumen.

235. Hyco Fuel Products Corporation, 30 Broad St., New York, N. Y.
Laboratory at Edgewater, N. J.

Research staff: Allen Rogers, 3 chemists, i engineer, i draftsman
and 5 assistants.

Research work : The plant is built for demonstration and research
on oil problems, especially as related to motor fuel.

236. Hynson, Westcott & Dunning, 423 N. Charles St., Baltimore,
Md. (Bacterial and bio-chemic therapeutic products.) Laboratory
at 16 E. Hamilton St, Baltimore, Md.

Research staff: Daniel Base, 2 chemists, i pharmacologist and i
bacteriologfist.

Research work: One-half time of 5 on preparation and pharma-
cological testing of new drugs.

237. Imperial Belting Company, Lincoln and Kinzie Sts., Chicago, 111.
(Belting and conveyors.) Laboratory at 400 N. Lincoln St., Chicago,
111.

Research staff: James A. Millner, i engineer and 2 chemists.
Research work: Approximately one-half time of 4 on oils, paints,
asphalts and textiles.

238. Industrial Chemical Institute of Milwaukee, 200 Pleasant St,
Milwaukee, Wis. (Consultants for chemical and engineering prob-
lems.)

Research staff : F. M. Dupont, i chemical engineer, i bacteriologist
and 4 chemists.

Research work: Full time of i chemist on food, beverage, mag-
nesite, lime, adhesives, antisepticides and general matters.

239. Industrial Research Corporation, 1025 Front St., Toledo, Ohio.
Research staff : C. P. Brockway and 2 engineers.

Research work: Full time of 3 on problems related to small
machine equipment and small devices in metal.

240. Industrud Research Laboratories, 190 N. State St., Chicago, 111.
F. Peter Dengler, Inc., proprietors. (General consulting and research
chemists, resource and chemical engineers.)

Research staff : F. Peter Dengler, 5 or 6 chemists and i engineer.

i

INDUSTRIAL RESEARCH LABORATORIES 43

Research work: Full time of staff on manufacturing and research
problems relative to cement, coal, corn, cotton seed, drugs, dairy, dyes,
foods, minerals, paints, paving, petroleum, paper, sewage, soap, steel,
sugar, tobacco, water, barley, conservation of waste material and
manufacture of non-alcoholic flavoring extracts including vanilla.

Equipment : Commercial equipment for production of coke and by-
products and for decolorizing and reclaiming cloth, mill ends, flour
bags, sugar bags, all cloth signs and rubber-coated textile materials.
Commercial equipment for extracting vegetable alkaloid from tea and
coflFee. Apparatus for the manufacture of non-alcoholic extracts is
being installed on a commercial scale for immediate use.

241. Industrial Testing Laboratories, 402 West 23rd St., New York,
N.Y.

Research staff: Emil Schlichting, director, H. Winther, chief
chemist, and 5 assistant chemists.

Research work : Part time of staff on problems related to beverage,
fermentation and food industries.

Equipment: For chemical, biological and microscopical analyses
of beverages and foods, their raw materials, by-products, and acces-
sories.

242. Industrial Works, Bay City, Mich.

Research staflF : R. H. Morgan, metallurgist, J. C. Wheat, develop-
ment engineer; chemists and assistants as required.

Research work: Heat treatment and properties of metals, proper-
ties of other materials, development and control of foundry practice
for iron, steel and bronze ; welding practice. Development of cranes
and accessories to meet needs of users ; statistical manufacturing and
executive control; standards of production and personnel, standard
times and routings.

Equipment: 150,000-pound Riehle testing machine. Shore sclero-
scope. Berry strain gauges, two proof testing machines of 500,000
pounds and 100,000 pounds capacity, for testing actual parts before
assembly.

243. Ingersoll-Rand Company, 11 Broadway, New York, N. Y.
(Rock drills, etc.)

Research staff: F. W. O'Neil and a number of assistants.

Research work : Full time of some and part time of others on drills,
pumps, pneumatic tools, compressors, blowers, condensers and oil
engines.

244. Inland Steel Company, Indiana Harbor, Ind.,

Research staff: J. C. Dickson, 29 chemists and 5 chemical
engineers.

Research work: Full time of 4 and part time of 30 on problems
connected with steel industry.

245. Institute of Industrial Research, The, 19th and B Sts., N. W.,
Washington, D. C.

Research staff: Allerton S. Cushman, chemists, physicists and as-
sistants as needed.

Research work : Varying part time of staff on physical testing of
cements, rocks, clays, brick, block, iron, steel, wood, rubber, and other
materials of construction. In Bitumen Laboratory petroleum and

44 INDUSTRIAL RESEARCH LABORATORIES

petroleum products, tars and tar products, creosoting oils, asphalts,
bituminous emulsions, bituminous aggregates, and all other types of
chemical road and paving materials, roofing materials, rubber, etc., are
examined and tested. Chemical examinations of rocks, clays, cements,
etc., are made and researches conducted on improvements in industrial
products and processes and utilization of waste products for road pur-
poses.
Equipment : For cement and bitumen.

346. International Filter Co., 38 S. Dearborn St., Chicago, 111. (Water
softening and filtration plants.) Laboratory at 333 W. 25th Place,
Chicago, 111.

Research staff : 2 to 5 workers.

Research work: Approximately one-half time on materials, meth-
ods and processes for purifying liquids.

347. International Nickel Company, The, Bayonne, N .J. Successors
to The Orford Copper Co.

Research staflF: raul D. Merica, 2 metallurgists, 2 assistant metal-
lurgists, 2 chemists, 2 laboratory assistants, and i machinist.

Research work: Metallurgy of copper and nickel; physical prop-
erties of nickel and Monel metal; uses of Monel metal and nickel
alloys.

Equipment : Laboratory electric furnace equipment ; dust and fume
sampling apparatus ; experimental electroplating plant.
948. International Shoe Co. (Burke Tannery), Morganton, N. C.

Research staff : J. S. Rogers, 2 trained assistants, and i helper.

Research work : Approximately one-half time of director and part
time of assistants on problems in extraction of tanning materials and
the tanning and finishing of sole leathers.

Equipment: Some special apparatus for small scale plant experi-
ments.

249. International Silver Company, Meriden, Conn.
Research staff : Chas. E. Skidgell and 2 chemists.
Research work : Small part time of 3 on electro-plating.

250. Interocean Oil Company, The, East Brooklyn, Baltimore, Md.
Research staff : H. R. Gundlach, 2 chemists and 6 assistants.
Research work : Approximately one-tenth time of 9 on development

of refining methods and testing ; recovering of waste products, etc.

Equipment: Laboratory scale refinery, also larger scale experi-
mental plant.
351. James Ore Concentrator Co., 35 Runyon St., Newark, N. J.

Research staff: U. S. James, i metallurgical engineer and 3 assist-
ants.

Research work : Full time of 3 on ore and coal testing.
353. Jaques Manufacturing Company, i6th and Canal Sts., Chicago,
111. (Manufacturers of K. C. baking powder.)

Research staff : J. R. Chittick and 3 chemists.

Research work : One-third time of 2 on leavening materials.

Jeffrey-Dewitt Co. See Champion Porcelain Company (p. 19).
353. Johnson & Johnson, New Brunswick, N. J. (Surgical supplies.)

Research staff : Fred B. Kilmer and 7 assistants.

Research work : One-third time of 8 in research on medical, surgical

INDUSTRIAL RESEARCH LABORATORIES 45

and hospital supplies (not equipment) and incidentally drugs and
commodities used therein.

254. Kalmus, Comstock & Wescott, Inc.» no Brookline Ave., Boston,
Mass. (Consulting, research and operating engineers.)

Research staff: A group of physicists, chemists, metallurgists and
chemical engineers of from 20 to 25 in number, directed by Herbert T.
Kalmus, Daniel F. Comstock and E. W. Westcott.

Research work : Full time of staff on mechanical, physical, chem-
ical, electrochemical, metallurgfical and photographic lines leading to
the development of processes, use of waste products, and through the
designing, constructing and early operating of plants.

Equipment : Specially designed equipment in the fields of ceramics,
abrasives, general chemical engfineering, metallurgy, photography,
motion pictures, and vegetable oils.

255. Kellogg Switchboard and Supply Co.» Adams and Aberdeen Sts.,
Chicago, 111.

Research staff: Wilbur J. Anglemyer, i electrical engineer and 5
assistants.

Research work: Full time of 6 on testing of materials including
analysis, tensile strength tests and magnetic characteristics ; checking
methods of manufacture and development of special testing instru-
ments and new products.

Equipment : Impregnating apparatus, 5000 and 50,000 volt testing
transformers, Burrous permeameter, Rowland dynamometer, G. E.
Co. oscillograph, centrifugal extractor, apparatus for testing textile
materials and paper and insulation testing equipment.

256. Keuffel & Esser Co., Hoboken, N. J. (Drawing materials and
mathematical and surveying instruments.)

Research staff: Carl Keuffel, i chemist, 2 assistant chemists, 2
optical engineers, and 2 assistants.

Research work: One-half time of 8 on optical glass and various
articles manufactured, including design of optical instruments and
calculation of optical systems.

Equipment : Special equipment for testing presence of small quan-
tities of iron in silicates, and for physical, chemical and microscopic
testing of papers. Optical laboratory equipped for general testing of
optical instruments.

257. Kidde, Walter, & Company, Incorporated, 140 Cedar St., New
York, N. Y. (Engineers and constructors.)

Research staff : Barzillai G. Worth and assistants as necessary.

Research work : Investigation for clients, such as electrolysis of
potassium and sodium compounds ; electrochemical extraction of oils ;
chemical salvage systems for tanneries ; sanitation of tannery effluent,
etc.

258. Kilbourne ft Clark Manufacturing Company, Seattle, Wash.
(Engineers and manufacturers of electrical and radio apparatus.)

Research staff: H. F. Jefferson and 5 men.

Research work : Time of staff as occasion requires, on testing and
investigating high-frequency circuits.

Equipment: Wave-meters, decremeters, sphere spark gap (25 cm.
sphere) for high voltage tests; condensers, variable and fixed, with
air, mica and oil dielectrics ; inductances in various forms for high and

46 INDUSTRIAL RESEARCH LABORATORIES

low voltage; 500-cycle meters for use in connection with audio-
frequency circuits in radio work.

Kistler, Lesh & Company. See International Shoe Co. (p. 44).

259. Klearflax Linen Rug Company, 63rd and Grand Aves., West,
Duluth, Minn. (Linen rugs and carpeting.)

Research staff : Charles F. Goldthwait and variable number of as-
sistants.

Research work : Full time of staff on use of flax fibre and its by-
products ; humidity, textiles and mechanism of dyeing process.

260. Kokomo Steel and Wire Co., Kokomo, Ind.
Research staff : R. K. Clifford, 2 chemists and 2 assistants.
Research work: One-third time of 5 on standardization of raw

materials, specifications and improvement of products in connection
with manufacture of open hearth steel, wire and wire products.

Equipment : 100,000-pound Olsen testing machine, Brinell machine,
electric furnace for heat treatments, metallographic equipment for
grinding, polishing and microphotography.

261. Kolynos Co., The, New Haven, Conn. (Dental cream.)
Research staff : L. A. Jenkins, 3 chemists and 2 bacteriologists.
Research work : One-half time of 6 on oral hygiene.

262. Koppers Company, The, Pittsburgh, Pa. (Designers and
builders of by-product coke and gas plants and apparatus for benzol
recovery, tar distillation and gas purification.)

Research staff: F. W. Sperr, Jr., 13 graduate chemists, i engineer
and 4 assistants.

Research work : Full time of I9 on coal carbonization, gas produc-
tion, and purification, by-product recovery, secondary treatment of
various by-products, general fuel research, refractories, pyrometry, in-
vestigation of coal properties.

Equipment: Special apparatus for coal carbonization at high and
low temperatures, coal washing, coke research, gas purification by
dry and liquid processes, furnaces for investigation of refractory ma-
terials at high temperatures, laboratories and experimental plant
fully equipped for semi-commercial tests, and plants available for
large scale tests in relation to coke and gas manufacture and by-prod-
uct recovery.

263. Kraus Research Laboratories, Inc., 130 Pearl St., New York,
N. Y. (Consulting engineers in refractories.)

Research staff : Charles E. Kraus, 2 ceramists, 2 research engineers
and 2 assistants.

Research work: Three-fourths time of 7 on ceramics and refrac-
tories.

Equipment: Equipped to make all standard tests on refractory
materials, both in raw and finished state.

264. Krebs Pigment and Chemical Co., The, Newport, Del.
Research staff : H. W. Fox, i chemical engineer, 2 chemists and 2

assistants.

Research work : Full time of 6 on properties of lithopone ; efficiency
of steps of process.

265. Kulhnan, Salz & Co., 603 Wells Fargo Building, San Francisco,
Calif. (Tanners.)

INDUSTRIAL RESEARCH LABORATORIES 47

Research staff: i chemist and i helper.

Research work: Variable amount of time of 2 on science of tan-
ning.

266. Laclede-Christy Clay Products Company, 4600 S. Kingshighway,
St. Louis, Mo.

Research staff: C. W. Berry and i assistant.

Research work : One-half time of 2 on development of refractories,
superior clays for use in paper, graphite crucibles, enamels; unusual
basic and neutral refractories, such as high aluminous materials, com-
binations of alumina and magnesia.

267. Lakeview Laboratories, 2 Jersey St., Buffalo, N. Y.
Research staff: A. L. Stevens and 2 assistants.

Research work : Four-fifths time of 3 on wood oils and tars.
a6S. Larkin Co., 680 Seneca St., Buffalo, N. Y. (Soap.)

Research staff : L. F. Hoyt, 4 chemists and 2 assistants.

Research work: Three-fourths time of 7 on soaps, fats and oils;
development along miscellaneous lines of new products for the com-
pany.

Equipment: Small experimental plant for producing soap.
269. Laucks, L F., Inc., 99 Marion St., Seattle, Wash. (Analytical
and consulting chemists, assayers and metallurgists.)

Research staff : L F. Laucks and H. P. Banks, 3 chemists, 3 chemi-
cal engineers, 2 agronomists and 3 inspection engineers.

Research work: One-fourth time of 8 on uses for raw materials
available in the Orient and adaptation of these materials to American
requirements; development of improvements in manufacturers' proc-
esses and development of coal by-products.

Equipment : Complete vegetable and fish oil refinery and complete
coal by-products plant.
370. Lee & Wight, 113 E. Franklin St., Baltimore, Md.

Research staff: Two chemists.

Research work: Part time of 2 on industrial and miscellaneous
problems.

271. Leeds & Northrup Company, 4901 Stenton Ave., Philadelphia,
Pa.

Research staff: Irving 6. Smith, 7 trained research workers and 3
mechanicians.

Research work: Full time of 11 on development of apparatus for
precise measurements in heat, electricity, magnetism and heat treat-
ment of steel; also for research and control in chemical industries.

Equipment : Apparatus for heat treatment of steel, instruments for
precise measurements in heat, electricity and n^agnetism.

372. Lehn & Fink, Inc., 192 Bloomfield Ave., Bloomfield, N. J. (An-
tiseptics, disinfectants, drugs, medicines, dentifrice, soaps, fine chemi-
cals.)

Research staff : C. Hinck, 14 chemists and 2 engfineers.
Research work: Full time of 17 on organic, biological and pharma-
ceutical problems.

373. Lemoine, Pierre, Cic., Inc., 294 Pearl St., New York, N. Y. (Es-
sential oils, aromatic chemicals.) Factory Laboratory at L. I. City,
N.Y.

AS INDUSTRIAL RESEARCH LABORATORIES

Research staff: 2 chemical engineers, i analytical and research
chemist and 2 associate chemists.

Research work : Part time of 5 on synthetic organic chemicals and
essential oils, perfumery oils and raw materials and flavors and flavor-
ing raw materials.

274. Lennox Chemical Co., The, 1205 E. 55th St., Cleveland, Ohio.
Laboratory at Euclid, Ohio.

Research staff : A. S. Allen and 2 assistants.

Research work : One-half time of 3 on carbonation as related to the
soft drink or beverage industry ; liquefaction, purification and drying
of commercial gases as oxygen, nitrous oxide, and carbon dioxide.

275. Lewis, F. J., Manufacturing Co., 2513 S. Robey St., Chicago,
111.

Research staff: W. 6. Murphy and 2 chemical engineers.
Research work : Part time on coal tar products.

Lewis, Gilman & Moore. See Metals & Chemicals Extraction
Corporation (p. 52).

276. Lilly, Eli, and Company, Indianapolis, Ind. (Pharmaceutical
and biologfical products.)

Research staff: G. H. A. Clowes, Frank R. Eldred, A. L. Walters
and about 40 chemists and pharmacologists.

Research work: Full time of 8 men and half time of 17 directed to
development of new therapeutic agents and to broad stud^ of mode
of action of drugs from physical, chemical and physiological stand-
points.

277. Lincoln, E. S., Inc., 534 Congress St., Portland, Me. (Consult-
ing engineers ; electrical laboratories.)

Research staff: E. S. Lincoln and 3 engineers.

Research work : Full time of 4 on electrical problems. Field work
a specialty.

Linde Air Products Company, The. See Union Carbide and
Carbon Research Laboratories, Inc. (p. 78).

278. Lindsay Light Company, 161 E. Grand Ave., Chicago, 111.

Research staff : H. N. McCoy, 8 chemists and i engineer.

Research work: Four-fifths time of 10 on improvements of proc-
esses of refining thorium nitrate, cerium compounds, organic prepara-
tions such as phenolphthalein and vanillin, preparation of dyes.

279. Little, Arthur D., Inc., 30 Charles River Road, Cambridge 39,
Mass. (Chemists, engineers, managers.)

Research staff: Earl P. Stevenson, director, and 8 research chem-
ists cooperating with 10 analytical chemists; 8 engineers, chemical,
mechanical, mining; i economic geologist; and special staff for valua-
tions and appraisals.

Research work : Full time of 10 on industrial research on lines de-
termined by requirements of clients and on special problems in adhe-
sives, ceramics, utilization of lumbering waste, paper and pulp, tex-
tiles, metallurgy, non-metallic minerals, and process developments.

Equipment : Complete experimental paper mill including a 30-inch
Fourdrinier machine. Semi-commercial equipment for miscellaneous
work.

INDUSTRIAL RESEARCH LABORATORIES 49

a8o. Littlefield Laboratories Co., Seattle, Wash.

Research staff: E. E. Littlefield, i electrochemist and electro-
physicist, I chemist and i mechanical engineer.

Research work : Full time of i and part time of 3 in chemical, elec-
trical and electrochemical fields ; development of special apparatus for
initiating and stopping flow of liquids by varying conductivity ; elec-
trical treatment of vegetation. Usually done in connection with large
industries in the United States and England.
aSz. Lockhart Laboratories, 33^ Auburn Ave., Atlanta, Ga.

Research staff: L. B. Lockhart.

Research work : Full time of i on lubricating oils and greases, spe-
cial soaps, varnishes, waterproofing, petroleum products, colloids and
emulsions.

aSa. Long, W. H., & Co., Inc., 244 Canal St., New York, N. Y.
(Wholesale druggists.)

Research staff: Charles H. Lewis, 2 chemists, i assistant and i
laboratory assistant.

Research work: Drugs, chemicals and dyes.

283. Ludlum Steel Company, Watervliet, N. Y.
Research staff: P. A. E. Armstrong and 4 trained men.
Research work: Full time of 5 on improvement of manufacturing

methods for ferro alloys and certain steels, such as magnet steel and
non-corrosive steels and methods of chemical analysis of steels and
ferro alloys.

284. Lumen Bearing Company^ Buffalo, N. Y. (Brass and bronze
foundry.)

Research staff: C. H. Bierbaum, metallurgist; B. Woiski, chief
chemist, and 2 assistants, and G. F. Comstock, consulting metallurgist.

Research work : Varying portion of time on problems having to do
with non-ferrous metallurgy and metallography, chemistry as applied
to non-ferrous metals, photomicrography of the non-ferrous metals.

Equipment : 50,000-pound Olsen universal testing machine, Brinell
hardness machine, scleroscope, microcharacter.

285. Lunkenbeimer Co., The, Cincinnati, Ohio. (Valves, pipe fit-
tings and other metal specialties.)

Research staff: George K. Elliott and 7 assistants.

Research work : Two-fifths time of 8 on metallurgical problems and
corrosion. Generation and handling of saturated and super-heated
steam; application of arc electric-furnace to production of malleable
cast iron, special gray irons, and other high-carbon iron alloys.

286. Ljrster Chemical Company, Inc., 61 Broadway, New York, N. Y.
Laboratory at Passaic Junction, N. J.

Research staff: William R. Lamar and 2 chemists.

Research work : Full time of 3 on utilization of former waste prod-
ucts in the rectification of wood tar oils for creosote and guaiacol;
organic compounds and photographic developers and perfumery chem-
icals.

287. Maas, A. R., Chemical Company, 308 E. 8th St., Los Angeles,
Calif.

Research staff: Arthur R. Maas, 3 analytical chemists, i research
chemist and i chemical engineer.

so INDUSTRIAL RESEARCH LABORATORIES

Research work: Manufacture of sulphites and other products,
chiefly those derived from alkali and sulfur dioxide.

Equipment: Absorption towers.
388. HacAndrews & Forbes Company, 3d St. and Jefferson Ave.,
Camden, N. J. (Licorice extract, natural dyestuffs, wallboard and
Foamite fire extin^ishers.)

Research staff: Fercy A. Houseman, 6 chemists and 3 helpers.

Research work : Approximately one-half time of 7 on constituents
of licorice root and extract and development of Foamite fire extin-
guishers.

Equipment: Copper extractors, percolators and vacuum pans of
laboratory size and semi-commercial size.

289. HaUinckrodt Chemical Works, St. Louis, Mo. (Chemicals for
medicinal, photographic, analytical and technical purposes.)

Research staff: W. N. Stull, 22 chemists, 2 chemical engineers and
I safety engineer.

Research work : Full time of 5, one-half time of 4 and part time of
others on improvement in processes of manufacture and methods of
analysis.

ago. Manhattan Rubber Mfg. Co., The, Passaic, N. J. (Mechanical
rubber goods.)

Research staff: W. L. Sturtevant, 6 chemists and 6 laboratory as-
sistants.

Research work : One-fourth time of 13 on rubber compounding and
vulcanization.

agz. Martin, Glen L., Company, The, 16800 St. Clair Ave., Cleveland,
Ohio. (Builders of airplanes.)

Research staff: Lessiter C. Milburn, i metallurgical engineer and
I chemist.

Research work : One-third time of 3 on new aircraft materials and
check of aircraft designs, aircraft performance tests, and general air-
craft development, metal construction, etc.

Equipment: Rib testing machine (transverse loading distributed
according to any pre-determined ratio). Combined pendulum tension
machine and impact test machine, with interchangeable hammers
(pendulums) and two ranges of capacity (200 and 1000 pounds).

292. Martinez Refinery, Shell Co. of California, Martinez, Calif.
Research staff: A. W. Jurrissen and 2 chemists.

Research work : Varying portion time of 3 on treatment and pro-
duction of petroleum products.

Equipment : Large scale cracking apparatus and treating plant.
Marvin-Davis Laboratories, Incorporated. See National Biscuit
Company (p. 55).

293. Matfaieson Alkali Works (Inc.), The, Niagara Falls, N. Y.
Research staff : R. E. Gegenheimer, 7 chemists and 4 assistants.
Research work: Full time of 6 on new process development and

investigation of problems of electrolytic chlorin and caustic plant
operation.

294. May Chemical Works, 204 Niagara St., Newark, N. J.
Research staff: Otto B. May and 2 assistants.

Research work : One-half time of 3 on azo-dyes and intermediates.

INDUSTRIAL RESEARCH LABORATORIES 51

395* Masmard, T. Poole, Atlanta, Ga. (Geological and industrial en-
gineering.)

Research staff: T. Poole Maynard, i chemical engineer, i mining
engineer and i civil engineer.

Research work: One-third time on clays, bauxites, fullers earth,
refractories, textiles, oil-cloth; recovery of potash from silicates, etc.
296. M. B. Chemical Co., Inc., Johnson City, Tenn.

Research staff : A. J. Buchanan and 2 chemists.

Research work: Large part time of i chemist on dyes and inter-
mediates.

397. Mcllhiney, Parker C, 50 E. 41st St., New York, N. Y.
Research staff: Parker C. Mcllhiney and 2 chemists.

Research work: One-half time of 3 on investigation of paints and
varnishes, hydrogenation processes, electrolytic processes, wood dis-
tillation processes, shellac and other resins and fats and oils.

398. McKesson & Robbins, Incorporated, 55 Berry St., Brooklyn,
N. Y. (Drugs and chemicals.) Laboratory at 97 Fulton St., New
York, N. Y.

Research staff: E. H. Gane and 2 pharmaceutical chemists.

Research work: Approximately one-half time of 3 on active prin-
ciples of vegetable drugs, new medicinal compounds and drug stand-
ards.

299. McLaughlin Gormley King Co., 1715 Fifth St., S. E., Minne-
apolis, Minn. (Drugs and herbs.)

Research staff: C. B. Gnadinger and 2 chemists.
Research work: Approximately one-half time of 2 on food prod-
ucts, crude drugs and insecticides.

300. McNab & Harlin Manufacturing Co., 55 John St., New York,
N. Y. (Valves, fittings, etc.) Laboratory at 440 Straight St., Pater-
son, N. J.

Research staff: Ernest G. Jarvis, i assistant, 5 chemists, 6 metal-
lurgists and 8 engineers.

Research work: Approximately one-half time of 21 on rare metals
and their uses in industrial alloys.

Equipment : Electric laboratory melting furnaces, Hoskins type F.
C. 106, miniature rolling mills and all necessary physical testing ma-
chines and equipment for testing sheets, rods, wire and castings, and
fully equipped metallographic department.

30X. Meigs, Bassett & Slaughter, Inc., 210 S. 13th St., Philadelphia,
Pa. (Chemical engineers.) Laboratory at Bala, Pa.

Research staff : Harry P. Bassett, i chemical engineer and 3 chem-
ists.

Research work : Full time of 5 on paper, paper pulp, plastics, cellu-
lose products, alkali and alkali salts.
30a. Merck & Co., 45 Park Place, New York, N. Y. (Chemists.)

Research staff : 2 trained chemists.

Research work: Full time of 2 on problems incident to manufac-
ture of the company's products.

Equipment: Standard equipment for research in connection with
manufacture of medicinal, analytical, photographic and technical
chemicals.

52 INDUSTRIAL RESEARCH LABORATORIES

m

303. Herrell. Wm. S., Company, The, 5th, Pike and Butler Sts., Cin-
cinnati, Ohio. (Manufacturing pharmacists.)

Research staff : 7 chemists, i chemical engineer and 4 pharmacists.

Research work: Approximately full time of 3 and part time of 2
on problems of manufacturing pharmaceuticals and pharmaceutical
products.

304. Merrell-Soule Laboratory, Ssrracuse, N. Y.

Research staff: R. S. Fleming, 2 chemists and i assistant. An
engineering department which does much work which might be classi-
fied as research.

Research work : Half time of 3 on food problems.

Equipment: Experimental drying plant.

305. Herrimac Chemical Company, North Woburn, Mass.
Research staff : Lester A. Pratt and 9 chemists.

Research work: Full time of staff on inorganic and organic re-
search problems.

Equipment: Industrial laboratory for carrying on large scale ex-
periments.

306. Hesabi Iron Company, Babbitt, Minn.

Research staff: W. G. Swart, 3 engineers, 2 metallurgists and i
chemist.

Research work : One-half time of 7 on magnetic separation of ores
and sintering.

Equipment: Magnetic cobbers, classifiers and log washers and
demagnetizers.

307. MetaUoth Co., N. Y., Susq. & Western R. R. and Garibaldi Ave.,
Lodi, N. J.

Research staff: Herbert B. Fenn.

Research work : Part time of i on mildewproofing, fireproofing and
waterproofing of cotton, flax and jute fabrics.

Equipment: Apparatus for processing materials under conditions
of actual commercial production.

308. Metals & Chemicals Extraction Corporation, 1014 Hobart Bldg.,
San Francisco, Calif. (Heavy chemicals.)

Research staff: L. H. Duschak and i chemical engineer.
Research work : Inorganic chemistry, including the manufacture of
heavy chemicals, potash, borax, barium compounds and acids.

309. Metz, H. A., Laboratories, Inc., 122 Hudson St., New York,
N. Y. Plant and laboratories, 642 Pacific St., Brooklyn, N. Y.

Research staff : A. E. Sherndal, C. N. Myers, C. W. Hooper, G. P.
Metz and 4 chemists.

Research work : Studies of chemical, pharmaceutical and medicinal
products; technical problems involved in their manufacture; path-
ological, biological and bacteriological investigations relative to their

use.

310. Meyer, Theodore, 213 S. loth St., Philadelphia, Pa.

Research staff: John K. Montgomery and 2 assistants.

Research work: One-fourth time of 3 on antiseptics and insecti-
cides.

Midvale Steel Company, The. See Mid vale Steel and Ordnance
Company.

INDUSTRIAL RESEARCH LABORATORIES 53

SIX. Midvale Steel and Ordnance Company, Nicetown Works, Phila-
delphia, Pa.

Research staff : A. H. Miller and 7 men.

Research work : One-half time of 8 on investigation of characteris-
tics of iron alloys, such as equilibrium diagrams, physical and mag-
netic qualities, etc.; also the investigation of new alloys of steel for
use in high service purposes.

Equipment: Apparatus for several methods of obtaining critical
temperatures, shock testing machines of Charpy and Izod t3rpes,
Brinell and Shore hardness testing apparatus, magnetic testing appa-
ratus of Koepsel and Burrows and experimental heat-treatment fur-
naces of both gas and electric types.

3za. Miller Rubber Co., The, Akron, Ohio. (Tires and other rubber
goods.)

Research staff : H. A. Morton and 3 chemists.

Research work : Full time of 4 on rubber and organic chemistry.

Equipment : Scott fabric tester, Curtis & Marbel fabric inspecting
apparatus, tire testing apparatus, etc.; compounding laboratory mill
and calendar, experimental press, etc.

3x3. Milliken» John T., and Co., 217 Cedar St., St. Louis, Mo. (Medi-
cines and pharmaceutical products.)

Research staff : Edsel A. Ruddiman and 2 assistants.

Research work : Part time of 3 on medicinal agents.

314. Milwaukee Coke & Gas Company, The, ist National Bank Build-
ing, Milwaukee, Wis.

Research staff: George H. Selke and a number of chemists.

Research work : Full time of i to increase efficiency of by-product
coke plant; includes heating of ovens, and recovery of light oil, am-
monia, gas, etc.

315. Mineral Refining & Chemical Corporation, Carondelet Station,
St. Louis, Mo. (Dry paint pigments.)

Research staff: B. B. McHan and 5 assistants.

Research work: Approximately one-fourth time of 6 on zinc and
cadmium hydrometallurgy in its relation to pigment manufacture, and
the separation and recovery of the impurities ; also barium compounds.

316. Miner Laboratories, The, 9 S. Clinton St., Chicago, 111. (Con-
sulting chemists ; pharmaceutical and food problems.)

Research staff: C. S. Miner, 9 chemists and 2 analysts.

Research work : Full time of 4 chemists and part time of 3 chemists
on utilization of oat hulls ; cause of rancidity of vegetable oils ; pre-
cooked cereals; yeast manufacture; dehydration of potatoes; also
many research problems are handled as a part of consulting service.
Supervision of research in molded insulation.

Equipment : Small scale cereal manufacturing equipment.
3x7. Minneapolis Steel and Machinery Co.» ^54 Minnehaha Ave.,
Minneapolis, Minn. (Tractors, threshers, structural steel work,
engines, hoists, etc.)

Research staff : C. S. Moody, 2 engineers and i assistant engineer,
3 chemists and 2 assistant chemists; A. W. Scarratt, automotive en-
gineer, I engineer and 3 assistants.

54 INDUSTRIAL RESEARCH LABORATORIES

Research work: One-fourth time of 14 on materials and construc-
tion.

Equipment : Izod impact testing machine, ioo,ooa-pound automatic
autographic Olsen testing machine, Brinell hardness machine, small
electric furnace for temperature up to 1800 degreee F., Leeds and
Northrup potentiometer, Leeds and Northrup optical pyrometer,
metallographical eaumment and Riehle testing machines, Sprague
dynamometer 100 H. P. at 500 R. P. M.

3x8. Hojonnier Bros. Co., 739 W. Jackson Boulevard, Chicago, 111.
(Scientific dairy apparatus and supplies ; milk testing.)

Research staff : Timothy Mojonnier and J. J. Mojonnier, i analyst,
3 chemists and 2 chemists and bacteriologists.

Research work: One-tenth time of 8 on scientific control of milk
and milk products, particularly in evaporated and condensed plants,
ice-cream plants and large dairies. Effect of preservatives on com-
posite milk samples ; culture, propagation, etc.

Equipment: Mojonnier Model D Milk Tester, containing rapid
cooling desiccators; the Mojonnier Model E Culture Controller for
the continual propagation and control of pure lactic cultures; sedi-
ment tester, acidity and salt tester.

319. Monroe Drug Company, Color Chemical Division, Bottom Road,
Quincy, 111.

Research staff: H. E. Kiefer and 4 assistants.
Research work: Approximately one-fourth time of 5 on direct
union colors and intermediates used in their manufacture.

320. Monsanto Chemical Works, 1800 South 2nd St., St Louis, Mo.
(Fine and medicinal chemicals, dye intermediates, sulphuric and other
technical acids, phenol and other heavy chemicals.)

Research staff: Jules Bebie, 30 chemists, 4 engineers and i safety
engineer.

Research work : Full time of 5 or 6 chemists on subjects related to
synthetic pharmaceuticals and fine chemicals, including intermediates.

Equipment: Semi-commercial scale experimental laboratory.
331. Morrill, Geo. H., Co., Norwood, Mass. (Printing and litho-
graphic inks.)

Research staff : Olney P. Anthony and 3 chemists.

Research work : Full time of 4 on ink research.

Equipment : Dye experimental apparatus.
322.^ Morris & Company, Union Stock Yards, Chicago, 111. (Packers
and provisioners.)

Research staff : J. J. Vollertsen, 3 chemical engineers, i chemist and
T bacteriologist.

Research work : Full time of 6 on industrial investigations of pack-
ing house problems and by-products.

333. Mulford, H. K., Company, Biological Laboratories, Glenolden,
Pa. (Manufacturing and biological chemists.)

Research staff : John Reichel and 9 persons ; in addition, dozens of
staff and laboratory assistants engage in some research.

Research work: One-third time of 10 and part time of laboratory
staff on problems connected with pharmacology, bacteriology, im-
munology and serology.

INDUSTRIAL RESEARCH LABORATORIES 55

Equipment : Specially equipped for dealing with problems relating
to pharmaceutical, biological, biological agricultural work and chem-
istry of soil, and for bacteriological and serological work.

334. Munn, W. Faitoute, 518 Main St., E. Orange, N. J.
Research staff: W. Faitoute Munn.

Research work: Nine-tenths time of i on electric furnace, color
photography and industrial lines in general.

335. Munmng, A. P., & Co., Matawan, N. J. (Electroplating and
buffing apparatus and supplies.)

Research staff : G. A. Cheney, i chemist and i consulting mechan-
ical and electrical engineer.

Research work : Approximately one-half time of i on problems in
connection with the electroplating of metals with the removal of
grease and dirt from metal surfaces, the polishing of various surfaces
and the compounds required for such polishes.

Equipment : Complete apparatus for electroplating.

336. Musher and Company, Incorporated, Baltimore, Md. Formerly
The Pompeian Co.

Research staff : Louis M. Roeg and 2 assistant chemists.

Research work : Full time of 3 along general lines of food products
with special attention to expression, care and utilization of vegetable
oils.

Equipment: Small scale food manufacturing operations, such as
expression and filtration of oils.

337. National Aniline & Chemical Company, Incorporated, 21 Burling
Slip, New York, N. Y. Research laboratories at Buffalo and Marcus
Hook, Pa. Dye laboratories at Buffalo and at various sales branches.

Research staff: G. C. Bailey and 9 chemists at Marcus Hook.
Varying number of chemists, engineers and other technical men at
other laboratories.

Research work : Almost entirely on dyes and intermediates.

Equipment: Semi-commercial scale equipment for testing pro-
cesses before putting them on a manufacturing basis.

338. National Association of Corrugated and Fibre Box Manufac-
turers, The, 1821 Republic Building, Chicago, 111.

Research staff: Fred D. Wilson and i assistant.

Research work : Full time of i on designing and testing corrugated
and solid fibre containers to develop the best container for the com-
modity experimented with.

Equipment : Revolving testing drum for fibre boxes.

339. National Biscuit Company, 409 W. Fifteenth St., New York,
N. Y. Formerly Marvin-Davis Laboratories, Incorporated.

Research staff: Clarke E. Davis, 4 chemists, i engineer, i baker
and I assistant.

Research work : Full time of 8 on food products, their packing and
distribution.

National Board of Fire Underwriters. See Underwriters' Lab-
oratories (p. 77).

330. National Canners Association, 1739 H St. N. W., Washington,
D. C.

Research staff: W. D. Bigelow, 4 chemists and 3 bacteriologists.

j

56 INDUSTRIAL RESEARCH LABORATORIES

Research work : Full time of i and part time of i on study of tin
plate from all standpoints ; causes of pinholing in tin cans ; influence
of composition and details of manufacture of steel on service value of
tin plate. Full time of i on study of heat penetration of canned food ;
study of various factors affecting penetration of heat to the center of
the can ; distribution of heat in sterilizing kettles in different systems
of management. Full time of 3 on study of microorganisms causing
spoilage; isolation of spoilage bacteria and study of their cultural
characteristics with special reference to thermal death point; study
of habitat of spoilage organisms in canning plants and farms where
raw products are grown. Full time of i and part time of i on study
of minor miscellaneous technological qeustions arisinp; from time to
time. Some of the most effective work has been done m collaboration
with other organizations. For instance, the tin plate investigations
are conducted in collaboration with manufacturers of steel, tin plate,
and cans.

Equipment: Special canning equipment with laboratory facilities.
Experimental small factory scale cannery and canning laboratory.

National Carbon Company. See Union Carbide and Carbon Re-
search Laboratories, Inc. (p. 78).
331. National Cash Register Company, The, Dayton, Ohio.

Research staff: A. B. Beaver, 12 chemists, 3 electrical engineers,
6 mechanical engineers and 2 metallurgists.

Research work: Full time of 8 and approximately one-tenth time
of others on chemical, mechanical, electrical, metallurgical and manu-
facturing problems.

Equipment: Special equipment for conducting endurance tests on
cash registers.

33a. National Cereal Products Laboratories, 1731 H St. N. W., Wash-
ington, D. C. (Chemical and technical advisors for The National
Macaroni Manufacturers' Association and The Alimentary Paste
Manufacturers' Association.)

Research staff: B. R. Jacobs and i chemist.

Research work : One-fourth time of 2 on standardization of cereal
products and raw materials entering into their composition, methods
of control in purchasing raw materials and containers for cereal
products.

333. National Gum & Mica Co., 12 West End Ave., New York, N. Y.
Research staff: S. Ginsburg, chemist, A. A. Haldenstein, chemical

engineer, and 3 assistants.

Research work : Four-fifths time of 5 on adhesives, colloids, gums,
starches, colors, sizings, finishings, etc., for paper and textiles.

334. National Laboratories, The, 1313 H St. N. W., Washington,
D. C.

Research staff : Ivan S. Hocker, 2 chemical engineers, i mechanical
engineer, 2 chemists, and i bacteriologist.

Research work : Gelatine, bacteriological dyes, by-products in acid
industries, yeast and fermentation problems, malt extracts and bread
improvers, glass, flotation oils and paints, cellulose and paper.
335-339« National Lamp Works of General Electric Company, Nda
Park, Cleveland, Ohio. Research Department: Edward P. Hyde,

INDUSTRIAL RESEARCH LABORATORIES 57

director of research, Francis E. Cady, manager, and J others. Instru-
ment shop, power plant, lamp shop and library. Renders service to
other research and development laboratories.

335. Nela Research Laboratories
Laboratory of Pure Science

Research staff: Directorship vacant; 3 physicists, i physical-
chemist, I psychologist, 2 biologists, 5 laboratory assistants, i student
on Brush Fellowship.

Research work: Full time of 13 on the physics, physiology, and
psychology of light, particularly in those phases which pertain to the
science of illumination ; the production of luminous energy ; the laws
of radiation ; and the effects of luminous and attendant radiation, par-
ticularly in connection with its physiological, psychological, biologi-
cal, and chemical action. Records of researches are presented before
scientific and technical societies and are published as contributions to
the technical journals.

Laboratory of Applied Science

Research staff: M. Luckiesh, 3 physicists, i engineer-physicist, 2
assistant physicists, i architect-engineer, i architect-designer, i light-
ing assistant, 4 laboratory assistants, 2 clerical workers.

Research work: Full time of 14 on spectrum analysis; light-pro-
duction ; spectrophotometry ; photometry ; various physical properties
and measurements pertaining to glass, metals, etc. ; physical, biolog-
ical, physiological, photo-chemical, and psychological aspects of light
utilization ; various phases of color.

336. Lamp Development Laboratory

Research staff: J. E. Randall, consulting engineer, W. L. Enfield,
manager, and I9 men.

Research work: Full time of 21 on development of processes of
manufacture of incandescent lamps ; investigation of quality of prod-
uct ; design of lamps ; development of new types of lamps ; investiga-
tions of raw materials for use in manufacture of lamps ; development
work on tungsten wire.

Equipment: Special equipment for use in lamp manufacture built
by National Lamp Works shop.

337. Experimental Engineering Laboratory
Research staff: Frank M. Dorsey and 33 assistants.
Research work : One-half time of 34 on a variety of problems.
Equipment: Adequate facilities for large-scale experiments,

whether on lamp making or chemical and metallurgical processes.

338. Glass Technology Department

Research staff: Wm. M. Clark, 7 technical men and 3 experienced
practical glassmen.

Research work : One-half time of 4 on development work on glass
parts used in connection with the manufacture of incandescent lamps.

Equipment: High temperature furnace equipment both gas and
electrically heated. Physical and optical apparatus for determining
the physical and optical properties of different glasses.

339. Engineering Department

Research staff: S. E. Doane, chief engineer, and 57 electrical
engineering graduates.

58 INDUSTRIAL RESEARCH LABORATORIES

Research work: One-half time of 15 on determining performance
and characteristic data on incandescent lamps and lamp accessories;
study of economics of light production; study of methods of light
utilization from standpoint of obtaining most satisfactory illumination
results.
340. National Lead Company. 129 York Street, Brooklyn, N. Y.

Research staff: Gustave W. Thompson, 3 assistants, 7 special in-
vestigators and analysts, 2 paint experts, i colorist, and necessary
assistants.

Research work: Large part of time of 34 on investigations con-
nected with manufacture and utilization of lead products (white lead,
lead oxides, alloys, etc.), other paint pigments, linseed oil and other
paint vehicles, paint technology, metallurgy of lead and of tin, physical
testing and metallography of white metal alloys, microphotography,
etc.

Equipment: Apparatus for testing of pigments, oils and metals,
including special apparatus for measuring whiteness of pigments;
opacity of paint films; fineness of pigments by classification; How-
land color photometer ; tension and hardness testing machines.
34Z. National Lime Association, 918 G St. N. W., Washington, D. C.

Research staff: M. E. Holmes, E. O. Pippin and 2 assistants. In
addition to the resident staff, there are 5 others in university and gov-
ernment laboratories working on fellowships.

Research work: Full time of 3 on properties and uses of lime in
the chemical, agricultural and construction fields.
34a. National Malleable Castings Company, The, 10600 Quincey
Ave., Cleveland, Ohio.

Research staff : H. A. Schwartz, 3 metallographers and chemists, i
physicist, i tester of materials and 3 assistants.

Research work: Full time of 9 on properties of ferrous alloys,
especially fatigue, alternating and impact stresses and resistance to
cutting; equilibrium conditions in non-carbon alloys, particularly in
stable system ; miscellaneous metallurgical investigations.

Equipment : One 50,000-pound for 6-foot specimens, and one 200,-
000-pound Olsen 3-screw testing machine; 00,000-inch-pound Olsen
torsion machine; Olsen universal efficiency testing machine; Charpy
impact machine ; Brinell machine, scleroscope ; inverted type Bausch &
Lomb metallographic microscope; automatic and autographic appa-
ratus for precision heat treatment of metals.

National Stain and Reagent Co. See Coleman & Bell Company,
The (p. 20).

343. National Tube Company, Frick Building, Pittsburgh, Pa. (Steel
and iron tubes and pipes.)

Research staff : F. N. Speller and 6 to 8 men.

Research work: Full time of staff on metallurgical and chemical
research work as applied to mill operations and various uses of tubular
material by consumers. Considerable portion of time devoted to the
problem of corrosion and protection of iron and steel from corrosion.

344. Naugatuck Chemical Company, The, Naugatuck, Conn.
Research staff: H. S. Adams, 3 chemists and 5 assistants.

INDUSTRIAL RESEARCH LABORATORIES 59

Research work : Full time of 9 on chemicals pertaining to the rub-
ber industry.

Nela Research Laboratories. See National Lamp Works of
General Electric Company (p. 56).

345. Nestli's Food Company, Incorporated, 130 William St., New
York, N. Y. (Condensed milk.) Laboratory also at Ithaca, N. Y.

Research staff: A. A. Scott, i bacteriologist and micologist and i
assistant ; 2 chemists and 2 assistants. F. E. Rice and i assistant at
Ithaca laboratory.

Research work : Full time of 3 on sweetened condensed and evap-
orated milk and other products that the company produces or may
produce.

Equipment : Experimental equipment for production of condensed
and evaporated milk.

346. Newark Industrial Laboratories, 96 Academy St., Newark, N. J.
(Conduct researches on an experimental as well as on a semi-com-
mercial scale.)

Research staff: Hubert Grunenberg and 3 assistant collegiate
chemists.

Research work: Development of synthetic flavoring matters, per-
fumes, drugs, and dyes.

347. New England Confectionery Company, 253 Summer St., Boston,
Mass.

Research staff : Edmund Clark and i chemist.
Research work : Nine-tenths time of 2 on problems connected with
the industry.

348. New Jersey Zinc Company, The, 160 Front St., New York, N. Y.
Research staff : J. A. Singmaster, manager of technical department,

F. G. Breyer, chief research division, 14 chemists, 8 physicists and 12
assistants.

Research work : Full time of 34 on mechanical and physical inves-
tigations connected wth metallurgy of zinc ; manufacture and use^ of
zinc oxide in rubber and paint industries ; manufacture and utilization
of sulphuric acid ; production and properties of worked metallic zinc
in shapes of strips, sheets, etc.

349. Newport Company, The, Pensacola, Fla.
Research staff : R. C. Palmer and 2 assistants.

Research work : Whole time of i and one-quarter time of 2 on prob-
lems relating to the technical and industrial development of terpenes
and terpene products, rosins and rosin products.

Newport Turpentine & Rosin Company of Florida. See New-
port Company, The.

350. New York Quebracho Extract Company, Incorporated, 80
Maiden Lane, New York, N. Y.

Laboratory at Greene and West Sts., Greenpoint, Brooklyn, N. Y.

Research staff : R. O. Phillips and 4 chemists.

Research work: One-half time of 5 on tannery operation, extract
manufacture and various problems in connection with the manufacture
and testing of leather.

Equipment: Experimental tannery.

60 INDUSTRIAL RESEARCH LABORATORIES

351. New York Quinine & Chemical Works, Incorporated, The, 135

William St., New York, N. Y.

Research staff : George L. Schaef er, 7 chemists and 2 engineers.

Research work: Approximately one-half time of 7 chemists on
organic products, alkaloids, and medicinal chemicals.
35a. New York Sugar Trade Laboratory, Inc., The, 79 Wall St., New
York, N. Y.

Research staff : C. A. Browne, S chemists and i helper.

Research work : One-fourth time of 7 on composition and deteriora-
tion of sugars; optical and chemical methods of sugar analysis;
influence of temperature and other conditions on polarization of
sugars ; composition and food value of syrups and molasses.

Equipment: Constant temperature laboratory for polarization of
sugars.

353. Niles Tool Works Company, The, 545 North Third St., Hamil-
ton, Ohio. (Machine tools.)

Research staff: J. W. Bolton, i experimental engineer, 2 routine
men and labor as desired.

Research work: One-fourth to three-fourths time of 4 on metal-
lurgy of grey iron, especially practical applications of metallography,
studies of changes produced by pouring temperatures, section size,
etc. Heat treatment, brass and bronze, core oils, etc.

Equipment : Completely equipped laboratory for study of grey iron.
354* Northwestern Chemical Co., The, Marietta, Ohio. (Chemical
automobile utilities.)

Research staff: A. S. Isaacs and 2 advisors.

Research work : One-half time of i on problems incident to auto-
mobile trade and news ink trade; cements, polishes, dressings and
enamels, printers' ink, oil and carbon black.

355. Norvell Chemical Corporation, The, 1 1 Cliff St., New York, N. Y.
Research staff: 4 chemists.

Research work : One-fourth time of 4 on mercurial products, phos-
phates, benzoate group, wood distillation derivatives, formaldehyde
condensation products, citric and oxalic acid derivatives, aniline de-
rivatives, phosgene condensation products and other pharmaceutical
and technical products.

356. Nowajc Chemical Laboratories, 518 Chemical Building, St. Louis,
Mo.

Research staff: C. A. Nowak.

Research work : On flavoring extracts used in soft drink manufac-
ture.

Equipment: Well equipped for brewery and other beverage and
food work.

357. Nulomoline Company, The, 11 1 Wall St., New York, N. Y.
(Glycerine substitutes.)

Research staff: M. A. Schneller, i chemist, i confectionery engi-
neer and I laboratory assistant.

Research work: Approximately one-half time of 3 on sugar and
sugar products.

358. Ohio Fuel Supply Company, The, 99 N. Front St., Columbus,
Ohio. Laboratory at Utica, Ohio.

INDUSTRIAL RESEARCH LABORATORIES 61

Research staff: George T. Koch, 2 chemists, 2 chemical engineers
and 2 routine men.

Research work: Approximately three-fourths time of 5 on petro-
leum, natural gas, gasoline, particularly the manufacture of synthetic
chemicals, such as amyl acetate, formaldehyde, formic acid, etc., from
the above natural products and absorption processes for gasoline.

359. Ohio Grease Co., The, Londonville, Ohio. (Lubricants.)
Research staff: i chemist.

Research work: Analysis of oils, fats and greases, such as are re-
quired in a grease factory.

360. Oliver Continuous Filter Co., 503 Market St., San Francisco,
Calif. Laboratories also at 226 E. 41st St., New York, N. Y., and
No. 9 Red Lion Passage, Holborn, London, W. C. I., England.

Research staff: E. L. Oliver in San Francisco, R. Gordon Walker
in New York and J. F. Mitchell-Roberts in London, with 3 engineers
and I chemist available for each laboratory.

Research work: Investigation of methods for increasing efficiencv
and reducing costs of filtration of all classes of chemical and metal-
lurgical products. No work done on drinking water filtration. Prin-
cipal products investigated are beet and cane sugar juices and saccha-
rate of lime ; lime sludges ; wood pulp ; sewage ; phosphoric acid ; cy-
anide slimes ; flotation concentrate ; clays of all kinds ; dyes, etc.

Equipment : Continuous vacuum filters, small intermittent vacuum
filters, various devices for treating filter "cake" during the filter cycle
to reduce moisture or increase washing- efficiency.

Orford Copper Co., The. See International Nickel Company,
The (p. 44).

361. Package Paper and Supply Corporation, 150 Birnie Ave., Spring-
field, Mass. (Waxed papers.)

Research staff: W. M. Bovard, 2 chemists, i engineer and i as-
sistant.

Research work : Approximately three-tenths time of S on wrapping
food products, especially for moisture protection, specializing on
waxed paper for automatic wrapping machine for wrapping soap,
cereals, food products and candy and developing special papers.
363. Packard Motor Car Company, Detroit, Mich. Engineering
laboratory.

Research staff : L. M. Woolson, 3 engineers and i chemist.

Research work : Full time of 5 on problems connected with' Liberty
motor, motor trucks and automobiles; automobile and truck chassis
development.

Equipment: Complete dynamometer equipment for testing truck,
car and airplane engines up to 500 H. P. Complete bench testing
equipment for all car, truck and airplane accessories. Automotive
power plant and accessories.

Page, Carl H. See Riverbank Laboratories (p. 68).
363. Palatine Aniline and Chemical Corporation, 81 N. Water St.,
Poughkeepsie, N. Y. (Dyestuffs and chemicals.)

Research staff: Felix Braude and 2 chemists.

Research work : Full time of 3 on intermediates and dyestuffs.

62 INDUSTRIAL RESEARCH LABORATORIES

364. Palmolive Company, The, Milwaukee, Wis.
Research staff: V. K. Cassady and 7 assistants.

Research work : Full time of i and approximately one-fourth time
of 6 on soaps and perfumes.

365. Pantasote Leather Company, The, Passaic, N. J.
Research staff: Edgar Josephson.

Research work : Full time of i on coatings for textiles, rubber coat-
ings for fabrics, oils, paints, varnishes and all closely related indus-
tries.

366. Parke, Davis & Company, Detroit, Mich. (Medicinal prepara-
tions.)

Research staff: J. M. Francis, chief chemist, Oliver Kamm, chief
of chemical research department, E. M. Houghton, chief of medical
research department and about 40 chemists, pharmacists, bacteriolo-
gists, botanists and pharmacologists.

Research work : Large part time of about 20 is devoted to the im-
provement in the constitution, or processes of manufacture, of sub-
stances now used as medicaments; and in the attempt to discover or
produce new therapeutic agents in both pharmaceutical and biologic
lines.

Patton Paint Company. See Pittsburgh Plate Glass Co. (p. 65).

367. Pea3e Laboratories, 39 West 38th St., New York, N. Y. (Suc-
cessors to Lederle Laboratories.)

Research staff : H. D. Pease and a number of chemists, bacteriolo-
gists and assistants.

Research work: Small part time of staff along sanitary, chemical
and bacteriological lines.

368. Peerless Color Company, Bound Brook, N. J.
Research staff: R. W. Comelison and 2 chemists.

Research work: Part time of 3 on problems dealing directly with
the manufacture of dyestuffs.

369. Peerless Drawn Steel Company, The, Massillon, Ohio.
Research staff: A. M. LeTellier and 4 assistants.

Research work : Approximately one-half time of 5 on effect of heat
treating and cold drawing on all grades of steel and development of
the cold drawing of steel.

Equipment: Apparatus for studying chemical and physical prop-
erties of steel, including full heat treating department as well as
metallography department.
369a. Peet Bros. Mfg. Co., Kansas City, Kans.

Research staff: W. J. Reese and 2 assistants.

Research work: Problems connected with the manufacture of
soaps and glycerin.

370. Penick & Ford, Ltd., Incorporated, New Orleans, La. (Sugar,
cane and corn products.) Laboratory at Marrero, La.

Research staff: F. W. Zerban, i chemist, i assistant chemist and
assistants.

Research work : Full time of 3 or more on manufacture and refin-
ing of the products of sugar cane, corn and other saccharine plants.

371. Pennsylvania Railroad Company, The, Altoona, Pa.
Research work: Small part time of staff on investigation of cause

of failure of steel rails ; locomotive design ; much work in preparation
of specifications for various materials; general field of lubrication;

INDUSTRIAL RESEARCH LABORATORIES 63

water treatment and purification ; paints and preservatives ; heat treat-
ment of metals, etc. Investigation of electrolysis in systems of under-
ground metallic structures; tests and investigations of the construc-
tion of various makes of transformers ; tests of various makes of pri-
mary and secondary battery cells; oscillo^aphic tests for linear and
angular velocity, wave forms, etc.; investigations of special cases of
electrical troubles ; development of an electrical method of measuring
the hardness and homogeneity of steel. Tests of locomotives on the
road or tests of equipment with special devices; tonnage rating of
trains and following up of all experimental appliances which are put
into service for test purposes. Methods for determination of elements
in plain-carbon steels, alloy steels and non-ferrous alloys used for
bearing backs and linings, packing-ring metal for different purposes,
etc. Examination of fuels, development of specifications for paint
products, lubricating and burning oils, boiler compounds, lacquers,
plush, car cleaners, cutting compounds, belt dressing, polishing com-
pounds, hydraulic- jack liquids, fuses, track caps, fire-extinguishing
preparations, the recovery of used or wasted products, etc.

Equipment: Six universal tension and compression testing ma-
chines, one of 1,000,000, two of 300,000, two of 100,000-pound and one
of iso,ooo-pound capacity; one vibratory endurance spring testing
machine of 75,000-pound capacity; one 43-foot and one 57-foot drop-
testing machine ; two vibrating staybolt testing machines ; one Brinell
hardness testing machine; one 2000-pound cement testing machine;
metallographic equipment.

Apparatus for testing hose: Six rubber stretching machines; one
friction test rack for rubber ; one hose mounting machine ; one vibrat-
ing test rack for hose; one continuous test rack for rubber; four ten-
sion testing machines for rubber ; one stretching machine for rubber
insulation ; one spring micrometer machine ; one vacuum gage testing
machine ; one arbor press specimen cutter ; one hydraulic gage testing
machine, capacity 25,000 pounds per square inch; one dead-weight
gage testing machine, capacity six gages; one wiggling testing ma-
chine for hose ; one bumping testing machine for gages ; one whipping
testing machine for g^ges; one hydraulic machine for testing gage
glasses.

Rubber, air-brake hose and miscellaneous laboratory, machines for
air-brake, signal and tank hose, and other miscellaneous tests.

Electrical laboratory, equipment for lamp tests consisting of three
photometers, lamp test rack of 1000 lamps capacity, with switchboard,
transformers and potential regulator equipment.

372. Pennsylvania Salt Manufacturing Co., Philadelphia, Pa.
Research staff: Director, chief chemist and 3 assistant chemists.
Research work: Problems relating to the manufacture of heavy

chemicals.

373. Permutit Company, The, 440 Fourth Ave., New York, N. Y.
(Water rectification systems.) Factory at Brooklyn, N. Y.

Research staff : T. R. Duggan, 7 chemists and 4 chemical engineers.
Research work: Full time of 3 entirely in connection with water
problems and the use and manufacture of artificial zeolites.

64 INDUSTRIAL RESEARCH LABORATORIES

374. Perolin Company of America, The, 2010 Peoples Gas Bldg., Chi-
cago, 111. Laboratory at 11 12 W. 37th St., Chicago, 111.

Research staff : E. L. Gross, chemical engineer.
Research work : Protection of metal surfaces against rust and pit-
ting and boiler scale removal and prevention.
Equipment : Beach-Russ vacuum pump and copper retorts.

375. Pettee, Charles L. W., Laboratories of, 112 High St., Hartford,
Conn. (Analytical and consulting chemist.)

Research staff: C. L. W. Pettee and i chemist.
Research work : Three-twentieths time of 2 on recovery and puri-
fication of precious metals.

376. Pfaudler Co., The, Rochester, N. Y.

Research staff: O. I. Chormann, i chemist, i metallurgist and i
helper.

Research work: Three-fourths time of 3 on enamels for steel and
cast iron ; packings ; resistivity of enamels, etc.

377. Pfister ft Vogel Leather Co., 447 Virginia St., Milwaukee, Wis.
(Tanners and curriers.)

Research staff: Louis E. Levi, 2 research chemists and 7 other
chemists.

Research work : Full time of 4 on problems related to leather, glue,
hair, gelatine, retarder, bitumen, paints, etc.

378. Pfizer, Chas., ft Co., Inc., 81 Maiden Lane, New York, N. Y.
(Manufacturing chemists.) Laboratory at 11 Bartlett St., Brooklyn,
N. Y.

Research staff: Richard Pastemack, 5 chemists and chemical engi-
neers and I engineer.

Research work: Full time of 7 on development of processes and
products.

Equipment: Complete laboratory and semi-plant equipment.
379* Pharma-Chemical Corporation, 1570 Wool worth Bldg., New
York, N. Y. Laboratory at Bayonne, N. J.

Research staff : Eugene A. Markush, 3 chemists and i engineer.

Research work: Dyes and pharmaceuticals.

380. Philadelphia Quartz Company, Philadelphia, Pa. (Silicate of
soda.)

Research' staff : James G. Vail, 4 chemists and i assistant.

Research work : One-half time of 6 on problems involving applica-
tion or manufacture of silicate of soda, study of its properties as an
adhesive, as an ingredient of acid-proof cement, grinding wheels, soap,
asbestos insulating material, coating materials for paper and wooden
packages, to prevent the absorption of grease, as an agent in refining
of vegetable oils, etc.

Equipment : Crushing and grinding apparatus, two gas-heated fur-
naces for experiments with fusion, one a small open hearth, and the
other a crucible furnace; apparatus for fusion, testing of adhesives,
cement, etc., and devices for making the usual commercial tests on
paper; small and semi-commercial autoclaves.

381. Ph]r8icians and Surgeons Laboratory, 605 Paxton Blk., Omaha,
Nebr.

INDUSTRIAL RESEARCH LABORATORIES 65

Research staff: Theodore M. Agnew, i chemist, i bacteriologist
and I pathologist and serologist.

Research work : . Variable amount time of 4 on bacteriological, path-
ological and serological problems.

382. Pierce-Arrow Motor Car Company, The» Elmwood Ave., Buffalo,
N. Y.

Research staff : J. Miller, metallurgist, and 2 assistants ; W. Slaght,
experimental engineer and 2 assistants.

Research work: Approximately one-fourth time of 8 on cause of
failure of parts, effect of impurities in metals, heat treatment, effect
of shocks, alternate stresses and efKciency of engines and transmis-
sions.

Equipment : Olsen testing machine, Avery impact testing machine,
Stanton impact testing machine, 150 H. P. electric dynameter and
engine test stand.

Pitcaim Varnish Co. See Pittsburgh Plate Glass Co.
383-384. Pittsburgh Plate Glass Co., Milwaukee, Wis. Laboratory
also at Newark, N. J.

383. Paiton-Pitcaim Division (Patton Paint Company and Pitcaim
Varnish Company).

Research staff: A. H. Woltersdorf and assistants at Milwaukee;
T. R. Collins and 2 assistants at Newark.

Research work : Part time of staff on problems connected with the
paint and varnish industry.

384. Corona Chemical Division (Corona Chemical Company).
Research staff: C. B. Dickey and assistants.

385. Pittsburgh Testing Laboratory, 616 Grant St., Pittsburgh, Pa.
Laboratories also in New York, N. Y., Birmingham, Ala., and Cin-
cinnati, Ohio.

* Research staff : Jas. O. Handy, director of special investigations,
H. H. Craver, manager chemical department, 26 chemists in Pitts-
burgh, 2 in New York, 3 in Birmingham and i in Cincinnati ; 3 me-
chanical and 3 civil en^^neers.

Research work : Variable amount of time of staff on food and drugs
(alcohol substitutes, etc.), oil refining (lubricating oil recovery), cor-
rosion-resisting metals, water purification, metal extraction from ores
and refractory materials (basic).

Equipment: Furnaces, special metallographic equipment, coal dis-
tillation apparatus (to be installed) and testing machines.

Pompeian Co., The. See Musher and Company, Incorporated

(p. 55)-

386. Porro Biological Laboratories, 625 Puget Sound Bank Bldg.,
Tacoma, Wash. (Successors to Staniford Laboratories.)

Research staff : Thomas J. Porro and John G. Scott.
Research work: Part time of 2 on chemical, serological and bac-
teriological problems.

387. Portage Rubber Co., The, Barberton, Ohio.
Research staff: R. M. Gage and 2 chemists.

Research work : One-half time of 3 on testing and compounding for
rubber goods.

66 INDUSTRIAL RESEARCH LABORATORIES

388. Porter, Horace C, 1833 Chestnut St., Philadelphia, Pa. (Con-
sulting chemist and chemical engineer.)

Research staff : Horace C. Porter and i assistant.

Research work : Coal carbonization, coking and by-products, "low
temperature" carbonization, shale distillation, application of fuels, re-
duction of wastes, coal storage problems and spontaneous combustion.

Equipment: Coal distillation retort (laboratory scale) and acces-
sories.

389. Powers-Weightman-Rosengarten Company, The, 916 Parrish
St., Philadelphia, Pa. (Chemists.)

Research staff: George D. Rosengarten and varying number of
assistants.

Research work : Variable amount time of staff on improvement of
present processes and investigation of new processes.

Prest-O-Lite Co., Inc., The. See Union Carbide and Carbon
Research Laboratories, Inc. (p. 78).

390. Procter & Gamble Co., The, Cincinnati, Ohio. (Soaps, glyce-
rine, candles, lard substitutes, refined oils, etc.) Laboratory at Ivory-
dale, Ohio.

Research staff: H. J. Morrison and 12 chemists.

Research work : Improvement of plant processes and products.

Equipment: Complete experimental plants for the various pro-
cesses.

39Z. Providence Gas Company, Incorporated, Providence, R. I.
Manufacturing Department.

Research staff: A. H. Meyer, i assistant chemist and 2 minor
chemists.

Research work: Small part time of 4 on problems arising in
manufacture.

Equipment : Laboratory is complete for gas plant operation.
39a. Pure Oil Company, Kanawha River Salt and Chemical Division,
Charleston, W. Va. Laboratory at Belle, W. Va.

Research staff : W. A. Borror and i chemist.

Research work : One-half time of i on salt industry, salt brine and
development of processes.

393. Pure Oil Company, Moore Oil and Refining Company Division,
York and McLean Aves., Cincinnati, Ohio.

Research staff : Frank Groodale and 2 assistants.
Research work: Full time of 3 on soaps, greases, polishes, lubri-
cating and soluble oils ; textile, boiler and cutting compounds.

394. Pyrolectric Instrument Company, 636 E. State St., Trenton, N. J.
(Pyrometric and electrical precision instruments.)

Research staff: H. L. Saums, i chemist, i electrical engineer and
I mechanical engineer.

Research work: Approximately one-fourth time of 4 on construc-
tion and adaptations of electrical instruments; special problems re-
quiring combination of mechanical and electrical development; tem-
perature measurement problems, problems in hydrogen-ion determi-
nations.

395. Pjrro-Non Paint Co., Inc., 505 Driggs Ave., Brooklyn, N. Y.
(Fire retarding paints and products.)

INDUSTRIAL RESEARCH LABORATORIES 67

Research staff : Ernest A. Marx, i chemical engineer and i chemist.
Research work: One-half time of 3 on technical paints and paint
products.

Equipment : Inflammability test apparatus.

396. Quinn, T. H., ft Comi>any» which includes: Lackawanna, Sus-
quehanna, Vandalia, Tonesta Valley, Keystone, Heinemann, Barclay,
Beerston Acetate Co., Smethport Chemical Companies and the Quinn
Laboratories Company. General office at Olean, N. Y. Laboratory
at E. Smethport, Pa.

Research staff: Edward E. Currier, 3 chemists, i engineer and
occasional assistance from other specialists.

Research work: Approximately one-third time of 5 on researches
on gases from wood, researches on the phenolic constituents of wood
oils and tars, formaldehyde and physical properties of charcoals.

Equipment : Destructive distillation plant and formaldehyde plant,
both on small scale.

397. Radiant Dye ft Color Works, 2837 W. 21st St., Brooklyn, N. Y.
Research staff: William Goldstein and i chemist.

Research work : Full time of 2 on triphenylmethane dyes and their
derivatives.

398. Radium Company of Colorado, Inc., The, i8th and Blake Sts.,
Denver, Colo.

Research staff: W. A. Schlesinger, 12 chemists and 4 engineers.

Research work: Approximately one-fifth time of 17 on radium,
uranium and vanadium.

399- Radium Limited, U. S. A., 2 W. 45th St., New York, N. Y.
(Radium emanation activators, radium ore, apparatus, etc.)

Research staff : Henry H. Singer, i chemist and 2 assistants.

Research work : One-half time of 4 on radium ore, radium, radium
emanation, radium luminous material and all other matters affiliated
with radium and similar products.

Equipment : Electrometers, fontactoscopes, spinthariscopes, ex-
perimental and demonstration outfits and exhibition of rare earth and
all kinds of luminous materials and paints.)
400. Ransom & Randolph Co., The, 518 Jefferson Ave., Toledo, Ohio.

Research staff : Thomas E. Moore, i chemist, i mechanical engineer
and 2 dentists.

Research work : Three-fourths time of 5 on dental materials.
40Z. Raritan Copper Works, Perth Amboy, N. J. Research Depart-
ment.

Research staff: S. Skowronski, 3 chemists and i physicist.

Research work: Full time of 5 on copper metallurgy, electrolytic
refining of copper, and recovery of by-products, gold, silver, platinum,
palladium, selenium, tellurium, arsenic, nickel, antimony.

402. Redlands Fruit Products Company, Redlands, Calif.
Research staff: H. P. D. Kingsbury and i chemist.

Research work : Small part time of 2 on fruit products, for example,
bottling orange juice.

403. Redmanol Chemical Products Co., 636 W. 22nd St., Chicago, 111.
(Acid- and heat-proof varnishes and lacquers, synthetic amber, mould-
ing compounds ; for electrical insulation and other uses.)

I» INDUSTRIAL RESEARCH LABORATORIES

Research staff: L. V. Redman, A. J. Wcith and F. P. Brock; 8
chemists and 6 chemical engineers.

Research work : Full time of 6 on electrical insulation from phenol,
condensation products and synthetic amber*like resins.

Equipment: Vacuum apparatus, rubber mixing rolls, beater mills,
kneading machines, hydraulic presses, stills, dephlegmators and higti
temperature kilns.

404. Reliance Aniline ft Chemical Co.» Incorporated, Poughkeepsie,
N. Y.

Research staff : Philip Kaplan and i chemist.

Research work : One-third time of 2 along lines of synthetic dyes.

405. Remington Arms, Union Metallic Cartridge Company, Bamum
Ave., Bridgeport, Conn. Research Division.

Research staff : 3 chemists, 3 assistant chemists, i metallographist,
I assistant metallographist and pyrometer expert, 2 engineers and 7
raicellaneous.

Rjcsearch work: One-eighth time of 15 on small arms ammunition.

406. Research Corporation, 25 W. 43rd St., New York, N. Y. Labora-
tory at St. Pauls Ave., Jersey City, N. J.

Research work: Problems of converting a work of completed re-
search to commercial or industrial application and use.

Equipment: Apparatus for developing the Cottrell electrical pre-
cipitation processes.

407. Rhode Island MaUeable Iron Works, Hillsgrove, R. I.
Research staff: M. M. Marcus, i chemist and i engineer.
Research work : Part time of 3 on furnace practice and testing.
Equipment : Commercial air furnaces, annealing furnaces and core

ovens.

408. Richards ft Locke, 69 Massachusetts Ave., Cambridge 39, Mass.
(Mining engineers.)

Research staff: Robert H. Richards and Charles E. Locke with
from I to 3 or 4 engineers and chemists.

Research work: Approximately full time on commercial problems
of ore concentration and allied subjects.

Equipment : Full ore testing equipment.

409. Richardson Company, The, Lockland, Ohio. Heppes Roofing
Division and laboratory at. 26th and Lake Sts., Melrose Park, 111.

Research staff : Robert Holz and 4 chemists.

Research work: One-half time on manufacture of asphalt and
roofing products.

4x0. Riches, Piver ft Co., 30 Church St., New York, N. Y. (Chemical
and color manufacturers and importers.) Laboratory at Hillside,
Elizabeth, N. J.

Research work : Insecticides, fungicides and the raw materials from
which they may be made.

411. Riverbank Laboratories, Geneva, 111. (Commercial research and
experimental laboratories.)

Research staff: Carl M. Page, several chemists, physicists and
other assistants.

Research work: Full time of director and part time of others on
physical, chemical and metallurgical problems ; rubber.

INDUSTRIAL RESEARCH LABORATORIES 69

Equipment: Apparatus for work on phenomena of high-potential
discharges and vacuum tubes; includes i6-plate static machine 36-
inch diameter, one 18-inch and one lo-inch spark X-ray coils with
electrolytic and mercury turbine interrupters, one 20,000-volt alter-
nating current transformer with rotary converter, vacuum tube oven,
assortment of special tubes, Gaede mercurial air-pump for high
vacuum with a Geryk oil-pump as auxiliary. Large special arc lamps
for ultra-violet rays ; apparatus for work in molecular transformations
of hydrocarbon oils; turbine-driven Sharpless super-centrifuge, with
many accessories of own design ; small shop for making special appa-
ratus.
4za. Rochester Button Company, 300 State St., Rochester, N. Y.

Research staff : J. F. Clark, i chemist, 2 engineers, i designer and
2 assistants.

Research work : Full time of 7 on investigation of plant processes,
materials and machinery used in manufacturing buttons.

413. Rodman Chemical Company, Verona, Pa. (Case hardening and
carbonizing compounds.)

Research staff : Hugh Rodman and 2 assistants.

Research work: Approximately full time of 3 on carburizing of
steel, investigation of carbonizing agents, special coking systems,
activated carbon and general research upon carbon.

414. Roeasler ft Hasslacher Chemical Company, The, Perth Amboy,
N.J.

Research staff: H. R. Carveth, technical director; M. J. Brown,
B. S. Lacy, Sterling Temple, E. A. Rykenboer, chief chemists; 10
research chemists with laboratory and engineering assistants.

Research work : Half time on problems connected specifically with
manufacture of caustic soda; inorganic and organic chlorine com-
pounds; formaldehyde and its compounds; precious metals used in
the arts, principally platinum, gold and silver; ceramic materials,
alkali metals, alkali cyanides, peroxides and persalts ; metal cyanides ;
also problems connected with utilization of products cited a)30ve in
plating ; in bleaching and finishing of textiles ; in enamelling, rubber
accelerators, fumigation.

415. Royster, F. S., Guano Company, Norfolk, Va.
Research staff : E. W. Magruder and 3 chemists.

Research work : Small part time of 4 on fertilizer problems entirely,
such as cause of hardening of acid phosphate, effects of different ma-
terials on each other when mixed, etc.

416. Rubber Trade Laboratory, The, 96 Academy St., Newark, N. J.
(An advisory organization conducting researches by request in indus-
trial establishments. Laboratory investigations are carried on at this
address.)

Research staff : Frederic Dannerth and 4 collegiate chemists.

Research work : Investigations for the industries using rubber and
related gums, paints, oils and varnishes. Investigations for the indus-
tries which make rubberized and water proof fabrics ; coal tar prod-
ucts.
4x7. Rumford Chemical Works, Providence, R. I. (Baking powder.

70 INDUSTRIAL RESEARCH LABORATORIES

yeast powder, bread preparation, phosphatic baking acid, acid phos-
phate, phosphoric acid solutions and similar products.)

Research staff: Augustus H. Fiske, 2 assistant chemists and 5 as-
sistants.

Research work : Equivalent to two-thirds time of i on improvement
of apparatus for manufacture of phosphoric acid and its salts; im-
provement of processes of manutacture and of methods of testing
products in laboratory.

Equipment : Gas-measuring devices for testing baking powder and
specially devised electrolytical apparatus for determination of ma-
terial by electrolysis.
417a. Sabine, Wallace Clement, Laboratcny, Riverbank, Geneva, 111.

Research staff: Paul E. Sabine, 3 physicists and i mechanician.

Research work : Full time of staff on transmission and absorption
of sound by standard constructions, structural materials; physical
characteristics of the ear ; absolute measurements in acoustics, special
problems in architectural design and acoustics.

Equipment: Sound chamber, calibrated sound sources, apparatus
for sound photography, telephonic and other devices for absolute
sound measurements.
4x8. Saginaw Salt Products Co., Saginaw, Mich.

Research staff : John P. Simons and 2 assistants.

Research work: Approximately one-fourth time of 3 on chemical
and engineering problems in connection with evaporators, removal of
impurities from salt brine, etc.
4x9. Sangamo Electric Con^mny, Springfield, 111.

Research staff: F. C. Holtz, i chemist, 3 electrical engineers, 2
assistants and 2 model makers.

Research work : One-third time of 7 on properties of magnet steels ;
endurance of material and precious stones used as bearings, paints,
varnishes, insulations, brass and steel, development of apparatus em-
ploying new principles of operation.

420. Schaeffer Brothers ft Powell Manufacturing Company, 189 N.
Clark St., Chicago, 111. Laboratory at 102 Barton St., St. Louis, Mo.
(Soap, oils, etc.)

Research staff: B. Nichols and 3 assistants.

Research work: One-third time of 4 on vegetable, animal and
mineral oil.

421. Schwarr Laboratories, 113 Hudson St., New York, N. Y. (Food
analyses and research; applied refrigeration; testing of fuels and
lubricants.)

Research staff: Robert Schwarz, 5 chemists, i biologist, i con-
sulting mechanical engineer and 2 assistants.

Research work: One-fifth time of 10 on food and beverage prob-
lems, both chemical and biological.

Equipment: Model brewery of 120 gallons capacity.

422. Scientific Instrument and Electrical Machine Company, The, 500
S. York and 221 West Coover Sts., Mechanicsburg, Pa.

Research staff : W. W. Strong and i or 2 skilled men.

Research work: Practically full time of 3 on ionization of gases.

INDUSTRIAL RESEARCH LABORATORIES 71

precipitation of fumes, deblooming oil, nitrogen fixation, diamond
surfaced glass, smoke and fume recorders and masks, etc.

Equipment: High voltage apparatus, gratings, ultra-violet appa-
ratus.

433. Scotty Ernest, & Companyy Fall River, Mass. (Engineers; ap-
paratus for saving industrial wastes; vacuum evaporators, vacuum
dryers, solvent extraction apparatus, ammonia stills, wood distillation
plants.)

Research staff: H. Austin and Robert W. Macgregor, 4 chemical
engineers.

Research work: One-tenth time of 6 on vacuum evaporation,
vacuum distilling and solvent extraction.

434. Scovill Manufacturing Company, Waterbury, Conn. (All
varieties of brass, bronze and German silver.)

Research staff: 3 metallurgists, i chief chemist and metallurgist
with staff of 27 assistants; 2 mechanical engineers, i electrical en-
gineer with 3 assistants, i plating and finishing expert with 2 as-
sistants.

Research work : About one-tenth time of technical staff is occupied
with research problems.

Equipment: Olsen 100,000-pound universal automatic and auto-
graphic testing machine, 3-screw t3rpe, motor drive, speed 0.025 inch
to 6.50 inches a minute ; Olsen 50,000-pound universal automatic and
autographic testing machine similar to the 100,000-pound machine;
Olsen 200,000-pound universal automatic testing machine; Riehle
2,000-pound testing machine, hand drive for tensile tests only;
Brinell hardness testing machine, capacity 3,000 kilograms pressure ;
Olsen and Erichsen sheet metal testers, for ascertaining ductility;
Shore scleroscope.

425. Sears, Roebuck and Co., Chicago, 111. (Diversified manufac-
turing and mail order business.)

Research staff: G. M. Hobbs, director testing department, C. H.
Higgfins, head chemical laboratory, Elizabeth Weirick, head textile
laboratory, and L. E. Wolgemuth, head mechanical research labora-
tory; 13 chemists, physicists and engineers.

Research work: Approximately one-fourth time of staff on de-
velopment of mechanical devices, methods, factory problems and the
standardization of merchandise.

Semet-Solvay Company. See Solvay Process Company, The

(P- 72).

426. Seydel Manufacturing Company, Jersey City, N. J. (Chem-
icals.)

Research staff : Paul Seydel and 4 to 6 assistants.
Research work : Pharmaceutical and textile chemicals.

427. Sharp & Dohme, Baltimore, Md. (Manufacturing chemists.)
Research staff: Herman Engelhardt, 5 research chemists, i phar-
macologist, I pharmacognosist and 10 pharmaceutical chemists.

Research work: One-half time of 5 on pharmaceutical chemistry,
crude drugs and synthesis of new compounds.

Skayef Ball Bearing Co. See S. K. F. Industries, Inc.

72 INDUSTRIAL RESEARCH LABORATORIES

438. S. K. F. Industries, Inc.» New York, N. Y. Research Labora-
tory, Front St. and Erie Ave., Philadelphia, Pa., also serves Hess-
Bright Manufacturing Co., Philadelphia, Pa., Atlas Ball Company,
Philadelphia, Pa., and Skayef Ball Bearing Co., Hartford, Conn.

Research staff: Haakon Styri, 4 mechanical engineers, i chemist,
2 metallurgists.

Research work : Full time of staff on ball bearing application and
endurance fatigue and improvement of material.

429. Skinner, Sherman & Bsselen, Incorporated, 248 Boylston St.,
Boston 17, Mass. (Chemists and engineers.)

Researeh staff: Gustavus J. Esselen, Jr., 9 chemists, 3 engineers
and 3 bacteriologists.

Research work: Approximately one-half time of 7 on paper, cellu-
lose and its esters, food and canning industries, industrial bacteri-
ology, adhesives and cement and building materials.

430. Solvay Process Company, The, and Semet-Solvay Company,
Syracuse, N. Y. (Alkali, coke and its by-products.) Do research
work also for By-Products Coke Corporation, South Chicago, 111.

Research staff : The Solvay Process Co., Carl Sundstrom, 10 chem-
ists, 5 chemical assistants, 5 clerks and mechanics. Semet-Solvay Co.,
A. C. Houghton, i2 chemists, i chemical engineer, 2 electro-chemical
engineers and 12 chemical assistants and routine men.

Research work : Four-fifths time of 20 and one-half time of 37 on
soda ash, caustic soda, bicarbonate of soda, lime and limestone,
cement, waste disposal, metal corrosion, new alkali products ; potash,
indigo, fixation of nitrogen, coal, light oils, causticizing, oxalic acid,
sulphonation of benzol, picric acid, salicylic acid, chlorination of
toluol, benzaldehyde, benzoic acid, and new products, such as di-
phenyl oxide, benzyl acetate, benzyl benzoate, aspirin, sodium sali-
cylate and cinnamic acid.

Equipment: Electric, steam and gas ovens and furnaces of nearly
all sizes up to 2x3x3 feet, capable of any temperature range up to
1500 degrees C; temperature measuring equipment ranging from
— 100 degrees C. to +1750 degrees C; laboratory kneading and mix-
ing machine.

43Z. Souther, Henry, Engineering Co., The, 11 Laurel St., Hartford,
Conn. (Consulting engineers.)

Research staff: J. A, Newlands, F. P. Gilligan, 7 technically trained
assistants and 4 others.

Research work : Part time of 6 on oils, waters and greases, ferrous
and non-ferrous metals, methods of heat-treatment, electro-plating,
foundry practice, boiler water treatment.

Equipment: Pyrometers, furnaces, lead pot for experimental heat
treatment; 100,000-pound Olsen physical testinp^ machine, Izod im-
pact tester and White-Souther endurance machmes; Emerson bomb
calorimeter.

432. Southern Cotton Oil Company, The, 120 Broadway, New York,
N. Y. Head laboratory at Savannah, Ga.

Research staff : Herbert S. Bailey and 6 or 7 assistants.

Research work : Problems pertaining to' the vegetable oil industry
such as improved methods of analyses, investigation of catalysers and

INDUSTRIAL RESEARCH LABORATORIES 73

their preparation, improvements in the methods of refining vegetable
oils, investigating and finding new uses for by-products.

433. Speciid Chemicals Company, Highland Park, 111.

Research staff: Carl Pfanstiehl, Robert S. Black and 3 assistants.

Research work: Rare carbohydrates, amino acids, rare organic
biological chemicals and industrial specialties.

Equipment : New Bates variable sensibility half-shade polariscope ;
use of bacteria as "living chemical reagents.'*

434. Speer Carbon Company, St. Marys, Pa. (Motor and generator
brushes.)

Research staff : M. S. May, 2 engineers, 2 chemists and 3 assistants.

Research work : Practically the entire chemical and electrical staff
devoted to the development of new products and the improvement of
present products.

435-436. Spencer Lens Company, Buffalo, N. Y. (Optical instru-
ments, optical glass.) Laboratory also at Hamburg, N. Y., in optical
glass factory.

435. Buffalo Laboratory

Research staff: Harry G. Ott and 7 trained assistants.

Research work : Half time of 8 on mathematical designing of lens
systems ; the other half on designing optical instruments and solving
the problems of the manufacture of lenses and optical instruments.

430. Hamburg Laboratory

Research staff: Donald E. Sharp and i trained assistant.

Research work : Full time of 2 on optical glass and problems con-
nected with its manufacture.

437. Sperry, D. R., ft Co., Batavia, 111. (Founders and engineers;
makers of filter presses and evaporators.) Sperry Filtration Labora-
tory.

Research staff: D. R. Sperry.

Research work: One-fourth time of i on systematic effort to de-
termine fundamental laws of filtration.
Equipment : Special filter presses.

438. Sprague, Warner ft Company, 600 West Erie St., Chicago, 111.
(Manufacturers and wholesalers of groceries.)

Research staff : Paul D. Potter and 2 trained chemists.

Research work : One-third time of 3 on problems relating to food.

439. Spreckels Sugar Company, 2 Pine St., San Francisco, Calif.
Research staff : K. E. Christie, i chief chemist, 3 assistant chemists

and 6 bench chemists through operating season of three months; i
chief chemist and i assistant chemist in off season of nine months.
Research work: Equivalent of time of i man for nine months on
extraction and purification of juices; minimization of sugar losses;
reduction of fuel-oil, lime and filter-cloth consumption; recovery of
potash soda and ammonia compounds from Steffen waste.

440. Squibb, E. R., ft Sons, New Brunswick, N. J. (Research and
biological laboratories.)

Research staff: John F. Anderson, 6 bacteriologfists and 3 chemists.
Research work : One-fourth time of 10 on biological and biochemi-
cal problems.

Equipment : For the production, for commercial purposes, of prod-

74 INDUSTRIAL RESEARCH LABORATORIES

ucts for theoretical research in the various phases of biological thera-
peutics.

441. Stamford Dyewood Company, Stamford, Conn.

Research staff: Roy H. Wisdom, i chemist and i engineer.

Research work : One-tenth time of 3 on improvement in manufac-
ture of dyewood extracts and economical methods of use of waste
products.

442. Standard Oil Company (New Jersey), 26 Broadway, New York,
N. Y. Central laboratory at Linden, N. J. Other laboratories at prin-
cipal plants of the Standard Oil Company in the United States and
abroad.

Research staff: Frank A. Howard, manager, C. I. Robinson, chief
chemist, C. O. Johns, director research laboratory, N. E. Loomis, di-
rector, experimental division.

Research work: Petroleum production, products and refining, nat-
ural and artificial gas.

443. Standard Oil Company of Indiana, Whiting, Ind.

Research staff : F. M. Rogers, 7 chemists and 6 assistants.

Research work : Full time of 7 on improvement of niethods of pe-
troleum refining; development of new products and new processes;
study of nature and properties of petroleum products.

Ecjuipment: Fully equipped experimental plant for carrying out
refinmg methods on a scale larger than is possible in the laboratory.

444. Standard Underground Cable Company, 26 Washington St.,
Perth Amboy, N. J.

Research staff: G. D'Eustachio and 2 assistants.
Research work: Approximately half time on insulating material
for electrical purposes.

Staniford Laboratories. See Porro Biological Laboratories.

(p. 65).

445. Stewart - Warner Speedometer Corporation, Chicago, 111.

(Speedometers, tachometers, vacuum gasoline systems, and other
automobile accessories.)

Research staff : F. G. Whittington, chief engineer, i assistant chief
engineer, i research engineer, 3 assistant research engineers, i elec-
trical engineer, 2 designers and inventors.

Research work: Full time of 5 on investigations of fuel feed sys-
tems, speedometers, tachometers, and other automobile equipment.

Equipment: For testing tachometer and speedometer indications
at varying temperatures, from —20 to 250* F. Sprague electric,
cradle type dynamometer, capacity 50 to 75 h. p. 4000 maximum
revolutions per minute ; torsion machines ; special flux meter for mag-
netic investigation work.

446. Stockham Pipe ft Fittings Co., Birmingham, Ala. (Cast iron
fittings.)

Research staff: R. E. Risley.

Research work: Full time of i on heat treatment of high speed
steel, molding sand selection and treatment and briquetting and re-
melting^ cast iron borings.

Equipment: Special equipment for physical testing of molding
sand.

INDUSTRIAL RESEARCH LABORATORIES 75

447. Stone & Webster^ Incorporated^ 147 Milk St., Boston, Mass.
(Engfineers, constructors, bankers, operators of public utilities.)

Research staff: 2 chemists, 2 mechanicians.

Research work: Full time of 4 on needs of industrial companies.

448. Strathmore Paper Company, Mittineague, Mass.
Research staff: Justus C. Sanborn and i assistant chemist.
Research work : One-fifth time of 2 on special paper mill problems.

449* Structural Materials Research Laboratory, Lewis Institute, 195 1
W. Madison St., Chicago, 111.

Research staff : Duff A. Abrams in charge of laboratory ; J. C. Witt,
chief research chemist, and 30 engineers, physicists and chemists.

Research work : Full time of 32 on the properties of concrete and
concrete materials, reinforced concrete and related topics. Research
is being carried on through a cooperative arrangement between the
Lewis Institute and the Portland Cement Association.

Equipment: One 300,000-pound, two 200,000-pound and one
40,000-pound screw-power universal testing machines, 20,000-pound
torsion testing machine, 4-unit Deval abrasion machine and standard
ball mill for tests of road materials, Ro-Tap sieve shaker for fineness
tests of materials, Talbot-Jones rattler for wear tests of concrete,
autoclave apparatus for high-pressure steam tests of cement.

450. Studebaker Corporation, The, Detroit, Mich. (Automobiles and
other vehicles.)

Research staff: E. J. Miles, 2 engineers and i mechanic in the
dynamometer department ; i electrical engineer and i assistant in the
electrical department; i chemist in the chemical department, i engi-
neer, I assistant and a staff of mechanics in the road testing depart-
ment, I engineer on special work.

Research work : One-half to two-thirds time of staff on power out-
put of motors, investigations of electrical appurtenances for automo-
biles, chemical studies of materials used in manufacture, road testing
of automobiles, special problems related to radiators, brakes, oil
pumps, fans and other equipment of an automobile.

Equipment : Research laboratory : 3 complete electric dynamome-
ter equipments for motors up to 80-horsepower output; completely
equipped for investigations of ignition apparatus, lighting and start-
ing apparatus, storage batteries and all other electrical appurtenances
of automobiles ; special equipment for investigating oils and grease.

451. Sun Chemical ft Color Co., 309 Sussex St., Harrison, N. J. (Dry
and pulp colors.)

Research staff: 2 chemists and i assistant.

Research work : One-half time of 3 on improving lake and pigment
colors.
45a. Swan-Myers Company, 219 N. Senate Ave., Indianapolis, Ind.

Research staff: Edgar B. Carter, director of biological division, A.
D. Thorburn, director of pharmaceutical division, 2 chemists and 5
bacteriologists and biological chemists.

Research work : Approximately one-fourth time of 9 on biological
products and organic synthetics used in medicine and pharmaceutical
products.

76 INDUSTRIAL RESEARCH LABORATORIES

453. Swenson Evaporator Company, 945 Monadnock Building, Chi-
cago, 111. Laboratory at Ann Arbor, Mich.

Research staff : W. L. Badger, i chemical engineer, assistants and
I helper.

Research work: Full time on design of evaporators and other
chemical engineering machinery ; trial of processes and theoretical re-
search on heat transmission in general.

Equipment: Large specially designed evaporators of all types.
Accessory equipment so that processes can be carried out on ton or
carload lots of material.

454. Swift ft Company, Chicago. 111.

Research staff: William D. Richardson and 9 assistants.

Research work: Full time of 10 on foods and dietetics, meat and
meat products, dairy products, oils and fats, soap and soap products,
glue and gelatin, fertilizers.

Equipment: Vacuum drying apparatus, agitator pressure tanks,
special chill rooms.
455* Tacony Steel Company, Philadelphia, Pa.

Research staff: H. A. Baxter and approximately 25 assistants.

Research work: On manufacture and use of special carbon and
alloy steels for high duty structural service.

456. Taggart and Yerza, 165 Division St., New Haven, Conn.
Research staff: Arthur F. Taggart, R. B. Yerxa, 3 chemists and 3

engineers.

Research work: Full time of 8 on flotation concentration of ores.

457. Takamine Laboratory, Inc., Takamine Bldg., 12 Dutch St., New
York, N. Y. (Manufactunng chemists.) Laboratory at Clifton, N. J.

Research staff: Jokichi Takamine, 4 chemists and i assistant.
Research work: Full time of 6 on biological, physiological and
organic chemistry.

458. Teeple, John E., 50 E. 41st St., New York, N. Y. (Consulting
chemist, chemical engineer.)

Research staff: John E. Teeple and 2 to 4 chemists.
Research work : Full time of 2 to 4 on investigations necessary for
directing research work in the laboratories of clients.

459. Telling-Belle Vernon Company, The, 3825 Cedar Ave., Cleve-
land, Ohio.

Research staff: W. O. Frohring, 2 bacteriologists and 2 chemists.
Research work : Three-fourths time of 5 on milk and milk products,
with large portion of time on ice cream and infant foods.

460. Tluic Industrial Products Corp., 58 Middle Rose St., Trenton,
N.J.

Research staff : A. I. Appelbaum and 2 assistants.

Research work : Part time of 3 on development of by-products.

461. Titanium Alloy Manufacturing Co., Niagara Falls, N: Y.
Research staff: L. E. Barton, chief chemist, 2 assistant chemists

and I helper. Physical testing laboratories, G. F. Comstock, metal-
lurgist and 2 metallographists.

Research work: On problems related to the manufacture and use
of ferro-carbon titanium and zirconium and zirconium products for
ceramic industries.

INDUSTRIAL RESEARCH LABORATORIES 77

46a. Titanium Pigment Co., Inc.» Niagara Falls, N. Y.

Research staff : L. E. Barton, chief chemist, research and technical
control of plant, and 4 assistant chemists.

Research work: On manufacture and use of titanium pigments,
titanium salts and other titanium products.

463. T. M. ft O. Chemical Co., 517 Cortlandt St., Belleville, N. J.
(Manufacturing chemists.)

Research staff: O. Ivan Lee and 3 assistants.

Research work : Approximately one-half time of 4 on development
of commercial processes for the manufacture of organic chemicals
with special reference to intermediates, dyes, and aromatic synthetics
for soaps and perfumes ; systematic study of the synthesis, separation
and purification of secondary and tertiary aromatic amines ; chlorina*-
tion products of aromatic hydrocarbons; and utilization of by-
products.

464. Toch Brothers, 320 Fifth Ave., New York, N. Y. (Paints, var-
nishes, colors, enamels ; acid, alkali and damp-proof coatings.)

Research staff : Maximilian Toch and 4 to 0 chemists.

Research work : • Problems related to water-proofing and protection
of Portland cement by inte^^l and surface coating methods ; water-
proofing of structural materials ; anti<<orro8ive paints and compounds.

465. Tolhurst Machine Works, Troy, N. Y. (Specialists in centrifu-
gals: hydro-extractors.)

Research staff: T. A. Bryson, usually i engineer and i or 2 as-
sistants.

Research work: One-sixth time of 3 on determination of profitable
methods of separation (and washing) of liquids from liquids or solids
by means of centrifugal force; apparatus for dewatering sewage
sludge ; separation of foots from oil, recovery of glycerine and salt in
soap industry, and improved methods of treating fish and fish oil.

Equipment: Centrifugal machines for filtration, extraction and
sedimentation, ranging from small hand-driven, tube and basket cen-
trifuges to higher speed 12 gallons basket capacity centrifugals, with
interchangeable baskets of various types for crystalline, granular or
fibrous materials, slimes and sludges.

466. Tower Manufacturing Co., Inc., 85 Doremus Ave., Newark, N. J
Research staff: C. P. Harris, 7 chemists and 2 engineers. .
Research work : Three-tenths time of 10 on processes for the manu-
facture of dyes and intermediates.

Equipment: Completely equipped semi-commercial plant.

467. Ultro Chemical Corporation, 41 Union Square, New York, N. Y.
(Colors and chemicals.) Laboratory at 236 40th St., Brooklyn, N. Y.

Research staff : A. £. Gessler, i chemist and i assistant.
Research work: Approximately full time of 3 on dry colors and
dyestuffs.

468. Underwriters' Laboratories, 207 E. Ohio St., Chicago, 111. Es-
tablished and maintained by National Board of Fire Underwriters.
Departments: Protection, electrical, gases and oils, chemical, cas-
ualty. Laboratory also at 25 City Hall Place, New York, N. Y.

Research staff : W. H. Merrill and 50 experts and necessary as-
sistants.

78 INDUSTRIAL RESEARCH LABORATORIES

Research work : A variable but large proportion of time of staff on
matters affecting performance and classincation devices, materials and
systems affecting the fire hazard or the personal accident hazard.
468a. Uniform Adhesive Company, Incorporated, foot of 39th St.,
Brooklyn, N. Y.

Research staff : Jerome and Walter Alexander.

Research work: Part time on adhesives, colloids, gums, starches,
colors, sizings, finishings, etc., for paper and textiles.

469. Union Carbide and Carbon Research Laboratories, Inc., Thomp-
son Ave. and Manley St., Long Island City, N. Y., a subsidiary of the
Union Carbide & Carbon Corporation, New York. Central Research
laboratory at Long Island City and branch research and development
laboratories at Long Island City and Buffalo, and two at Niagara
Falls, N. Y.; two at Cleveland and one at Fremont, Ohio; one each
at Indianapolis and Kokomo, Ind., and Clendenin, W. Va.

Research staff: Central laboratory has a staff of over 40, and
branches combined, over 30, including chemists, chemical, metallur-
gical and electrical engineers and physicists.

Research work : Full time of staff on metallurgical and other elec-
tric furnace products, calcium carbide, compressed gases, carbon
products, dry batteries and storage batteries, flashlights, organic
chemicals and equipment for using the above products.

Equipment: Electric furnaces of various types; alloy testing and
pyrometric equipment; gas compressing and testing equipment; ap-
paratus for making and testing dry batteries, storage batteries, arc
light carbons and brushes for electric motors and generators.

470. Union Switch ft Signal Company, Swissvale, Pa. (Railway sig-
nal equipment.) Materials laboratories are maintained separately
under the direction of H. C. Loudenbeck, with 3 chemists.

Research staff: L. O. Grondahl, 2 engineers and i assistant in
charge of standardizing laboratory.

Research work: Two-thirds time of 4 on development of iron for
electro-magnets, heat treatments, methods of test, electrical contacts,
insulators, impregnation of coils and of wood.

Equipment: Oscillographs; standardizing equipment for electrical
instruments; 50,000-volt insulation testing transformer; Heissler im-
pact testing machine ; an experimental impregnating plant, oil heated,
with vacuum and pressure pump ; and salt spray tester.

471. United Alloy Steel Corporation, Canton, Ohio. (Open hearth
and electric steels, bars, slabs, billets, blooms, universal plates.)

Research staff : M. H. Schmid, i metallurgical engineer, i assistant
metallurgical engineer, i laboratory foreman, 10 assistants and i en-
gineer of tests ; m the Electric Furnace, i chief and 2 recorders ; in
the Open Hearth Furnace, i chief and 8 recorders; in the Rolling
Mills, I chief and 4 recorders.

Research work: One-half time of 32 on investigations connected
with production and use of steel.

Equipment : Heat treatment : 4 Hoskins' electric furnaces, i Amer-
ican gas furnace for pieces up to 20 inches length and 5 inches diame-
ter. Physical testing: equipped for tensile, torsion, cold bend, vibra-
tory, Izod, Brinell, scleroscope, staybolt, etc. ; also Leeds & Northrup

INDUSTRIAL RESEARCH LABORATORIES 79

permeameter for determining magnetic permeability of steel and i
Leeds & Northrup recalescence instrument for determining critical
points of steel.

472. United Chemical and Organic Products Co., W. Hammond, 111.
(Successors to Hirsh, Stein & Company.)

Research staff: Jay Bowman and 4 chemists.
Research work : One-half time of 5 on problems arising in connec-
tion with plant processes.

Equipment: Semi-manufacturing scale equipment.

473. United Drug Company, Boston, Mass.
Research staff : Edward C. Merrill and 10 chemists.

Research work: One-half time of 10, largely on pharmaceutical
investigations and research, and independent problems covering mis-
cellaneous subjects.

474. United Gkis Improvement Co., The, 3101 Passyunk Ave., Phila-
delphia, Pa.

Research staff: Edward J. Brady and 3 assistants.

Research work: Problems dealing only with the manufacture,
purification, measurement and combustion of gas and the development
of instruments in connection with the above.

Equipment : Laboratory water gas plant ; laboratory blue gas gen-
erator; the use of a separate and complete commercial-sized experi-
mental plant available at times ; furnaces for refractory testing ; high
pressure gas equipment; complete physical equipment for high tem-
peratures; high gas pressures; evacuating spectroscopy; electrical
standards ; radiation measurements ; photometry and color.

475. United Shoe Machinery Corporation, Boston, Mass. Laboratory
at Beverly.

Research staff: Walter Gould Bullard and assistants.
Research work: Examination of raw materials; tests on core oils
and compounds, systematic investigation on improvement in antiseptic

Suality of cutting compounds and on pickling steel bars and plates,
ome work on reclamation of waste materials and in attempts to im-
prove methods of manufacturing shoe-factory supplies of all kinds.

476. United States Bronze Powder Works, Inc., Closter, N. J.
Research staff: Everett S. Landman and 2 chemical engineers.
Research work : One-fifth time of 3 on oxidation and reduction of

finely divided copper, properties and composition of bronzing liquids,
non-tamishable bronze powders and anti-fouling boat bottom com-
positions; pulverized copper and alloys for manufacture of electrical
brushes.

U. S. Conditioning and Testing Co. See U. S. Testing Co., Inc.
(p. 80).

477. U. S. Food Products Corp., Peoria, 111.
Research staff: J. K. Dale and 2 chemists.

Research work : Full time of 3 on food development problems.

478. United States Glue Co., Milwaukee, Wis.
Research staff : C. R. McKee and 3 trained men.

Research work : One-half time of 4 on improvements in technology
in glue and gelatine industry, particularly development of processes
to produce glue and gelatine for various specific purposes, such as

go INDUSTRIAL RESEARCH LABORATORIES

gelatine with various photographic properties, food gelatine, marsh-
mallow gelatine and special glue.
Equipment: Complete miniature glue and gelatine factory.

479. U. S. Industrial Alcohol Company, 27 William St., New York,
N. Y. Laboratory at South Baltimore, Md.

Research staff : A. A. Backhaus, 12 chemists, 2 bacteriologists, 10
assistant chemists and 2 chemical engineers.

Research work: Full time of staff on research in connection with
the development of alcohol products, utilization of by-products of
alcohol manufacture, improvement in the manufacture of alcohol,
study of yeasts and bacteria.

480. United States Metals Refining Co., Chrome, N. J.

Research staff: H. D. Greenwood, in charge of chemical depart-
ment, W. C. Smith in charge of metallurgical department; about 42
assistants.

Research work : Part time of staff on maintaining a high standard
in plant metallurgy and discovering new and improved methods.

481. United States Smelting, Rdi^g ft Mining Company, 55 Con-
gress St., Boston, Mass. (Silver, gold, lead, copper, zinc, iron, arsenic,
bismuth, cadmium, and tellurium.) Plants and research laboratories
located at various points in the United States and Mexico.

Research staff : Galen H. Clevenger and 20 engineers, chemists and
other specialists.

Research work : Full time on metallurgy, industrial chemistry and
mining in the development of new processes, improvements in ex-
istent processes, investigation of new processes submitted and ex-
amination and improvement of products.

Equipment: Thirty-liter-per-hour liquid oxygen machine, equip-
ment for investigating liquid oxygen explosives and for determining
the volatilization losses of the precious metals during melting, reduc-
ing kiln of 50 tons daily capacity, experimental bag house, and ex-
perimental farm for the study of the effect of smelter fume upon grow-
mg crops and animal life.

United States Steel Corporation. See Carnegie Steel Company
(p. 18).

482. U. S. Testing Co., Inc., 316 Hudson St., New York, N. Y.
Research staff: W. F. Edwards, 5 chemists, 3 engineers and i

physico-chemist.

Research work : One-half time of 10 on investigations of problems
arising in textile and allied industries.

Equipment : Apparatus for investigation of effect of light on dyed
textiles.

483. Universal Aniline Dyes and Chemical Co.» nth and Davis Sts.,
S. Milwaukee, Wis.

Research staff : A. H. Schmidt and 2 assistants.
Research work : Approximately one-half time of 3 on intermediates
and dyes.
Equipment: Complete miniature plant equipment.

484. Upjohn Company, The, Kalamazoo, Mich. (Fine pharmaceutic
cals.)

INDUSTRIAL RESEARCH LABORATORIES 81

Research staff : Frederick W. Heyl, 4 or 5 chemists, i pharmacolo-
gist, I bacteriologist.

Research work: Part time of 7 on estimation of nitroglycerine;
analyses of two Echinacea roots; standardization of commercial pa-
pain; some constituents of the roots of Brauneria augustifolia ; some
constituents of Sunbul root; standardization of the mercurials; Al-
genta root ; some constituents of jambul ; analysis of ragweed pollen ;
chemical examination of the leaves of Adonis vernalis; protein ex-
tract of ragweed pollen ; yellow coloring substance of ragweed pollen ;
some constituents of Viburnum prunifolium, stability of Digitalis leaf
extracts and infusions ; pharmacological action of Adonis vernalis,
485. Utah Copper Company, Deseret Bank Bldg., Salt Lake City,
Utah. Laboratory at Garfield. Utah.

Research staff: Thomas AJ Janney, 6 chemists, 4 engineers and
3 assistants.

Research work : Three-fifths time of 16 on treatment of ores by the
flotation process, gravity concentration, lixiviation and related inves-
tigations, flotation oils and reagents.
4M. UtEih-Idaho Sugar Company, Salt Lake City, Utah.

Research staff: E. G. Titus, i research assistant, i agricultural
chemist, and i laboratory assistant.

Research work: Approximately one-half time of 4 on agricultural
problems, beet-seed breeding, crop improvement, seed testing, soil
reclamation and analysis, fertilizer experiments, and insect^ disease
and weed control.

Equipment: Special beet testing machinery, seed germination ap-
paratus.

487. Utility Color ft Chemical Co., The, 395 Frelinghuysen Ave.,
Newark, N. J.

Research staff : Joel Taub and 2 assistants.

Research work: One-half time of 3 on development of colors.

488. Vacuum Oil Company, Incorporated, 61 Broadway, New York,
N. Y. (Refiners of petroleum for all purposes ; manufacturers of ship-
ping containers and of products used in the leather industry, etc.)
Works and laboratories at Rochester and Olean, N. Y., Paulsboro and
Bayonne, N. J.

Research staff : Florus R. Baxter, 3 chemists at Rochester ; i chem-
ist each at Olean, Paulsboro and Bayonne, also 12 assistants.

Research work: One-fifth time of 7 studying improvements in
manufacturing methods; causes of deterioration of oils in service;
utilization of by-products, properties of petroleum to determine suit-
ability for specific uses.

Equipment: Fire, steam and vacuum stills, lead lined agfitators,
fully equipped, wax presses, super-centrifuges, photomicrographic set,
apparatus for measurements of specific resistance, di-electric loss and
di-electric strength, etc..

489. Vanadium- Alloys Steel Co., The,^ Latrobe, Pa. (High speed,
alloy and carbon steels.)

Research staff: James P. Gill and 9 assistants.
Research, work : Approximately one-third time of 10 on alloy tool
steels, high speed and special steels.

82 INDUSTRIAL RESEARCH LABORATORIES

490. Vanadium Corporation of America, 120 Broadway, New York,
N. Y. Laboratory at Bridgcville, Pa.

Research staff: B. D. Saklatwalla, 5 chemists, i chemical engineer
and I electrochemical engineer.

Research work: One-half time of 8 on metallography of alloy
steels, development of metallurgical processes for alloying elements
and development of electro-thermic methods of reducing metals.
49Z. Van Schaack Brotiiers Chemical Works, Inc., 3358 Avondale
Ave., Chicago, 111. (Amyl acetate, soluble cotton, etc.)

Research staff : R. H. Van Schaack, Jr., and 4 assistants.

Research work: Approximately one-half time of 5 on nitrocellu-
lose solvents.

492. Ventura Refining Company, Fillmore, Calif.
Research staff : J. W. Weir and 7 assistants.

Research work : One-seventh time of 8 on petroleum refinery prob-
lems.

493. Vesta Battery Corporation, 2100 Indiana Ave., Chicago, 111.
(Storage batteries, auto dynamos, etc.)

Research staff: Chester M. Angell, i chemist, i battery engineer
and I assistant.

Research work: Approximately one-fourth time of 4 on electro-
chemistry, practical engineering features and improvement of parts
and materials used in manufacture of the lead plate storage battery.

494. Victor Chemical Works, Fisher Building, Chicago, 111. Large
laboratory for factory control and general work and two smaller ones
for research.

Research staff: L. D. Mathias, 5 chemists and i engineer.
Research work : Full time of 6 and one-half time of 2 on problems
connected with manufacturing activities.

495. Wadsworth Watch Case Co., Incorporated, The, Dayton, Ky.
Research work : Approximately full time of 5 on alloys of precious

metals and some of the brasses.

496. Wahl-Henius Institute, Incorporated, 1135 Fullerton Ave., Chi-
cago, 111.

Research staff : Max Henius, 3 experts, i chief analytical chemist,
I chief research chemist, 2 assistant chemists and 3 assistants.

Research work : Full time of chief research chemist and about one-
half time of I assistant chemist on fermentation and packing-house
problems.

Equipment : Apparatus for testing products of fermentation indus-
tries and for carrying out experimental work on semi-commercial
scale (experimental brewery, bottlery, etc.). Apparatus for testing
solid and liquid fuel, and lubricants; differential refractometer
(Tomoc's).

497. Wallace ft Tieman Co., Inc., Box 178, Newark, N. J. (Chlorine
control apparatus.)

Research staff: C. F. Wallace, G. C. Baker, 3 chemists and 2
engineers.

Research work: Lar^ part time of 7 on chlorine gas control and
applications in sterilization, bleaching and other lines ; flour bleaching ;
carburetor laws and mechanical applications; and food products.

INDUSTRIAL RESEARCH LABORATORIES 83

Equipment: Carburetor testing outfit complete and chlorine con-
trol equipment.

498. Wallace, Joseph H., ft Co., 5 Beekman St., New York, N. Y.
(Industrial engineers.) Laboratory at Webbs Hill, Stamford, Conn.,
R. F. D. 29.

Research staff: F. E. Greenwood, i consulting engineer and i
chemist.

Research work : Full time of 3 on cellulose and by-products, pulp,
paper, naval stores, etc.

Equipment : Semi-commercial plant for pulp, paper and by-products.

499. Waltham Watch Company, Waltham, Mass.
Research staff : F. P. Flagg and 3 chemists.

Research work: Full time of 2 on investigation of the properties
of enamel used on watch dials and study of the properties of metals
and their relation to watch production.

500. Warner, William R., ft Company, Incorporated, 113 W. i8th St.,
New York, N. Y. (Manufacturing pharmaceutists.) v

Research staff: Frederick J. Austin, Charles Costa.

Research work: Chemical and pharmaceutical research which has
for its object the improvement of products as regards physiological
activity, permanence, elegance, etc., together with original work lead-
ing to the development of new preparations and new methods of manu-
facture.

Warren, S. D., Co. See Cumberland Mills (p. 24).
50Z. Washburn-Crosby Co., Minneapolis, Minn. (Flour mills.)

Research staff : Frank W. Emmons, 3 chemists, i specially trained
physical laboratory man, i expert baker and various assistants.

Research work : Full time of i on problems relating to wheat flour.

502. Wayne Oil Tank and Pump Co., Ft. Wayne, Ind. (Tanks,
pumps and underground storage outfits.)

Research staff: R. E. Langston, i chemist and 2 engineers.

Research work : Approximately three-fourths time of 3 on devising
improved methods of handling, storing and using volatile and non-
volatile liquids, such as gasoline, paint oil, varnish, lubricating oil,
fuel oil, kerosene, etc. ; methods of reclaiming used auto oil and puri-
fication of used engine oil.

503. Wedge Mechanical Furnace Company, 1000 Widener Bldg.,
Philadelphia, Pa. (Roasting furnaces.) Laboratory at Greenwich
Point, Philadelphia, Pa.

Research staff: Carl S. Fogh and a variable number of assistants.
Research work: Full time on roasting ores, concentrates, mattes,
mixtures and various materials for smelters and chemical plants.

504. Weld and Liddell, 2 Rector St., New York, N. Y. (Consulting
engfineers.) Laboratory at 961 Frelinghuysen Ave., Newark, N. J.

Research staff: Donald M. Liddell, 3 trained men and 2 untrained
assistants.

Research work : Variable amount of time of 6 on stucco, zinc oxide,
oil shale and petroleum. Balance of time on research problems of
The Gray Industrial Laboratories.

Equipment: Completely equipped for pressure and steam distilla-
tions on oil shales or any bituminous or oily products.

S4 INDUSTRIAL RESEARCH LABORATORIES

505. Wellt, Raymond, Homer, N. Y. (Chemist and technologist.)
Research staff: Raymond Wells and 2 assistants.

Research work : One-half time of 3 on animal and vegetable oils ;
fertilizers, soap, candles and glycerine ; abattoir by-products ; garbage
and sewage disposal ; lubrication oils and greases ; wire mill soaps and
drawing compounds; textile soaps and oils and agricultural insecti-
cides and fungicides.

Equipment: Commercial scale equipment for research in oils, etc.

506. Welsbach Company, Gloucester, N. J. (Mantles for illuminating
gas.)

Research staff: Harlan S. Miner and 6 trained men.

Research work : One-half time of 7 directed especially to economic
production of rare earth chemicals, especially thorium and cerium;
manufacture of special rare earth salts, nitration of cellulose, produc-
tion of mesothorium ; radio-chemistry.

Equipment: Especially for the study of problems connected with
development of incandescent gas mantles.

507. Western Electric Company, Incorporated, 463 West St., New
York, N. Y., known as the Research Laboratories of the American
Telephone and Telegraph Company and the Western Electric Com-
pany, conducts research and engineering activities for the Bell Tele-
phone System.

Research staff: F. B, Jewett, chief engineer, E. B. Craft, K H.
Colpitts and W. F. Hendry, assistant chief engineers ; heads of func-
tional activities : H. D. Arnold, J. J. Lyng, R. L. Jones, A. F. Dixon,
J. W. Harris, L. Keller, H. C. Snook, J. B. Harlow, G. A. Aoderegg
and H. E. Shreeve who have under their direction approximately 825
research physicists, chemists and engineers, and approximately 750
assistant engineers, draftsmen, etc.

Research work: Full time of 1575 devoted to original investiga-
tion and development of new forms and improvement of existing
forms of apparatus and equipment for electrical communication. The
problems include research m thermionic emission and conduction,
vacuum tube performance, microphonic conduction, radio transmis-
sion, the physical basis of speech, wave and impulse propagation and
the physical and chemical properties of a great variety of materials ;
the development and design of full-mechanical and semi-mechanical
telephone switchboards and systems in preparation for a comprehen-
sive service transformation from the present manually operated sys-
tem; the development and design of high frequency carrier systems
with their associated generators, oscillators, modulators, hybrid coils,
repeaters, loading coils, demodulators, amplifiers, and other special
apparatus; development of new forms of local and long distance
cables, submarine cables, transmitters, receivers, automatic printing
telegraph apparatus, lightning arresters, protective fuses, current rec-
tifiers, ringers, ringing systems, precision apparatus for high fre-
quency measurements, marine radio sets, portable radio sets, trans-
mitter life test methods, test methods for transmission efficiency, dry
cells, storage cells, farm-light sets, household appliances and numer-
ous problems in the design of keys, cords, plugs, switches, relays, con-
tacts, loading coils, impedance coils, repeaters, transformers, con-

INDUSTRIAL RESEARCH LABORATORIES 85

densers, insulators, lamps, and kindred details of communication ap-
paratus and systems.

A thirteen story building of 400,000 square feet floor area. Physical
research laboratory, transmission research laboratory, chemical re-
search laboratory and physical testing laboratory completely equipped
with all facilities necessary for this work; also completely equipped
shop for the construction of working models and special equipment
used in conducting research and development work.

508. Western Gas Construction Company, The, 1429 Buchanan St.,
Ft. Wayne, Ind. (Designers and builders of water, coal and gas ap-
paratus, gas holders and special equipment.)

Research staff: F. Salathe, 2 chemists and 3 engineers.
Research work : One-fourth time' of 6 on oils, gas, general organic,
mechanical and chemical engineering.

509. Western Precipitation C<mipany, 1016 W. Ninth St., Los An-
geles, Calif. (Chemical engineers.)

Research staff : H. V. Welch, i physicist, i engineer, 3 chemists.

Research work: Three-fourths to nine-tenths time of 6 on prob-
lems centering around the Cottrell Processes of electrical precipitation.

Equipment: i5o,ooo»Yolt transformer, 50,000-volt direct current
generator, high potential mechanical rectifiers, potash laboratory, di-
gestion and filtration apparatus and special apparatus adapted for
study of equilibrium conditions in solutions.

510. Western Research Corporation, Incorporated, 514 i8th St., Den-
ver, Colo.

Research staff : James M. McClave, i chemist and i oil chemist.

Research work: One-half time of 3 on investigation of minerals
and non minerals, oils and shales; special attention to working out
treatment methods and the construction of plants.
51 z. Western Sugar Refinery, Foot 23d St., San Francisco, Calif.

Research staff: S. C. Meredith, i chief chemist, 3 engineers and
3 assistant chemists.

Research work : Two-fifths time of 8 on investigations of sugar
losses, sugar machinery and materials.

5x2-5x3. Westinghouse Electric ft Manufacturing Company, East
Pittsburgh, Pa. (Electrical apparatus of all kinds.) Laboratory also
at Essington, Pa.

512. East Pittsburgh Laboratory

Research staff, scientific: C. £. Skinner, manager of research de-
partment, 10 chemists, 28 physicists, and an operating staff of 23, in-
cluding plant engineer, office staff, glass blowers, instrument makers,
etc.

Research work: Chemical division, organic materials, inorganic
chemical research and analytical chemistry; division of physics and
metallurgy, magnetic testing and research, magnetic materials, metal-
lurgical preparations, metallurgical testing and research, electrolytic
condensers, power condenser research, insulating materials, electrical
porcelain, radio bulbs, thermal conductivity and expansivity, resister
materials, etc.

Equipment: Electric furnaces and rolls for metallurgical prepara-
tions, high vacuum apparatus, special magnetic testing apparatus.

86 INDUSTRIAL RESEARCH LABORATORIES

thermal conductivity and expansivity apparatus and conductivity of
dielectrics.

Research staff, technical : R. P. Jackson, manager of materials and
processes department, 8 chemists, 2 physical test men, 10 specialists
on materials and their uses and 35 technical and other assistants.

Research work : 5 research laboratories in which work is conducted
on technical problems connected with manufacture and testing of raw
materials and finished products.

Equipment: High tension testing, special oscillographs and test*
ing machines for determining physical properties.

Standard house: O. B. Riley with staff of about 11 engaged in
checking and testing standard instruments and apparatus, chiefly
electrical.

513. Essington Laboratory

Research staff : A. T. Kasley and 4 assistants.

Research work: Problems connected with heat and power.
5x4. Westinghouse Lamp Co.« Bloomfield, N. J. Engineering and de-
velopment laboratories under the direction of R. E. Myers with a
staff of 85.

Research staff: H. C. Rentschler, 3 physicists, 3 assistant physi-
cists, 2 chemists and i assistant chemist.

Research work: Full time of 9 on study of radiation from solids
and gases and vapors; also high vacua phenomena.

Equipment : Apparatus for obtaining and measuring high vacua,
for producing high potential rectified current and for photometric and
optical pyrometer measurements. High frequency electric furnace.
Liquid air available at all times. Rare gases for study of their prop-
erties and uses are available.

515. Weston, Byron, Co., Dalton, Mass. (Ledger and record paper.)
Research staff : P. W. Codwise and i assistant.

Research work : Varying amount time of 2 on problems connected
with paper making.

516. Weston ft Sampson, 14 Beacon St., Boston, Mass. (Consulting
engineers; water sanitation.)

Research staff: Robert S. Weston, i chemist and i engineer.
Research work: Part time of 3 on water, sewage and sanitation.

517. Wheeler ft Woodruff, 280 Madison Ave., New York, N. Y.
Research staff : T. L. Wheeler, 2 chemical engineers, i mechanical

engineer, i draftsman, i chemist and i helper.

Research work : One-half time of 7 on manufacture and regenera-
tion of bone black, manufacture, use and revivification of decolorizing
carbons, production and use of gas absorbing carbons, manufacture
and treatment of carbon black, refining, deodorizing and hydrogena-
tion of vegetable oils, filtration of water, manufacture of hydrogen,
electroplating and electro-chemical problems and corrosion of metals
and metal finishes.

Equipment: Semi-commercial size apparatus for study of carbon,
etc.

518. White Tar Company of New Jersey, Inc., The, 56 Vesey St..
New York, N. Y. (Chemicals.) Laboratory at Newark, N. J.

Research staff : Herbert W. Hamilton and i assistant.

INDUSTRIAL RESEARCH LABORATORIES 87

Research work : Approximately one-third time of 2 on purification
of naphthalene and the development of sanitary products.

519. Whitten, J. O., Company, The, Cross St., Winchester, Mass.
(Gelatines.)

Research staff: G. R. Whitten and 3 assistants.
Research work : One-half time of 4 on treatment of bone and hide
preparatory to the manufacture of gelatine.

520. Wilbur White Chemical Co., The, 62 Temple St., Owego, N. Y.
Research staff : J. A. Bridgman and 3 chemists.

Research work: One-half time of 4 on new processes for organic
intermediates.

52Z. WUckes, Martin, Wilckes^ Company, Head of Pine St., Camden,
N. J. (Lampblacks, carbonblacks, etc.)

Research staff : A. Malmstrom, 4 chemists and i engineer.

Research work: Full time of i or 2 chemists on phosphoric acid
compounds and baking powders.

522. Wiley ft Company, Inc., 904 N. Calvert St., Baltimore, Md.
(Analytical and consulting chemists.)

Research staff : Samuel W. Wiley and 7 assistants.

Research work : Full time of i and part time of others on problems
connected with the fertilizer industry ; cellulose and paper ; coal, oils
and coke; beverages.

523. Wilson ft Co., Chicago, 111. (Packers and provisioners.) Lab-
oratories at Chicago, 111., Chattanooga, Tenn., Oklahoma City, Okla.,
and Kansas City, Kansas.

Research staff: L. M. Tolman and 5 assistants.

Research work : One-half time of 6 on problems connected with
fermentation, spoilage, etc.; hydrogenation of oils, refining and han-
dling of oils and by-products.

524. Winchester Repeating Arms Co., New Haven, Conn. (Rifles,
shotguns, small arms ammunition, fishing tackle, skates, cutlery, flash-
lights and tools.)

Research staff : J. S. Gravely, 4 research chemists, 2 metallurgists
and metallographists, 2 electrochemists and engineers and 8 assistants.

Research work: Three-fifths time of 17 on materials and processes
involved in the manufacture of small arms and ammunition, cutlery,
tools, hardware and sporting goods, dry batteries, flashlights, etc.

525. Zinsser ft Co., Hastings-on-Hudson, N. Y. (Manufacturing
chemists.)

Research staff: J. S. Zinsser, 5 chemists, i dyer and i analyst.
Research work : Full time of 8 on anthraquinone color work.

526. Zobel, Ernst, Company, Inc., 104 2d Ave., Brooklyn, N. Y. (Dis-
tillers and manufacturers of pine products and coal tar products ; ad-
hesive pitch, etc.)

Research staff: F. C. Zobel and 2 assistants.

Research work: Asphaltum, resin and oil; and coal tar distillate.

INDUSTRIAL RESEARCH LABORATORIES

INDEX TO SUBJECT CLASSIFICATION OF LABORATOiaBS

PAGX

^B R A S I V E S (carbonindum,
emery, grinding, polishing, sand-
paper) 94

Acetylene, see gas, fuel and illumi-
nating 109

Acids, see chemicals, heavy 97

Acoustics, see sound 118

Adding machines, see office equip-
ment 115

Adhesives (glue, paste, sizing) 94

Aeronautics, see aircraft 94

Agitators, see chemical engineering

equipment 96

Agricultural equipment and engi-
neering (land drainage, threshing

machines, tractors) 94

Agricultural problems (entomology,
genetics, pathology, etc See also

soils and fertilizers) 94,1 18

Air (air-driven machines, air prod-
ucts, compressed air, liquid air,

pneumatics) 94

Air conditioning (ventilation) 94

Aircraft and accessories (see also in-
ternal combustion motors) . . . • 94,1 1 1
Alcohol, see fuels; see also chemi-
cals fine (including solvents) and

liquors 97,108,112

Alimentary pastes, see foods 107

Alkalies, see chemicals, heavy 97

Aluminum, see non-ferrous metals.. 114
Ammunition, see military and naval

equipment 114

Ammeters, see electrical equipment. . 105
Apparatus and instruments, chemi-
cal and physical (astronomical in-
struments, autoclaves, balances,
compasses, gages, lenses, micro-
scopes, survejring instruments, tele-
scopes, transits) 94

Argon, see gases, except fuel and

illuminating 109

Armor, see military and naval equip-
ment 114

Artificial ice, see refrigeration 117

Asphalt, see building materials 95

PAGE

Astronomical instruments, see appa-
ratus and instruments 94

Autoclaves, see apparatus and in-
struments 94

Automobiles, Ke automotive vehicles 95

Automotive vehicles, equipment and
accessories (automobiles, tanks,
tractors, trucks) 95

BACTERIOLOGY, see chemistry,

biological 98

Bakelite, see plastics 116

Bakery, see foods 107

Baking powder, see foods 107

Balances, see apparatus and instru-
ments 94

Ball bearings, see mechanics, general 113
Bearing metals, see non-ferrous

metals 114

Bearings, see mechanics, general. ... 113

Beer, see liquors 112

Beverages, noo-alcoholic 95

Biological equipment and suppHes. . . 95
Biokigy, see chemistry, biological. ... 98

Biscuit, see foods 107

Blowers, see chemical engineering

equipment 96

Boilers, see fuel utilization; see also

steam power 109,118

Boots and shoes, including machin-
ery, see leather Ill

Bottle seals, see containers 104

Brass, see non-ferrous metals 114

Bricks, see ceramics 96

Bronze, see non-ferrous metals 114

Building materials (asphalt, cement,
concrete, lime, marble, road mate-
rials, slate. See also iron and

steel) 95,111

Butter, see foods 107

Buttons, see textiles. 119

By-products from wastes 96

QABLE, see electrical conununica-

tion; see also insulation 105,110

Calorimetry, see heat HO

Cameras, see photography 116

INDUSTRIAL RESEARCH LABORATORIES

»

PAGE

Candy, see foods 107

Cannrng and jireserviiig, see foods.. 107

Cans, see OQotainers 104

Carbon, ace chemistry, inorganic; see

also hibrkants .99,112

Carbomndmi, see abrasives 94

Cars, see raibroad equipment 117

Cash registers, see office equipment. 115
Casting, see foundry equipment; see

also plastics 106, 116

Cast iron, see iron and steel Ill

Cellulose, see pulp and paper 117

Cement, see building materials 95

Centrifuges, see chemical engineermg

equq>ment 96

Ceramics (bricks, china, glass, mag-
nesite, pottery, porcelain, refrac-
tories) 96

Charcoal, see fuels 108

Chemical engineering equipment
(agitators, blowers, centrifuges,
compressors, concentrators, con-
densers, dryers, evaporators, filter
presses, pulverizers, pumps, sepa*

rators) 96

* Chemicals, fine, including solvents.. 97
Chemicals, heavy (acids, alkalies,

fungicides, insecticides, salts).... 97
Chemistry, biological (bacteriology,

biology) 98

Chemistry, inorganic (carbon, graph-
ite, etc.) 99

Chemistry, mineralogical and geo-

U^gical (quartz, etc.) 100

Chemistry, organic (fermentatkm,

starch, vegetable oils, etc.) 100

Chemistry, pharmaceutical (cos-
metics, dentifrice, drugs, disinfec-
tants, medicines) 102

China, see ceramics 96

Chlorine, see gases, except fuel and

illuminating 109

Classi6ers, see metallurgy and

metallography 113

Gothing, see textiles 119

Coal, sec fuels 108

Coke, see fuels 108

Cold storage, see foods 107

Omipasses, see apparatus and in-
struments 94

PAGE

Compressed air, see air 94

Compressors, see chemical engineer-
ing equipment 96

Concentration of ores (see also

chemical engineering equipment )96, 102
Concentrators, see chemical engi-
neering equipment 96

Concrete, see building materials 95

Condensers, see chemical engineer-
ing equipment 96

Condensite, see plastics 116

Consulting research laboratories 103

Containers, including bottle seals

(cans, fiber-board containers, etc.) 104
Copper, see non-ferrous metals.... 114

Cordage, see insulation 110

Cosmetics, see chemistry, pharma-
ceutical 102

Cdtton and its products, see textiles. 119
Cutlery, see machine tools and hard-
ware 112

J^^^'^^L equipment and supplies,
see surgical, dental and hospital

equipment and supplies 119

Dentifrice, see chemistry, pharma-
ceutical 102

Developers, see photography 116

Die casting, see foundry equipment. . 108
Diesel engines, see internal combus-
tion motors • Ill

Disinfectants, see chemistry, phar-
maceutical 102

Drill-press, see machine tools. ...». 112
Drugs, see chemistry, pharmaceuti-
cal 102

Dryers, see chemical engineering
equipment; see also paints, oib

and varnishes. 96, 115

Dyes, natural and artificial (mks,
intermediates, pigments, ribbons) . 104

Dynamite, see explosives 106

Dynamos, see electric power 105

£^CONOMIZ£RS, see steam power. 118
Electrkal communication (cable,
telegraph, telephone, wireless) .... 105

Electrical equipment and instruments
(ammeters, lamps, voltmeters,
wattmeters) 105

90

INDUSTRIAL RESEARCH LABORATORIES

PAGE

Electricity, general (economics, util*
ization) 105

Electric power (conversion, distribu-
tion, dynamos, generation, motors,
power plants, transmission) 105

Electrochemistry (electrochemical
processes, electrodes, storage bat-
teries) 106

Electrodes, see electrochemistry 106

Electro-plating 106

Emery, see abrasives 94

Enamels, see paints, oils and var-
nishes 115

Engines, see steam power; see also
internal combustion motors... Ill, 118

Entomology, see agricultural prob-
lems 94

Evaporators, see chemical engineer-
ing equipment 96

Explosives and explosions (dyna-
mite, powder, TNT) 106

Extinguishers, see fire prevention... 107

pATS, fatty oils and soaps 106

Fermentation, see chemistry, or-
ganic 100

Ferrous alloys, see iron and steel.. Ill
Fertilizers, see soils and fertilizers.. 118
Fiber -board containers, see con-
tainers 104

Films, see photography 116

Filter presses, see chemical engineer-
ing equipment 96

Filtration 107

Fire prevention (extinguishers,

sprinklers) 107

Fittings, see metal manufactures,

miscellaneous 113

Flavoring extracts, see f oodst 107

Flour, see foods 107

Foods (alimentary pastes, bakery,
baking powder, biscuit, butter,
candy, canning and preserving,
cold storage, flavoring extracts,
flour, gelatine, meat and meat
products, milk, oils, preservatives,

wheat, yeast, etc.) 107

Foundry equipment, materials and
methods (casting, die casting,
moulding) 108

PAGB

Fuels (alcohol, charcoal, coal, coke,
gasoline, kerosene, oil, peat. See
also gas, petroleum and wood)... 108

Fuel utilization (boilers, furnaces,
gas-producers, radiators, stokers) . 109

Fungicides, see chemicals, heavy. ... 97

Furnaces, see fuel utilization 109

Q-AGES, see apparatus and instru-
ments 94

Gas, fuel and illuminating, including

mantles (acetylene, hydrogen) .... 109
Gases, except fuel and illuminating,
including generating apparatus
(argon, chlorine, helium, neon,
nitrogen, oxygen, poisonous gases) 109

Gasoline, see fuels 108

Gasoline engines, see internal com-
bustion motors Ill

Gas-producers, see fuel utilization. . . 109

(jelatine, see foods 107

Glass, see ceramics. 96

Glue, see adhesives 94

(jold, see non-ferrous metals 114

Graphite, see chemistry, inorganic;

see also lubricants 99, 112

Graphophones, see phonographs and

graphophones 116

Grinding, see abrasives 94

Gutta-percha, see rubber and rubber
goods 117

fjAIR, curled, etc 110

Hardware, see machine tools and
hardware 112

Heat (calorimetry, pyrometry, ther-
mal physics, thermometry) 110

Heating 110

Helium, see gases, except fuel and
illuminating 109

Hospital equipment and supplies, see
surgical, dental and hospital equip-
ment and supplies 119

Hydraulics (waterworks, water
power) 110

Hydrogen, see gas, fuel and illumi-
nating 109

JLLUMINATION, electric, gas and

other 110

Inks, see dyes 104

INDUSTRIAL RESEARCH LABORATORIES

91

PAGE

Insecticides, see chemicals, heavy. . . 97
Insulation, electrical and thermal
(cable, cordage, non-conductors,

insulated wire) 110

Intermediates, see dyes 104

Internal combustion motors (Diesel
engines, gasoline engines, motors,

oil engines) Ill

Iron and steel (cast iron, ferrous al-
loys, pipe, wrou{;ht iron) Ill

KEROSENE, see fuels 108

X^ACQUERS, see paints, oils and

varnishes 115

Lamps, see electrical equipment; see

also illumination 96, 110

Land drainage, see agricultural

equipment and engineering 94

Lathes, see machine tools 112

Lead, see non-ferrous metals 114

Leather and leather goods (boots,
shoes, including machinery, leather

substitutes, tanning) Ill

Lenses, see apparatus and instru-
ments ; see also light 94, 112

Light (optical instruments, optics.

See also illumination) 1 10, 112

Lime, see building materials 95

Linen, see textiles 119

Liquid air, see air 94

Liquors, fermented and distilled (al-
cohol, beer, wine) 112

Locomotives, see railroad equip-
ment 117

Lubricants (carbon, graphite, oil,
petroleum) 112

lyfACHINE tools and hardware
(cutlery, drill - presses, lathes,

planers, shapers) 112

Magnesite, see ceramics 96

Magnetism 112

liiantles, see gas, fuel and illu-
minating 109

Marble, see building materials 95

Marine engineering (ships) 112

Matches 113

Meat and meat products, see foods. . 107

PAcn

Mechanics, general (bearings, ball,
roller, etc.) 113

Medicines, see chemistry, pharma-
ceutical 102

Metal manufactures, miscellaneous
(fittings, pipes, valves) 113

Metallurgy and metallography, in-
cluding equipment 113

Microscopes, see apparatus and in-
struments 94

Military and naval equipment (am-
munition, armor, ordnance, small
arms, torpedoes) 114

Mining, general (testing drills,
ropes, tools; ore dressing) 114

Motors, see electric power; see also
internal combustion motors... 105, HI

Moulding, see foundry equipment;
see also plastics 108, 116

Moving-picture equipment, see pho-
tography 116

j^ATURAL gums, see rubber and

rubber goods 117

Neon, see gases, except fuel and

illuminating 109

Nickel, see non-ferrous metals 114

Nitrates, see soils and fertilizers... 118
Nitrogen, see gases, except fuel and

illuminating 109

Non-conductors, see insulation 110

Non-ferrous metals (aluminum,
bearing metals, brass, bronze, cop-
per, gold, lead, nickel, platinum,
silver, tin, titanium, zinc) 114

Office equipment (adding ma-
chines, cash registers) 115

Oil engines, see internal combustion

motors Ill

Oils, see fats, foods, fuels, lubri-
cants, paints).. . 106, 107, 108, 112, 115

Optical instruments, see light 112

Optics, see light 112

Ordnance, see military and naval

equipment 1 14

Ore dressing, see mining, general. . . 114
Oxygen, see gases, except fuel and
illuminating 109

92

INDUSTRIAL RESEARCH LABORATORIES

FACE

PAINTS, oila and vmraislies (dry-
ers, enasncU, lacqtwrs, pignentt,

putty, resins, rust-proofiing) 115

Pltper, see pulp and paper 117

Paste, see adhesives 94

Peat, see fuels 108

Petroleum and its products (see also

lubricants) 112, 115

Phonoffraphs and grapbophones H^^

Phosphates, see soils and fertiUaers. 118
Photograplqr (cameras, developers,

films, movingwpicture equipment,

plates) 11<^

Pigments, see dyes; see also paints,

oils and varnishes 104, 115

Pipe, see iron and steel; see also

metal manufactures* misc Ill, 113

Planer, see machine tools 112

Plant genetics and patfudogy; see

agrknhural proUema. ^

Plastics (bakelite, condenstte, red-

manol; fat*"^g and moulding of

phutics) "*

Plates, see photography H^

Platinum, see noii*ferrous metato... H^

Pneumatics, see air ^

Poisonous gases, see gases, except

fuel and illuminatmg 1^

Polishing, see ahrasivcs •*

Porcefadn, see ceramics ^

Potash, see soils and fertilixers; see

also chemicals, heavy 118, 97

Pottery, see ceramics 96

Powder, see explosives 106

Power plants, see electric power. ... 105

Preservatives! see foods 107

Properties of engineering materials. 116

Public utilities 117

Pulp and paper (cellulose) 117

Pulverisers, see chemical engineer-
ing equipment 96

Pumps, see chemical engineering

equipment 96

Putty, see paints, oils and varnishes. 115
Pyrometry, see heat 110

QUARTZ, see chemistry, mineral-
ogical and geological 100

FACE

RADIATORS, see fuel ultlisatioa. 109
Radio, see electrical communica**
tion; see also subatomic phe-
nomena 105, 118

Railroad equipment (cars, loeomo-

tives, signals, etc.) 117

Razors 117

Reagenu, see biological equipment

and supplies 95

Redmanol, see plastics 116

Refractories, see ceramics 96

Refrigeration (artificial ice) 117

Resins, see paints, oils and varnishes 115

Ribbons, typewriter, see dyes 104

Road materials, see building mate-
rials 95

Roasters, see metallurgy and metal-
lography 113

Rubber and rubber goods, including
other natural gums (gutta-
percha) 117

Rust-proofing, see paints, oils and
varnishes 116

Salts, see chemicals, heavy 97

Sandpaper, see abrasives 94

Sanitation* sec water, sewage and

sanitation 119

Separators, see chemical engineering

equipment 96

Sewage, see water, sewage and sani-
tation 119

Shaper, see machine tools 112

Ships, see marine engineering 112

Shoes and boots, including machin-
ery, see leather Ill

Signals, see railroad equipment 117

Silver, see non-ferrous metals 114

Sizing, see adhesives; see also pulp

and paper 94, 117

Slate, see building materials 95

Small arms, see military and naval

equipment 114

Soaps, see fats 106

Soils and fertilisers (nitrates, phos-
phates, potash) 118

Solvents, see chemicals, fine 97

Sorghums, see sugar 118

INDUSTRIAL RESEARCH LABORATORIES

93

PAGB

Sound (acoustics) 118

Spriaklers, see fire prevention 107

Starch, see chemistry, ors^nic; see

also foods 100, 107

Steam power (boilers, economizers,
en^es, turbines. See also inter-
nal combustion motors) 118, 111

Steel, see iron and steel Ill

Stokers, see fuel utilization 109

Storage batteries, see electrochemis-
try 106

Subatomic phenomena and radio-
activity 118

Sugar (sorghums, syrups) 118

Surgkal, dental and hospital equip-
ment and supplies 1 19

Surveying instruments, see appa-
ratus and instruments 94

Syrups, see sugar 118

fANKS, see automotive vehicles.. 95
Tanning, see leather llf

Tar and its products 119

Telegraph, see electrical communica-
tion 105

Telephone, see electrical communica-
tion 105

Telescopes, see apparatus and in-
struments 94

Textiles, including machinery (but-
t<»is, clothing, cotton and its prod-
ucts, linen, wool ; waterproofing) . . 1 19

Thermal physics, see heat 110

Thermometry, see heat 110

Threshing machines, see agricultural

equipment and engineering 94

Tin, see non-ferrous metals 114

Titanium, see non-ferrous metals 114

TNT, see explosives 106

PAGE

Torpedoes, see military and naval
equipment 114

Tractors, see agricultuial equip-
ment and engineering; see also
automotive vehicles 94, 95

Transits, see apparatus and instru-
ments 94

Trucks, see automotive vehicles 95

Turbines, see steam power 118

V^ALVES, see metal manufactures,
miscellaneous 113

Varnishes, see paints, oils and var-
nishes 115

Vegetable oils, see chemistry, or-
ganic; see also foods 100, 107

Ventilation, see air conditioning 94

Voltmeters, see electrical equip-
ment 105

"VV ATER, sewage and sanitation. . 119
Water power, see hydraulics.. 110

Waterproofing, see textiles 119

Wattmeters, see electrical equip-
ment 105

Welding, autogenous, gas, electric,

forge 120

Wheat, see foods 107

Wine, see liquors 112

Wire 120

Wireless, see electrical communica-
tion 105

Wood products, other than cellulose
and paper (sec also containe 104, 120

Wool, see textiles 119

Wrought iron, see iron and steel. ... Ill

YEAST, see foods

107

^INC, see non-ferrous metals 114

94

INDUSTRIAL RESEARCH LABORATORIES

SUBJECT CLASSIFICATION OF LABORATORIES

Abrasives (carborundum, emery,
grinding, polishing, sandpaper)

Armour Glue Works

Armour Sandpaper Works

Bausch & Lomb Optical G>.

Carbortmdum G>mpany, The

Dorr Company, The

Gillette Safety Razor Co.

Kalmus, Comstock & Wescott, Inc.

Maynard, T. Poole

Metals & Chemicals Extraction
Corporation

Union Carbide and Carbon Re-
search Laboratories, Inc.

Adhesives (glue, paste, sizing)

Abbott, William G., Jr.

Armour Glue Works

Banks & Craig

Bausch & Lomb Optical Co.

Bloede, Victor G., Co.

Brunswick-Balke-Collender Co.

Carborundum Company, The

Cudahy Packing Co., The

Cumberland Mills

Dewey & Almy Chemical Com-
pany

Dextro Products, Inc.

Emerson Laboratory

Feculose Co. of America

Grosvenor, Wm. M.

Little, Arthur D., Inc.

Morris & Company

National Gum & Mica Co.

Pfister & Vogcl Leather Co.

Philadelphia Quartz Company

Seydel Manufacturing Company

Skinner, Sherman & Esselen, In-
corporated

Swift & Company

Thac Industrial Products Corp.

Uniform Adhesive Company, In-
corporated

United Chemical and Organic
Products Co.

U. S. Food Products Corp.

United States Glue Co.

Zobel, Ernst, Company, Inc.

Agricultural equipment and engi-
neering (land drainage, thresh-
ing machines, tractors)

American Beet Sugar Company

Banks & Craig

Minneapolis Steel and Machinery

Co.
Utah-Idaho Sugar Company

Agricultural problems (entomol-
ogy, genetics, pathology, etc.
See also soils and fertilizers)

American Agricultural Chemical

Company, The
American Beet Sugar Company
National Lime Association
Utah-Idaho Sugar Company

Air (air - driven machines, air
products, compressed air, liquid
air, pneumatics)

Abbott, William G., Jr.

Ingersoll-Rand Company

Union Carbide and Carbon Re-
search Laboratories, Inc.

Air conditioning (ventilation)

American Blower Company
American Radiator Company
U. S. Testing Ca, Inc.

Aircraft and accessories (see also
internal combustion motors)

Curtiss Aeroplane & Motor Cor-
poration
General Motors Research Corpo-
ration
Industrial Research Corporation
Martin, Glen L, Company, The
Packard Motor Car Company

Apparatus and instruments,
chemical and physical (astron-
omical instruments, autoclaves,
balances, compasses, gages,
lenses, microscopes, surveying

INDUSTRIAL RESEARCH LABORATORIES

95

instruments, telescopes, tran-
sits)

Baldwin Locomotive Works, The

Bausch & Lomb Optical Co.

Brown & Sharpe Mfg. Co.

Central Scientific Company

Coming Glass Works

Cutler-Hammer Mfg. Co., The

Eastman Kodak Company

Eimer & Amend

Electrical Testing Laboratories

Grosvenor, Wm. M.

Gurlcy, W. & L. E.

Kellogg Switchboard and Supply
Co,

Kilboume & Qark Manufacturing
Company

Keu£Fel & Esser Co.

Mojonnier Bros. Co.

Munn, W. Faitoute

Pyrolectric Instrument Company

Riverbank Laboratories

Sangamo Electric Company

Scientific Instrument and Electri-
cal Machine Company, The

Tolhurst Machine Works

Wallace & Tieman Co., Inc.

Waltham Watch Company

Automotive vehicles, equipment
and accessories (automobiles,
tanks, tractors, trucks)

Abbott. William G., Jr.
Boyer Chemical Laboratory Com-
pany
Champion Ignition Company
Diamond Chain & Manufacturing

Company
Dodge Brothers
Electrical Testing Laboratories
Fansteel Products Company, Inc
General Motors Research Corpo-
ration
Holt Manufacturing Company, The
Industrial Research Corporation
Lunkenheimer Co., The
Minneapolis Steel and Machinery

Co.
Northwestern Chemical Co., The
Packard Motor Car Company

Pierce- Arrow Motor Car Com-
pany, The

Stewart- Warner Speedometer
Corporation

Studebaker Corporation, The

Wallace & Tieman Co., Inc.

Beverages, non-alcoholic

California Fruit Growers Ex-
change

Dehls & Stein

Industrial Chemical Institute of
Milwaukee

Industrial Testing Laboratories

Lennox Chemical Co., The

Nowak Chemical Laboratories

Schwarz Laboratories

Skinner, Sherman & Esselen, In-
corporated

U. S. Food Products Corp.

Wahl-Henius Institute, Incorpo-
rated

Biological equipment and supplies

Baker, J. T., Chemical Co.
Beebe Laboratories, Inc
Central Scientific Company
Coleman & Bell Company, The
Dean Laboratories, Inc.
Digestive Ferments Co.
Eimer & Amend
Lilly, Eli, and Company
Mulford, H. K., Company
Swan-Myers Company

Building materials (asphalt, ce-
ment, concrete, lime, marble,
road materials, slate. See also
iron and steel)

Barber Asphalt Paving Company,

The
Beaver Board Companies, The
Borrowman, George
Conwell, W. L., ft Co., Inc.
Hunt, Robert W., and Co.
Institute of Industrial Research,

The
Interocean Oil Company, The
Lewis, F. J., Manufacturing Co.
Maynard, T. Poole

96

INDUSTRIAL RESEARCH LABORATORIES

National Lime Association

Pennsylvania Railroad G>mpany»
The

Richardson Company, The

Skinner, Sherman ft Esselen, In-
corporated

Standard Oil Company (New
Jersey)

Structural Materials Research
Laboratory

Toch Brothers

Weld and Liddell

By-products from wastes

Abbott, William G., Jr.

Anaconda Copper Mining Co.

California Fruit Growers Ex-
change

Davison Chemical Company, The

Emerson Laboratory

Federal Products Company, The

Grosvenor, Wm. M.

Harrison Mfg. Co., The

Kidde, Walter, & Company, In-
corporated

Koppers Company, The

Lakeview Laboratories

Laucks, L F., Inc.

Ljrster Chemical Company, Inc.

Maynard, T. Poole

Morris ft Company

Research Corporation

Scott, Ernest, ft Company

Stamford Dyewood Company

Swenson Evaporator Company

Teeple, John E.

Thac Industrial Products Corp.

Vacuum Oil Company, Incorpo-
rated

Weld and Liddell

Western Precipitation Company

Wheeler ft Woodru£F

White Tar Company of New
Jersey, Inc., The

Wilson ft Ca

Ceramics (bricks, china, glass,
magnesite, pottery, porcelain,
refractories)

American Window Glass Co.
Anaconda Copper Mining Co.

Andrews, A. B.

Babcock ft Wilcox Co., The

Bausdi ft Lomb Optical Co.

Beaver Falls Art Tile Company

Buckeye Qay Pot Co.

Carborundum Company, The

Celite Products Company

Champion Porcelain Company

Coming Glass Works

Dorite Manufacturing Company,
The

Ellis-Foster Company

FitzGerald Laboratories, Inc., The

Fry, H. C, Glass Company

Glass Container Association of
America

Harbison-Walker Refractories
Company

Kalmus, Comstock ft Wescott, Inc.

Keuffel ft Esser Co.

Koppers Company, The

Kraus Research Laboratories, Inc.

Laclede-Christy Clay Products
Company

Little, Arthur D., Inc.

Maynard, T. Poole

National Laboratories, The

National Lamp Works of General
Electric Company

Pfaudler Co., The

Pittsburgh Testing Laboratory

Ransom ft Randolph Co., The

Roessler ft Hasslacher Chemical
Company, The

Spencer Lens Company

Thac Industrial Products Corp.

Titanium Alloy Manufacturing Co.

Union Carbide and Carton Re-
search Laboratories, Inc.

Waltham Watch Company

Weld and Liddell

Western Gas Construction Com-
pany, The

Chemical engineering equipment
(agitators, blowers, centrifuges,
compressors, concentrators,
condensers, dryers, evapora-

INDUSTRIAL RESEARCH LABORATORIES

97

tors, filter presses, pulverizers,
pumps, separators)

Abb^ Engineering G)mpany

Abbott, William G.. Jr.

American Blower Company

Anaconda Copper Mining Co.

Andrews, A. B.

Bethlehem Shipbuilding Corpora-
tion, Ltd.

Bu£Falo Fomidry and Machine Ca

Cramp, William, ft Sons Ship ft
Engine Building Co., The

Deister Concentrator Company,
The

DeLaval Separator Co., The

Dorr Company, The

IngersoU-Rand 'Company

International Filter Co.

Oliver Continuous Filter Co.

Scott, Ernest, ft Company

Sperry, D. R., ft Co.

Swenson Evaporator Company

Tolhurst Machine Works

United States Bronze Powder
Works, Inc.

Wayne Oil Tank and Pump Co.

Western Gas Construction Com-
pany, The

Chemicals, fine, including sol-
vents

Abbott Laboratories, The
Atlantic Dyestuff Company
Baker, J. T., Chemical Co.
Barrett Company, The
Calco Chemical Company, The
Cams Chemical Company
Central DyestuflF and Chemical Co.
Central Scientific Company
Chemical Economy Company
Chemical Products Company
Coleman ft Bell Company, The
Cosmos Chemical Co., Inc.
Dehls ft Stein
Digestive Ferments Co.
Eastman Kodak Company
Eppley Laboratory
Federal Products Company, The
Florida Wood Products Co.
General Chemical Company

Harrison Mfg. Co., The

Heyden Chemical Company of
America, Inc.

Hsmson, Westcott & Dunning

Lakeview Laboratories

Lehn ft Fink, Inc.

Lemoine, Pierre, Cie., Inc.

Lindsay Light Company

Long & Co., Inc.

Lyster Chemical Company, Inc.

Mallinckrodt Chemical Works

McKesson ft Robbins, Incorporated

McLaughlin Gormley King Co.

Merck ft Co.

Monroe Drug Company

Monsanto Chemical Works

Newark Industrial Laboratories

New York Quinine ft- Chemical
Works, Incorporated, The

Norveil Chemical Corporation, The

Ohio Fuel Supply Company, The

Palmolive Company, The

Parke, Davis ft Company

Peet Bros. Mfg. Co.

Pfizer, Chas., ft Co., Inc.

Pharma-Chemical Corporation

Powers - Weightman - Rosengarten
Company, The

Radium Company of Colorado.
Inc., The

Radhim Limited, U. S. A.

Roessler ft Hasslacher Chemical
Company, The

Seydel Manufacturing Company

Sharp & Dohme

Special Chemicals Company

Squibb, K R., ft Sons

Thac Industrial Products Corp.

T. M. & G. Chemical Co.

Tower Manufacturing Co., Inc.

Union Carbide and Carbon Re-
search Laboratories, Inc.

U. S. Industrial Alcohol Company

Universal Aniline Dyes and Chem-
ical Co.

Wilbur White Chemical Co., The

Chemicals, heavy (acids, alkalies,
fungicides, insecticides, salts)
American Cyanamid Company
American Trona Corporation

98

INDUSTRIAL RESEARCH LABORATORIES

Anaconda Copper Mining Co.
Ansul Chemical Company
Armour Ammonia Works
Atlas Powder Ca
Bowker Insecticide Company
Brown Company

Buchanan, C. G., Chemical Com-
pany
Butterworth-Judson Corporation
California Fruit Growers Ex-
change
Carborundum Company, The
Carus Chemical Company
Charlotte Chemical Laboratories,

Inc.
Condensite Company of America
Davison Chemical Company, The
Detroit Testing Laboratory, The
Drackett, P. W., & Sons Co., The
du Pont, K I., de Nemours ft

Company
Eagle-Picher Lead Company, The
Eastern Manufacturing Company
Federal Phosphorus Company
General Chemical Company
Glidden Company, The
Grasselli Chemical Company
Great Western Electro-Chemical

Company
Grosvenor, Wm. M.
Harrison Mfg. Co., The
Hooker Electrochemical Company
Industrial Chemical Institute of

Milwaukee
Maas, A. R., Chemical Company
Mallinckrodt Chemical Works
Mathieson Alkali Works, Inc., The
McLaughlin Gormley King Co.
Merrimac Chemical Company
Metals & Chemicals Extraction

Corporation
Meyer, Theodore
Monsanto Chemical Works
National Lead Company
Naugatuck Chemical Company,

The
New Jersey Zinc Company, The
New York Quinine ft Chemical

Works, Incorporated, The
Norvell Chemical Corporation, The
Peet Bros. Mfg. Co.

Pennsylvania Salt Manufacturing
Co.

Philadelphia Quartz Company

Pittsburgh Plate Glass Co.

Powers - Weightman - Rosengartcn
Company, The

Pure Oil Company, Kanawha River
Salt and Chemical Division

Riches, Piver ft Co.

Rodman Chemical Company

Roessler & Hasslacher Chemical
Company, The

Saginaw Salt Products Co.

Seydel Manufacturing Company

Solvay Process Company, The

Swen^pn Evaporator Company

Titanium Pigment Co., Inc.

Union Carbide and Carbon Re-
search Laboratories, Inc.

United Chemical and Organic
Products Co.

United States Metals Refining Co.

Utah-Idaho Sugar Company

Victor Chemical Works

Welsbach Company

Wilckes, Martin, Wilckes Com-
pany

Chemistry, biological (bacteri-
ology, biology)

•

Abbott Laboratories, The
American Beet Sugar Company
American Hominy Company
American Institute of Baking
Banks & Craig
Beebe Laboratories, Inc.
Bridgeman-Russell Company
Coleman & Bell Company, The
Dean Laboratories, Inc.
Dearborn Chemical Company
Dehls ft Stein
Digestive Ferments Co.
Freed, H. E, Co., The
Gallun, A. F., ft Sons Co.
Glass Container Association of

America
Hochstadter Laboratories
Industrial Testing Laboratories
Kolynos Co., The
Lehn & Fink, Inc.

INDUSTRIAL RESEARCH LABORATORIES

99

Merrell-Soale Laboratory

Metz, H. A., Laboratories, Inc.

Miner Laboratories, The

Morris & G)mpany

Mulford, H. K., Company

National Canners Association

National Laboratories, The

New York Quinine & Chemical
Works, Incorporated, The

Parke, Davis & Company

Pease Laboratories

Physicians and Surgeons Labora-
tory

Porro Biological Laboratories

Schwarz Laboratories

Seydel Manufacturing Company

Skinner, Sherman & Esselen, In-
corporated

Special Chemicals Company

Sprague, Warner & Company

Squibb, E. R., & Sons

Swan-Myers Company

Takamine Laboratory, Inc.

Telling-Beele Vernon Company,
The

United States Glue Co.

U. S. Industrial Alcohol Company

Upjohn Company, The

Weston & Sampson

White Tar Company of New Jer-
sey, Inc., The

Wilson & Co.

Chemistry, inorganic (carbon,
graphite, etc.)

Acheson Graphite Company
American Agricultural Chemical

Company, The
American Chemical Paint Com-
pany
American Cyanamid Company
American Trona Corporation
Ansbacher, A. B., & Company
Ansul Chemical Company
Atlas Powder Co.
Baker ft Co., Inc.
Beaver Falls Art Tile Company
Borrowman, George
Borromite Co. of America, The
Bowker Insecticide Company

Brown Company

Buchanan, C. G., Chemical Com-
pany

Burdett Manufacturing Company

Carus Chemical Company

Celite Products Company

Charlotte Chemical Laboratories,
Inc.

Chase Metal Works

Childs, Charles M., & Co., Inc.

Condensite Company of America

Dearborn Chemical Company

Detroit Testing Laboratory, The

Diamond Match Co., The

Dorite Manufacturing Company,
The

Drackctt, P. W., & Sons Co., The

Eagle-Picher Lead Company, The

Eimer & Amend

Emerson Laboratory

FitzGerald Laboratories, Inc.

General Chemical Company

Glysyn Corporation, The

Grasselli Chemical Company

Great Western Electro-Chemical
Company

Harrison Mfg. Co., The

Heyden Chemical Company of
America, Inc.

Hochstadter Laboratories

Hooker Electrochemical Company

Industrial Chemical Institute of
Milwaukee

Jaques Manufacturing Company

Kalmus, Comstock & Wescott, Inc

Laucks, I. F., Inc.

Lee & Wight

Lennox Chemical Co., The

Lindsay Light Company

Maas, A. R., Chemical Company

Mallinckrodt Chemical Works

McNab & Harlin Manufacturing
Co.

Merck & Co.

Merrimac Chemical Company

Metals & Chemicals Extraction
Corporation

Mineral Refining & Chemical Cor-
poration

Munning, A. P., & Co.

100

INDUSTRIAL RESEARCH LABORATORIES

National Aniline & Qiemical Com-
pany, Incorporated

National Laboratories, The

National Lead Company

National Lime Association

Niles Tool Worics Company, The

Northwestern Chemical Co., The

Norvell Chemical Corporation, The

Pennsylvania Salt Manufacturing
Co.

Permutit Company, The

Perolin Company of America, The

Pfizer, Chas., & Co., Inc.

Pittsburgh Testing Laboratory

Pure Oil Company, Kanawha River
Salt and Chemical Division

Pyro-Non Paint Co., Inc.

Radium Company of Colorado^
Inc., The

Radium Limited, U. S. A.

Ransom ft Randolph Co., The

Rhode Island Malleable Iron
Works

Riches, Piver ft Co.

Rodman Chemical Company

Speer Carbon Company

Squibb, E. R., ft Sons

Teeple, John E.

Union Carbide and Carbon Re-
search Laboratories, Inc.

United States Bronze Powder
5 Works, Inc.

United States Smelting, Refinuig
ft Mining Company

Wadsworth Watch Case Ca, In-
corporated, The

Waltham Watch Company

Wedge Mechanical Furnace Com-
pany

Weld and Liddell

Wheeler ft Woodruff

White Tar Company of New Jer-
sey, Inc., The

Wilckes, Martin, Wilckes Com-
pany

Wiley ft Company, Inc.

Chemistry, mineralogical and
geological (quartz, etc.)

Celite Products Company

Charlotte Chemical Laboratories,
Inc.

Hirsch Laboratories, Inc., The

Lee ft Wight

Little, Arthur D., Inc.

Philadelphia Quartz Company

United States Smelting, Refining
& Mining Company

Wedge Mechanical Furnace Com-
pany

Western Research Corporation, In-
corporated

Chemistry, organic (fermenta-
tion, starch, vegetable oils, etc.)

American Beet Sugar Company
American Chemical and Manufac-
turing Corporation
American Chemical Paint Com-
pany
American Cyanamid Company
American Dianuilt Company
American Hominy Company
Atlantic Dyestuff Company
Avri Drug ft Chemical Company,

Inc.
Barrett Company, The
Beckman and Linden Engineering

Corporation
Bennetts' Chemical Laboratory
Betz, Frank S., Company
Bloede, Victor G., Co.
Brown Company
Calco Chemical Company, The
California Fruit Growers Ex-
change
California Ink Company, Inc.
Cams Chemical Company
Central Dyestuff and Chemical Co.
Charlotte Chemical Laboratories,

Inc.
Chemical Economy Company
Chemical Products Company
Chemical Service Laboratories,

Inc., The
Coleman ft Bell Company, The
Com Products Refining Company
Cosmos Chemical Co., Inc.
Cudahy Packing Co., The
Davis Chemical Products, Inc.

INDUSTRIAL RESEARCH LABORATORIES

101

Dearborn Chemical Company

Defals ft Stein

Detroit Testing Laboratory, The

Dewey ft Almy Chemical C6mi»any

Dextro Products, Inc

Dicks David Company, Incorpo-
rated

Digestive Ferments Ca

du Pont, £. I., de Nemours ft
Company

Dye Products & Chemical Com-
pany, Inc.

Eastman Kodak Company

Eimer ft Amend

Ellis-Foster Company

Emerson Laboratory

Feculose Co. of America

Federal Products Company, The

Foster-Heaton Company

Garfield Aniline Works, Inc

General Bakelite Company

General Chemical Company

Glysyn Corporation, The

Grasselli Chemical Company

Harrison Mfg. Co^ The

Heap, William, ft Sons

Heyden Chemical Company of
America, Inc.

Hirsch Laboratories, Inc, The

Hochstadter Laboratories

Hynson, Westcott ft Dunning

Industrial Chemical Institute of
Milwaukee

Industrial Testing Laboratories

Lakeview Laboratories

Laucks, I. F., Inc.

Lee ft Wight

Lehn ft Fink, Inc

Lemoine, Pierre, Cie., Inc

Lewis, F. J., Manufacturing Co.

Long ft Co., Inc.

Mallinckrodt Chemical Works

May Chemical Works

M. B. Chemical Co., Inc.

McLaughlin Gormley King Co.

Merck & Co.

Metz, H. A., Laboratories, Inc

Miner Laboratories, The

Monroe Drug Company

Musher and Company, Incorpo-
rated

National Aniline ft Chemical Com-
pany, Incorporated
National Laboratories, The
New York Quebracho Extract

Company, Incorporated
New York Quinine & Chemical

Works, Incorporated, The
New York Sugar Trade Labora-
*tory, Inc, The

Norvell Chemical Corporation, The
Nulomoline Company, The
Ohio Fuel Supply Company. The
Ohio Grease Co., The
Palmolive Company, The
Pfizer, Chas., ft Co., Inc.
Pharma-Chemical Corporation
Pittsburgh Testing Laboratory
Procter & Gamble Co., The
Pure Oil Company, Moore Oil and

Refining Company Division
Quinn, T. H., & Company
Radiant Dye ft Color Works
Schaeffer Brothers ft Powell

Manufacturing Company
Schwarz Laboratories
Sears, Roebuck and Co.
Seydel Manufacturing Company
Sharp ft Dohme

Skinner, Sherman & Esselen, In-
corporated
Southern Cotton Oil Company,

The
Special Chemicals Company
Squibb, E. R., ft Sons
Stamford Dye wood Company
Standard Oil Company (New

Jersey)
Standard Oil Company of Indiana
Swan-Myers Company
Swift ft Company
Takamine Laboratory, Inc.
Teeple, John K
Telling-Belle Vernon Company,

The
T. M. ft G. Chemical Ca
Tower Manufacturing Co., Inc
Ultro Chemical Corporation
U. S. Food Products Corp.
U. S. Industrial Alcohol Company
Universal Aniline Dyes and Chem-
ical Co.

102

INDUSTRIAL RESEARCH LABORATORIES

Utility Color & Chemical Co^ The

Van Schaack Brothers Chemical
Works, Inc.

Wallace, Joseph H., & Co.

Wells, Raymond

Western Gas Construction Com-
pany, The

Western Sugar Refinery

White Tar Company of New Jer-
sey, Inc., The

Whitten, J. O., Company, The

Wilbur White Chemical Co., The

Zinsser & Co.

Zobel, Ernst, Company, Inc.

Chemistry, pharmaceutical (cos-
metics, dentifrice, drug^s, disin-
fectants, medicines)

Abbott Laboratories, The

Avri Drug & Chemical Company,

Inc.
Betz, Frank S., Company
Bowker Insecticide Company
Boyer Chemical Laboratory Com-
pany
Calco Chemical Company, The
Carus Chemical Company
Caulk, L. D., Company, The
Central Dyestuff and Chemical Co.
Corn Products Refining Company
Cudahy Packing Co., The
Dean Laboratories, Inc.
Heinrich Laboratories of Applied

Chemistry
Heyden Chemical Company of

America, Inc.
Hirsch Laboratories, Inc., The
Hoehstadter Laboratories
Hynson, Wcstcott & Dunning
Industrial Chemical Institute of

Milwaukee
Industrial Testing Laboratories
Johnson & Johnson
Kolynos Co., The
Lakeview Laboratories
Larkin Co.
Lehn & Fink, Inc.
Lilly, Eli, and Company
Long & Co., Inc.
Lyster Chemical Company, Inc.

MacAndrews ft Forbes Company
Mallinckrodt Chemical Works
McKesson & Robbins, Incorpo-
rated
McLaughlin Gormley King Co.
Meigs, Bassett & Slaughter, Inc.
Merck ft Co.

Merrell, Wm. S., Company, The
Metz, H. A. Laboratories, Inc.
Meyer, Theodore
Milliken, John T., and Co.
Miner Laboratories, The
Monsanto Chemical Works
Mulford, H. K., Company
Newark Industrial Laboratories
New York Quinine ft Chemical

Works, Incorporated, The
Norvell Chemical Corporation, The
Parke, Davis ft Company
Pfizer, Chas., ft Co., Inc.
Pharma-Chemical Corporation
Physicians and Surgeons Labora-
tory
Pittsburgh Testing Laboratory
Sears, Roebuck and Co.
Seydel Manufacturing Company
Sharp & Dohme
Squibb, E. R., ft Sons
Standard Oil Company (New

Jersey)
Swan-Myers Company
Takamine Laboratory, Inc.
Thac Industrial Products Corp.
Union Carbide and Carbon Re-
search Laboratories, Inc.
United Drug Company
U. S. Food Products. Corp.
Upjohn Company, The
Warner, William R., & Company,

Incorporated
White Tar Company of New Jer-
sey, Inc., The
Wilckes, Martm, Wilckes Com-
pany
Zinsser & Co.

Concentration of ores (see also
chemical engineering equip-
ment)

Anaconda Copper Mining Co.

INDUSTRIAL RESEARCH LABORATORIES

103

Deister Concentrator Company,
The

Dorr Company, The

General Engineering Company,
Incorporated, The

Grasselli Chemical Company

Grosvenor, Wm. M.

James Ore Concentrator Co.

Maynard, T. Poole

Mesabi Iron Company

National Laboratories, The

Richards & Locke

Taggart and Yerxa

United States Smelting, Refining
& Mining Company

Utah Copper Company

Wedge Mechanical Furnace Com-
pany

Western Research Corporation,
Incorporated

Consulting research laboratories

Abbott, William G., Jr.

Andrews, A. B.

Babcock Testing Laboratory

Banks & Craig

Beckman and Linden Engineering
Corporation

Beebe Laboratories, Inc.

Bennetts' Chemical Laboratory

Borrowman, George

Cabot, Samuel, Inc.

Case Research Laboratory

Chemical Service Laboratories,
Inc., The

Cleveland Testing Laboratory Co.,
The

Commercial Testing and Engi-
neering Co.

Conwell, E. L., & Co., Inc.

Dean Laboratories, Inc.

Detroit Testing Laboratory, The

Dorr Company, The

Dunham, H. V.

Durfee, Winthrop C.

Electrical Testing Laboratories

Electrolabs Company, The

Ellis-Foster Company

Emerson Laboratory

Eppley Laboratory, The

Eustis, F. A.

Fahy, Frank P.

FitzGerald Laboratories, Inc., The

Fort Worth Laboratories

General Engineering Company, In-
corporated, The

Gray Industrial Laboratories, The

Grosvenor, Wm. M.

Hayes, Hammond V.

Heinrich Laboratories of Applied
Chemistry

Hirsch Laboratories, Inc., The

Hochstadter Laboratories

Howard Wheat and Flour Testing
Laboratory, The

Hunt, Robert W., and Co.

Industrial Chemical Institute of
Milwaukee

Industrial Research Corporation

Industrial Research Laboratories

Industrial Testing Laboratories

Institute of Industrial Research,
The

Kalmus, Comstock & Wescott,
Inc.

Kidde, Walter, & Company, In-
corporated

Kraus Research Laboratories

Lakeview Laboratories

Laucks, I. F., Inc.

Lee & Wight

Lincoln, E. S., Inc.

Little, Arthur D., Inc.

Littlefield Laboratories Co.

Lockhart Laboratories

Maynard, T. Poole

Mcllhiney, Parker C.

Meigs, Bassett & Slaughter, Inc.

Miner Laboratories, The

Munn, W. Faitoute

National Laboratories, The

Newark Industrial Laboratories

New York Sugar Trade Labora-
tory, Inc., The

Pease Laboratories

Pettee, Charles L. W., Labora-
tories of

Physicians and Surgeons Labora-
tory

Pittsburgh Testing Laboratory

104

INDUSTRIAL RESEARCH LABORATORIES

Porro Biological Laboratories
Porter, Horace C.
Quinn, T. H., & G)mpany
Research Corporation
Richards & Locke
Riverbank Laboratories
Rubber Trade Laboratory, The
Sabine, Wallace Qement, Labora-
tory

Schwarz Laboratories

Skinner, Sherman & Esselen, In-
corporated

Souther, Henry, Engineering G>.,
The

Structural Materials Research
Laboratory

Taggart and Yerxa

Takamine Laboratory, Inc.

Teeple, John E.

Wahl-Henius Institute, Incorpo-
rated

Weld and Liddell

Wells, Raymond

Western Precipitation G>mpany

Western Research Corporation,
Incorporated

Weston & Sampson

Wiley & Company, Inc.

Containers, including bottle seals
(cans, fiber-board containers,
etc.)

American Can Company

Bond Manufacturing Corporation

Chicago Mill and Lumber Com-
pany

Dewey & Almy Chemical Com-
pany

Glass Container Association of
America

Lehn & Fink, Inc.

National Association of Corru-
gated and Fibre Box Manufac-
turers, The

National Canners Association

Package Paper and Supply Cor-
poration

Vacuum Oil Company, Incorpo-
rated

Wheeler & Woodruff

Dyes, natural and artificial (inks,
intermediates, pigments, rib-
bons)

Amoskeag Manufacturing Com-
pany

Arlington Mills

Atlantic Dyestuff Company

Ault & Wiborg Company, The

Banks & Craig

Butterworth-Judson Corporation

Calco Chemical Company, The

California Ink Company, Inc.

Central Dyestuff and Chemical Co.

Coleman ft Bell Company, The

Dicks David Company, Incorpo-
rated

du Pont, E. I., de Nemours &
Company

Durfee, Winthrop C.

Dye Products & Chemical Com-
pany, Inc.

Eastern Finishing Works, Inc.

Eavenson & Levering Co.

Emerson Laboratory

Foster-Heaton Company

Garfield Aniline Works, Inc.
Garrison Mfg. Co., The
Grasselli Chemical Company
Grosvenor, Wm. M.
Hirsch Laboratories, Inc., The
Hodcer Electrochemical Company
Klearflax Linen Rug Company
Little, Arthur D., Inc.
Lockhart Laboratories
Long & Co., Inc.
MacAndrews & Forbes Company
May Chemical Works
M. B. Chemical Co., Inc.
Merrimac Chemical Company
Monroe Drug Company
Monsanto Chemical Works
Morrill, Gea H., Co.
National Aniline & Chemical Com-
pany, Incorporated
National Laboratories, The
Naugatuck Chemical Company,

The
Newark Industrial Laboratories
Northwestern Chemical Co., The
Oliver Continuous Filter Co.

INDUSTRIAL RESEARCH LABORATORIES

105

Palatine Aniline and Chemical
Corporation

Peerless Color Company

Pharma-Chemical Corporation

Pittsburgh Plate Glass Co.

Radiant Dye & Color Works

Reliance Aniline & Chemical Ca,
Incorporated

Sears, Roebuck and Co.

Seydel Manufacturing Company

Stamford Dyewood Company

Sun Chemical & Color Co.

T. M. & G. Chemical Co.

Tower Manufacturing Co., Inc.

Ultro Chemical Corporation

U. S. Testing Co., Inc.

Universal Aniline Dyes and Chem-
ical Co.

Utility Color & Chemical Co., The

White Tar Company of New Jer-
sey, Inc., The

Wilbur White Chemical Ca, The

Zinsser & Co.

Electrical communication (cable,
telegraph, telephone, wireless)

American Radio and Research
Corporation

Belden Manufacturing Company

Coming Glass Works

General Electric Company

Hayes, Hammond V.

Industrial Research Corporation

Kellogg Switchboard and Supply
Co.

Kilbourne & Qark Manufacturing
Company

Munn, W. Faitoute

Western Electric Company, Incor-
porated

Electrical equipment and instru-
ments (ammeters, lamps, volt-
meters, wattmeters)

Abbott, William G., Jr.

Allen-Bradley Co.

American Radio and Research

Corporation
Commonwealth Edison Company
Cooper Hewitt Electric Company

Coming Glass Works
Cutler-Hammer Mfg. Co., The
Edison, Thomas A., Laboratory
Electrical Testing Laboratories
Fansteel Products Company, Inc.
General Electric Company
Hoskins Manufacturing Company
Kilbourne & Clark Manufacturing

Company
Kellogg Switchboard and Supply

Ca
Leeds & Northrup Company
Munn, W. Faitoute
National Lamp Works of General

Electric Company
Pyrolectric Instrument Company
Sangamo Electric Company
Scientific Instrument and Electri-
cal Machine Company, The
Speer Carbon Company
Union Carbide and Carbon Re-
search Laboratories, Inc.
Western Electric Company, Incor-
porated
Westinghouse Electric & Manu-
facturing Company
Westinghouse Lamp Co.

«

Electricity, general (economics,
utilization)

Belden Manufacturing Company
Cutler-Hammer Mfg. Co., The
Edison, Thomas A., Laboratory
Electrical Testing Laboratories
General Electric Company
Hayes, Hammond V.
Kilbourne & Clark Manufacturing

Company
Western Electric Company, Incor-
porated
Westinghouse Electric & Manu-
facturing Company

Electric power (conversion, dis-
tribution, dynamos, generation,
motors, power plants, trans-
mission)

American Radio and Research

Corporation
Commonwealth Edison Company

106

INDUSTRIAL RESEARCH LABORATORIES

Cutlcr-Hammcr Mfg. Co., The
Detroit Edison Company, The
General Electric Company
General Motors Research Corpo-
ration
Imperial Belting Company
Industrial Research Corporation
Lincoln, E. S., Inc.
S. K. F. Industries, Inc.
Union Carbide and Carbon Re-
search Laboratories, Inc.
Vesta Battery Corporation

Electrochemistry (electrochem-
ical processes, electrodes, stor-
age batteries)

Acheson Graphite Company
Anaconda Copper Mining Co.
Andrews, A. B.
. Beckman and Linden Engineering
Corporation
Carborundum Company, The
Eastern Manufacturing Company
Edison, Thomas A., Laboratory
Electro Chemical Company, The
Elcctrolabs Company, The
Eppley Laboratory
FitzGerald Laboratories, Inc., The
Grasselli Chemical Company
Great Western Electro-Chemical

Company
Grosvenor, Wm. M.
Hirsch Laboratories, Inc., The
Hooker Electrochemical Company
International Silver Company
Kidde, Walter, & Company, In-
corporated
Leeds & Northrup Company
Littlefield Laboratories Co.
Mathieson Alkali Works, Inc., The
Mcllhiney, Parker C.
National Lamp Works of General

Electric Company
Prest-O-Lite Co., Inc., The
Riverbank Laboratories
Speer Carbon Company
Union Carbide and Carbon Re-
search Laboratories, Inc.
United States Smelting, Refining
& Mining Company

Vesta Battery Corporation
Weld and Liddell

Electro-plating

Bausch & Lomb Optical Co.

Columbia Graphophone Manufac-
turing Company

Crompton & K n o w 1 e s Loom
Works

Gillette Safety Razor Co.

Gurley, W. & L. E.

Munn, W. Faitoute

Munning, A. P., & Co.

Sears, Roebuck and Co.

Union Carbide and Carbon Re-
search Laboratories, Inc.

Waltham Watch Company

Wheeler & Woodruff

Explosives and explosions (dyna-
mite, powder, TNT)

Atlas Powder Co.

Barrett Company, The

Davis Chemical Products, Inc.

du Pont, E. I., de Nemours &

Company
Grasselli Chemical Company
Hercules Powder Co.
Remington Arms, Union Metallic

Cartridge Company
Van Schaack Brothers Chemical

Works, Inc.

Fats, fatty oils and soaps

American Chemical and Manufac-
turing Corporation
Armour Glue Works
Armour Soap Works
Babcock Testing Laboratory
Corn Products Refining Company
Cudahy Packing Co., The
Dunham, H. V.
Electrolabs Company, The
Fort Worth Laboratories
Globe Soap Company, The
Industrial Testing Laboratories
Kalmus, Comstock & Wcscott, Inc.
Kidde, Walter, & Company
Larkin Co.

, INDUSTRIAL RESEARCH LABORATORIES

107

Laucks, I. F., Inc.

Lehn & Fink, Inc.

Lockhart Laboratories

Mcllhiney, Parker C

Miner Laboratories^ The

Mojonnier Bros. Co.

Musher and Company, Incorpo-
rated

National Lead Company

Ohio Grease Co., The

Palmolive Company, The

Peet Bros. Mfg. Co.

Procter & Gamble Co., The

Pure Oil Company, Moore Oil and
Refining Company Division

Riverbank Laboratories

Schaeffer Brothers & Powell
Manufacturing Company

Schwarz Laboratories

Seydel Manufacturing Company

Skinner, Sherman & Esselen, In-
corporated

Souther, Henry, Engineering Co.,
The

Southern Cotton Oil Company,
The

Swift & Company

Wells, Raymond

Wheeler & WoodruflF

Wiley & Company, Inc.

Filtration

Celite Products Company
DeLaval Separator Co., The
Dorr Company, The
International Filter Co.
Oliver Continuous Filter Co.
Sperry, D. R., & Co.

Fire prevention (extinguishers,
sprinklers)

Factory Mutual Laboratories
MacAndrews & Forbes Company
Underwriters' Laboratories

Food8*(aIimentary pastes, bakery,
baking powder, biscuit, butter,
candy, canning and preserving,
cold storage, flavoring extracts,
flour, gelatine, meat and meat

products, milk, oils, preserva-
tives, wheat, yeast, etc.)

American Can Company
American Hominy Company
American Institute of Baking
Banks & Craig
Brach, E. J., and Sons
Bridgeman-Russell Company
Brown Company

California Fruit Growers Ex-
change
Carus Chemical Company
Cleveland Testing Laboratory Co.,

The
Com Products Refining Company
Cudahy Packing Co., The
Dunham, H. V.
Emerson Laboratory
Forth Worth Laboratories
Frees, H. K, Co., The
Gibbs Preserving Company
Glass Container Association of

America
Hochstadter Laboratories
Hooker Electrochemical Company
Howard Wheat and Flour Test-
ing Laboratory, The
Industrial Research Laboratories
Industrial Testing Laboratories
Jaques Manufacturing Company
Lehn & Fink, Inc.
Long & Co., Inc.
McLaughlin Gormley King Co.
Merrell-Soule Laboratory
Miner Laboratories, The
Mojonnier Bros. Co.
Morris & Company
Musher and Company, Incorpo-
rated
National Biscuit Company
National Canners Association
National Cereal Products Labora-
tories
National Laboratories, The
Nestle's Food Company, Incorpo-
rated
Newark Industrial Laboratories
New England Confectionery Com-
pany

Nowak Chemical Laboratories

10b

INDUSTRIAL RESEARCH LABORATORIES

Pease Laboratories

Penick & Ford, Ltd., Incorporated

Pittsburgh Testing Laboratory

Procter & Gamble Co., The

Redlands Fruit Products G)m-
pany

Rumford Chemical Works

Schwarz Laboratories

Sears, Roebuck and Co.

Seydel Manufacturing Company

Skinner, Sherman & Esselen, In-
corporated

Southern Cotton Oil Company,
The

Sprague, Warner ft Company

Swift & Company

Takamine Laboratory, Inc.

Telling-Belle Vernon Company,
The

United Chemical and Organic
Products Co.

U. S. Food Products Corp.

United SUtes Glue Co.

Wahl-Henius Institute, Incorpo-
rated

Wallace ft Tieman Co., Inc.

Washburn-Crosby Co.

Wheeler ft WoodruflF

Whitten, J. O., Company, The

Wilckes, Martin, Wilckes Com-
pany

Wilson ft Co.

Foundry equipment,
and methods (casting, die cast-
ing, moulding)

American Brass Company, The

Crane Co.

Doehler Die-Casting Co.

General Motors Research Corpo-
ration

Gurley, W. ft L. E.

Lunkenheimer Co., The

Niles Tool Works Company, The

Pettee, Charles L. W., Labora-
tories of

Rhode Island Malleable Iron
Works

Stockham Pipe ft Fittings Co.

Union Carbide and Carbon Re-
search Laboratories, Inc.

United States Bronxe Powder
Works, Inc.

Fuels (alcohol, charcoal, coal,
coke, gasoline, kerosene, oil,
peat See also gas, petroleum
and wood)

American Can Company

American Radiator Company

Anaconda Copper Mining Co.

Andrews, A. B.

Atlantic Refintng Company, The

Babcock ft Wikox Co., The

Barrett Company, The

Bridgeman-Russell Company
. Chemical Service Laboratories,
Inc., The

Commercial Testing and Engineer-
ing Co.

Consolidated Gas Company of
New Yoric

Dearborn Chemical Company

Detroit Testing Laboratory, The

Dodge Brothers

Doherty Research Company, Em-
pire Division

Electrical Testing Laboratories

Emerson Laboratory

Federal Products Company, The

General Motors Research Corpo-
ration

Gulf Pipe Line Company

Hyco Fuel Products Corporation

Industrial Chemical Institute of
Milwaukee

Industrial Testing Laboratories

Interocean Oil Company, The

James Ore Concentrator Co.

Koppers Company, The

Laucks, I. F., Inc.

Lewis, F. J., Manufacturing Co.

Little, Arthur D., Inc.

Lockhart Laboratories

Martinez Refinery, Shell Co. of
California

Meigs, Bassett ft Slaughter, Inc.

Milwaukee Coke ft Gas Company,
The

Ohio Fuel Supply Company, The

INDUSTRIAL RESEARCH LABORATORIES

109

Porter, Horace C.

Providence Gas G>mpany, Incor-
porated

Quinn, T. H., ft G>nipany

Rhode Island Malleable Iron
Works

Rodman Chemical Company

Schwarz Laboratories

Sears, Roebuck and Co.

Souther, Henry, Engineering Co.,
The

Standard Oil Company (New
Jcrs^)

U. S. Industrial Alcohol Company

United States Smelting, Refining
ft Mining Company

Wayne Oil Tank and Pump Co.

Western Gas Construction Com-
pany, The

Western Research Corporation,
Incorporated

Wheeler ft Woodruff

Wiley ft Company, Inc.

Fuel utilization (boilers, furnaces,
gas - producers, radiators,
stokers)

American Blower Company

American Radiator Company

Brooklyn Union Gas Company,
The

Celite Products Company

Champion Porcelain Company

Cochrane, H. S. B. W., Corpora-
tion

Commercial Testing and Engineer-
ing Co.

Consolidated Gas Company of
New York

Consolidated Gas, Electric Light
and Power Company of Balti-
more

Doherty Research Company, Em-
pire Division

Hunt, Robert W., and Co.

Kidde, Walter, & Company

Koppers Company, The

Porter, Horace C.

Rhode Island Malleable Iron
Works

Western Gas Construction Com-
pany, The
Wheeler ft Woodruff

Gas, fuel and illuminating^ in-
cluding mantles (acetylene,
hydrogen)

Brooklyn Union Gas Company,
The

Chemical Service Laboratories,
Inc.

Consolidated (jas Company of
New York

Consolidated Gsls, Electric Light
and Power Company of Balti-
more

Cosden ft (^mpany

Detroit Edison Company, The

Gulf Pipe Line Company

Harrison Mfg. Co., The

Koppers Company, The

Little, Arthur D., Inc.

Milwaukee Coke ft Gsls Company,
The

Ohio Fuel Supply Company, The

Porter, Horace C.

Providence Gas Company, Incor-
porated

Standard Oil Company (New
Jersey)

Union Carbide and Carbon Re-
search Laboratories, Inc.

United Gas Improvement Co., The

Welsbach Company

Western Gas Construction Com-
pany, The

Wheeler ft Woodruff

Gases, except fuel and illumina-
ting, including generating ap-
paratus (argon, chlorine,
helium, neon, nitrogen, oxygen,
poisonous gases)

Burdctt Manufacturing Company
Electrolabs Company, The
Florida Wood Products Co.
Great Western Electro-CHiemical
Company

Hooker Electrochemical Company
Lennox Chemical Co., The

110

INDUSTRIAL RESEARCH LABORATORIES

Mathieson Alkali Works, Inc., The

Union Carbide and Carbon Re-
search Laboratories, Inc.

Wallace & Tiernan Co., Inc.

Western Gas Construction Com-
pany, The

Hair, curled, etc.

Armour Curled Hair Works
Cudahy Packing Co., The
Pfister & Vogel Leather Co.

Heat (calorimetry, pyrometry,
thermal physics, thermometry)

Celite Products Company
Commonwealth Edison Company
General Motors Research Corpo-
ration
Koppers Company, The
Leeds & Northrup Company
Munn, W. Faitoute
Pyrolectric Instrument Company
Rhode Island Malleable Iron

Works
Swenson Evaporator Company
Union Carbide and Carbon Re-
search Laboratories, Inc.
Wahl-Henius Institute, Incorpo-
rated

Heating

American Blower Company
American Radiator Company
Cochrane, H. S. B. W., Corpora-
tion
Detroit Edison Company, The
Hoskins Manufacturing Company

Hydraulics (waterworks, water
power)

Cochrane, H. S. B. W., Corpora-
tion

Cramp, William, & Sons Ship &
Engine Building Co., The

Illumination, electric, gas and
other

Brooklyn Union Gas Company,
The

Commonwealth Edison Company

Consolidated Gas Company of
New York

Consolidated Gas, Electric Light
and Power Company of Balti-
more

Cooper Hewitt Electric Company

Coming Glass Works

Harrison Mfg. Co., The

National Lamp Works of General
Electric Company

Ohio Fuel Supply Company, The

Providence Gas Company, Incor-
porated

Union Carbide and Carbon Re-
search Laboratories, Inc.

United Gas Improvement Co., The

Welsbach Company

Westinghouse Electric & Manu-
facturing Company

Westinghouse Lamp Co.

Insulation, electrical and thermal
(cable, cordage, non-conduc-
tors, insulated wire)

Allen-Bradley Co.

Belden Manufacturing Company

Boonton Rubber Manufacturing

Company
Carborundum Company, The
Celite Products Company
Champion Ignition Company
Condensite Company of America
Electrical Testing Laboratories
General Bakelite Company
Habirshaw Electric Cable Com-
pany, Inc.
Kellogg Switchboard and Supply

Co.
Kilbourne ft Clark Manufacturing

Company
Redmanol Chemical Products Co.
Sangamo Electric Company
Standard Underground Cable

Company
Vacuum Oil Company, Incorpo-
rated

INDUSTRIAL RESEARCH LABORATORIES

111

Internal combustion motors (Die-
sel engines, gasoline engines,
motors, oil engines)

Abbott, William G., Jr.

Bethlehem Shipbuilding Corpora-
tion, Ltd.

General Motors Research Corpo-
ration

Ingersoll-Rand Company

Standard Oil Company (New
Jersey)

Studebaker Corporation, The

Iron and steel (cast iron, ferrous
alloys, pipe, wrought iron)

American Chemical Paint Com-
pany

American Rolling Mill Co., The

American Sheet and Tin Plate
Company

Barber-Colman Company

Borrowman, George

Buffalo Foundry and Machine Co.

Byers, A. M., Company

Carnegie Steel Company

Chase Metal Works

Cleveland Testing Laboratory Co.,
The

Crane Co.

Crompton & Knowlcs Loom Works

Crucible Steel Company of Amer-
ica

Diamond Chain & Manufacturing
Company

Dodge Brothers

Duriron Company, Inc., The

Eastern Malleable Iron Company

Fahy, Frank P.

Fansteel Products Company, Inc.

General Motors Research Corpo-
ration

Gillette Safety Razor Co.

Houghton, £. F., & Co.

Hunt, Robert W., and Co.

Industrial Works

Inland Steel Company

Kokomo Steel and Wire Co.

Ludlum Steel Company

Lunkenheimer Co., The

Maynard, T. Poole

McNab & Harlin Manufacturing

Co.
Mesabi Iron Company
Midvale Steel and Ordnance Com-
pany
Minneapolis Steel and Machinery

Ca
National Malleable Castings Com-
pany, The
National Tube Company
Nilcs Tool Works Company, The
Peerless Drawn Steel Company,

The
Pettee, Charles L. W., Labora-
tories of
Pierce-Arrow Motor Car Com-
pany, The
Rhode Island Malleable Iron

Works
Rodman Chemical Company
Sangamo Electric Company
Stockham Pipe & Fittings Co.
Tacony Steel Company
Titanium Alloy Manufacturing Co.
Union Carbide and Carbon Re-
search Laboratories, Inc.
United Alloy Steel Corporation
United States Smelting, Refining

& Mining Company
Vanadium- Alloys Steel Co., The
Vanadium Corporation of America
Waltham Watch Company
Western Gas Construction Com-
pany, The

Leather and leather goods (boots,
shoes, including machinery,
leather substitutes, tanning)

Atlas Powder Co.

Carus Chemical Company

Dennis, Martin, Company, The

Durfee, Winthrop C.

Gallun, A. F., & Sons Co.

Houghton, E. F., & Co.

International Shoe Co.

Kidde, Walter, & Company

Kullman, Salz & Co.

New York Quebracho Extract

Company, Incorporated
Pantasote Leather Company, The

112

INDUSTRIAL RESEARCH LABORATORIES

Pfister & Vogel Leather Co.

United Shoe Machinery Corpora-
tion

Vactram Oil Company, Incorpo-
rated

Light (optical instruments, optics.
See also illumination)

American Optical Company
Bausch & Lomb Optical Co.
Case Research Laboratory
Cooper Hewitt Electric Company
Coming Glass Works
Eastman Kodak Company
Gurley, W. & L. E.
Keuffel & Esser Co.
National Lamp Works of General

Electric Company
Spencer Lens Company

Liquors, fermented and distilled
(alcohol, beer, wirie)

Frees, H. R, Co., The
Industrial Chemical Institute of

Milwaukee
Industrial Testing Laboratories
National Laboratories, The
Nowak Chemical Laboratories
Wahl-Henius Institute, Incorpo-
rated
Wiley & Company, Inc.

Lubricants (carbon, graphite, oil,
petroleum)

Acheson Graphite Company

Chase Metal Works

Columbia Graphophone Manufac-
turing Company

Commercial Testing and Engineer-
ing Co.

Dearborn Chemical Company

Dodge Brothers

Doherty Research Company, Em-
pire Division

Gray Industrial Laboratories, The

Industrial Testing Laboratories

Interocean Oil Company, The

Laucks, I. F., Inc.

Lf ickhart . Laboratories

Martinez Refinery, Shell Co. of
California

Maynard, T. Poole

Minneapolis Steel and Machinery
Co.

Ohio Grease Co., The

Pittsburgh Testing Laboratory

Pure Oil Company, Moore Oil and
Refining Company Division

Schwarz Laboratories

S. K. F. Industries, Inc.

Speer Carbon Company

Standard Oil Company (New
Jersey)

Union Carbide and Carbon Re-
search Laboratories, Inc.

Vacuum Oil Company, Incorpo-
rated

Ventura Refining Company

Wayne Oil Tank and Pump Ca

Weld and Liddell

Wells, Raymond

Western Gas Construction Com-
pany, The

Western Research Corporation,
Incorporated

Wheeler & Woodruff

Wiley & Company, Inc.

Machine tools and hardware (cut-
lery, drill-presses, lathes,
planers, shapers)

Barber-Colman Company
Brown & Sharpe Mfg. Co.
Niles Tool Works Company, The
Rochester Button Company
Stockham Pipe & Fittings Co.
United Shoe Machinery Corpora-
tion
Winchester Repeating Arms Co.

Magnetism

Electrical Testing Laboratories
Kilboume & Clark Manufacturing

Company
Leeds & Northrup Company

Marine engineering (ships)

Cramp, William, & Sons Ship ft
Engine Building Co., The

INDUSTRIAL RESEARCH LABORATORIES

113

Matches

Diamond Match Co., The

Mechanics, general (bearings,
bail, roller, etc.).

Minneapolis Steel and Machinery

Co.
National Cash Register Company,

The
S. K. F. Industries, Inc.

Metal manufactures, miscellan-
eous (fittings, pipes, valves)

Byers, A. M., Company

Crane Co.

Grasselii Chemical Company

Lmikenheimer Co., The

McNab & Harlin Manufacturing
Ca

National Tube Company

Scovill Manufacturing Company

Stockham Pipe & Fittings Co.

Western Gas Construction Com-
pany, The

Winchester Repeating Arms Co.

Metallurgy and metallography, in-
cluding equipment

American Brass Company, The
American Optical Company
American Sheet and Tin Plate

Company
Anaconda Copper Mining Co.
Babcock & Wilcox Co., The
Bennetts' Chemical Laboratory
Borrowman, George
Bridgeport Brass Company
Buffalo Foundry and Machine Co.
Byers, A. M., Company
Calumet ft Hecla Mining Com-
pany
Carnegie Steel Company
Chase Metal Works
Qeveland Testing Laboratory Co.,

The
Crane Co.

Crompton ft Knowles Loom Works
Detroit Testing Laboratory, The
Dodge Brothers

Dorr Company, The

Duriron Cotiapany, Inc., The

Eastern Malleable Iron Company

Eustis, F. A.

Fansteel Products Company, The

FitzGerald Laboratories, Inc., The

General Electric Company

General Engineering Company, In-
corporated, The

General Motors Research Corpo-
ration

Gillette Safety Razor Co.

Hirsch Laboratories, Inc., The

Hoskins Manufacturing -Company

Hunt, Robert W., and Co.

Industrial Works

International Nickel Company,
The

James Ore Concentrator Co.

Kalmus, Comstock ft Wescott, Inc.

Kokomo Steel and Wire Co.

Lumen Bearing Company

Lunkenheimer Co., The

McNab ft Harlin Manufacturing
Co.

Metals & Chemicals Extraction
Corporation

Midvale Steel and Ordnance Com-
pany

Minneapolis Steel and Machinery
Ca

National Cash Register Company,
The

National Lamp Works of General
Electric Company

National Lead Company

National Malleable Castings Com-
pany, The

Niles Tool Works Company, The

Oliver Continuous Filter Co.

Peerless Drawn Steel Company,
The

Pierce-Arrow Motor Car Com-
pany, The

Raritan Copper Works

Research Corporation

Rhode Island Malleable Iron
Works

Rodman Chemical Company

Scovill Manufacturing Company

114

INDUSTRIAL RESEARCH LABORATORIES

Sears, Roebuck and Co.

S. K. F. Industries, Inc.

Souther, Henry, Engineering Co.,
The

Stewart - Warner Speedometer
Corporation

Studebaker Corporation, The

Titanium Alloy Manufacturing Co.

Union Carbide and Carbon Re-
search Laboratories, Inc.

United Alloy Steel Corporation

United States Metals Refining Co.

United Sutes Smelting, Refining
& Mining Company

Vanadium Corporation of Amer-
ica

Wadsworth Watch Case Co., In-
corporated, The

Wahham Watch Company

Wedge Mechanical Furnace Com-
pany

Westinghouse Electric & Manu-
facturing Company

Wheeler & Woodruff

Military and naval equipment
(ammunition, armor, ordnance,
small arms, torpedoes)

Abbott, William G., Jr.

Remington Arms, Union Metallic
Cartridge Company

Union Carbide and Carbon Re-
search Laboratories, Inc.

Winchester Repeating Arms Co.

Mining, general (testing drills,
ropes, tools ; ore dressing)

Deister Concentrator Company,

The
Doherty Research Company, Em-
pire Division
Dorr Company, The
IngersoU-Rand Company
James Ore Concentrator Co.
Maynard, T. Poole
National Lead Company
Oliver Continuous Filter Co.
United States Smelting, Refining
& Mining Company

Non-ferrous metals (aluminum,
bearing metals, brass, bronze,
copper, gold, lead, nickel, plati-
num, silver, tin, titanium, zinc)

Aluminum Company of America

American Brass Company, The

American Can Company

American Sheet and Tin Plate
Company

Anaconda Copper Mining Co.

Baker & Ca, Inc.

Bethlehem Shipbuilding Corpora-
tion, Ltd.

Bridgeport Brass Company

Calumet ft Hecla Mining Company

Chase Metal Works

Cramp, William, & Sons Ship ft
Engine Building Co., The

Crane Co.

Crompton & Knowles Loom Works

Dodge Brothers

Doehler Die-Casting Co.

Eagle-Picher Lead Company, The

Fansteel Products Company, Inc.

General Motors Research Corpo-
ration

Glidden Company, The

Grasselli Chemical Company

Grosvenor, Wm. M.

Gurley, W. ft L. E.

Hochstadter Laboratories

Industrial Works

International Nickel Company, The

International Silver Company

Lumen Bearing Company

Lunkenheimer Co., The

McNab ft Harlin Manufacturing
Co.

Metals ft Chemicals Extraction
Corporation

Mineral Refining ft Chemical Cor-
poration

National Canners Association

National Lamp Works of General
Electric Company

National Lead Company

New Jersey Zinc Company

Niles Tool Works Company, The

Pettee, Charles L. W., Labora-
tories of

INDUSTRIAL RESEARCH LABORATORIES

115

Radium Company of Colorado,
Inc., The

Radium Limited, U. S. A.

Raritan Copper Works

Remington Arms, Union Metallic
Cartridge Company

Roessler & Hasslacher Chemical
Company, The

Scovill Manufacturing Company

Union Carbide and Carbon Re-
search Laboratories, Inc.

United States Bronze Powder
Works, Inc.

United States Metals Refining Co.

United States Smelting, Refining
& Mining Company

Vanadium Corporation of Amer-
ica

Wadsworth Watch Case Co., In-
corporated, The

Waltham Watch Company

Weld and Liddell

0£Ece equipment (adding ma-
chines, cash registers)

Abbott, William G., Jr.
National Cash Register Company,
The

;, oils and varnishes (dryers,
enamels, lacquers, pigments,
putty, resins, rust-proofing)

Abbott, William G.. Jr.

Acme White Lead & Color Works

American Chemical and Manufac-
turing Corporation

American Chemical Paint Com-
pany

Andrews, A. B.

Ansbacher, A. B., & Company

Atlas Powder Co.

Ault & Wiborg Company, The

Babcock Testing Laboratory

Berry Brothers, Inc.

Borrowman, George

Boyer Chemical Laboratory Com-
pany

Buchanan, C. G., Chemical Com-
pany

Cabot, Samuel, Inc.

Cams Chemical Company

Chase Metal Works

Childs, Charles M., & Co., Inc.

Chemical Products Company

Condensite Company of America

Davis Chemical Products, Inc.

Dodge Brothers

Drackett, P. W., & Sons Co., The

du Pont, £. I., de Nemours &
Company

Eagle-Picher Lead Company, The

Glidden Company, The

Grosvenor, Wm. M.

Hunt, Robert W., and Co.

Imperial Belting Company

Industrial Chemical Institute of
Milwaukee

Industrial Testing Laboratories
' Krebs Pigment and Chemical Co.

Lakeview Laboratories

Little, Arthur D., Inc.

Lockhart Laboratories

Mcllhiney, Parker C.

Mineral Refining & Chemical Cor-
poration

National Laboratories, The

National Lead Company

Newport Company, The

Perolin Company of America, The

Pfister & Vogel Leather Co.

Pittsburgh Plate Glass Co.

Pyro-Non Paint Co., Inc.

Redmanol Chemical Products Co.

Richardson Company, The

Rubber Trade Laboratory, The

Sangamo Electric Company

Sears, Roebuck and Co.

Skinner, Sherman & Esselen, In-
corporated

Titanium Pigment Co., Inc.

Toch Brothers

Ultro Chemical Corporation

United States Bronze Powder
Works, Inc.

Wayne Oil Tank and Pump Co.

Wells, Raymond

Wheeler & Woodruff

Zobel, Ernst, Company, Inc.

Petroleum and its products (see
also lubricants)

Atlantic Refining Company, The

116

INDUSTRIAL RESEARCH LABORATORIES

Babcock Testing Laboratory

Barber Asphalt Paving G>mpany,
The

Charlotte Chemical Laboratories,
Inc.

Cosden & Company

Doherty Research Company, Em-
pire Division

Dunham, H. V.

Gray Industrial Laboratories

Gulf Pipe Line Company

Institute of Industrial Research,
The

Interocean Oil Company, The

Little, Arthur D., Inc.

Lockhart Laboratories

Martinez Refinery, Shell Co. of
California

Ohio Fuel Supply Company, The

Richardson Company, The

Schaeffer Brothers ft Powell
Manufacturing Company

Standard Oil Company (New
Jersey)

Standard Oil Company of Indiana

Union Carbide and Carbon Re-
search Laboratories, Inc.

Vacuum Oil Company, Incorpo-
rated

Ventura Refining Company

Wayne Oil Tank and Pump Co.

Weld and Liddell

Western Gas Construction Com-
pany, The

Western Research Corporation, In-
corporated

Wheeler ft Woodruff

2^bel, Ernst, Company, Inc.

Phonographs and graphophones

Columbia Graphophone Manufac-
turing Company
Edison, Thomas A., Laboratory

Photography (c a m e r a a, de-
velopers, films, moving-pic-
ture equipment, plates)

Ansco Company

Chemical Economy Company

Coming Glass Works

Eastman Kodak Company

Grosvenor, Wm. M.

Heinrich Laboratories of Applied

Chemistry
Hirsch Laboratories, Inc., The
Kalmus, Comstock & Wescott, Inc.
Munn, W. Faitoute
National Lead Company
United States Glue Co.
Zinsser ft Co.

Plastics (bakelite, condensite, red-
manol ; casting and moulding of
plastics)

Abbott, WilUam G., Jr.

Boonton Rubber Manufacturing
Company

Champion Ignition Company

Columbia Graphophone Manufac-
turing Company

Condensite Company of America

du Pont, E. L, de Nemours ft
Company

General Bakelite Company

Heap, William, ft Sons

Meigs, Bassett ft Slaughter, Inc.

Redmanol Chemical Products Co.

Rubber Trade Laboratory, The

Properties of engineerhig ma-
terials

American Brass Company, The

Borrowman, George

Carborundtun Company, The

Chicago Mill and Lumber Com-
pany

Columbia Graphophone Manufac-
ing Company

Electrical Testing Laboratories

General Electric Company

Hunt, Robert W., and Co.

Industrial Works

Institute of Industrial Research,
The

Kokomo Steel and Wire Co.

Maynard, T. Poole

National Association of Corru-
gated and Fibre Box Manufac-
turers, The

INDUSTRIAL RESEARCH LABORATORIES

117

Pennsylvania Railroad Company,
The

Pierce-Arrow Motor Car Com-
pany, The

Scovill Manufacturing Company

Skinner, Sherman & Esselen, In-
corporated

Stewart- Warner Speedometer Cor-
poration

Swenson Evaporator Company

Union Carbide and Carbon Re-
search Laboratories, Inc.

United Shoe Machinery Corpora-
tion

Public utilities

Detroit Edison Company, The
Doherty Research Company, Em-
pire Division
Stone & Webster, Incorporated

Pulp and paper (cellulose)

American Writing Paper Co.

Andrews, A. B.

Atlas Powder Co.

Babcock Testing Laboratory

Beaver Board Companies, The

Brown Company

Carborundum Company, The

Chemical Economy Company

Chemical Products Company

Chicago Mill and Lumber Com-
pany

Crane ft Co.

Cumberland Mills

Davis Chemical Products, Inc.

Dill & Collins Co.

du Pont, E. I., de Nemours &
Company

Eastern Manufacturing Company

Eastman Kodak Company

Emerson Laboratory

Glysyn Corporation, The

Grosvenor, Wm. M.

Hammerslcy M'f'g Co., The

Heap, William, & Sons

Hooker Electrochemical Company

Industrial Testing Laboratories

Little, Arthur D., Inc.
. MacAndrews ft Forbes Company

Meigs, Bassett ft Slaughter, Inc.

Metals & Chemicals Extraction
Corporation

Munn, W. Faitoute

National Association of Corru-
gated and Fibre Box Manufac-
turers, The

National Laboratories, The

Oliver Continuous Filter Co.

Package Paper and Supply Cor-
poration

Richardson Company, The

Skinner, Sherman ft Esselen, In-
corporated

Strathmore Paper Company

Van Schaack Brothers Chemical
Works, Inc.

Wallace, Joseph H., ft Co.

Weston, Byron, Co.

Wiley & Q>mpany, Inc.

Railroad equipment (cars, loco-
motives, signals, etc.)

Baldwin Locomotive Works, The
Hunt. Robert W., and Ca
Industrial Works
Niles Tool Works Company, The
Pennsylvania Railroad Company,
The

Union Switch ft Signal Company

Razors

Gillette Safety Razor Ca

Refrigeration (artificial ice)

American Radiator Company
Ansul Chemical Company
General Motors Research Corpo-
ration
Industrial Research Corporation
Industrial Testing Laboratories
International Filter Ca

Rubber and rubber goods, includ-
ing other natural gums (gutta-
percha)

Abbott, William G., Jr.
Belden Manufacturing Company
Boonton Rubber Manufacturing
Company

118

INDUSTRIAL RESEARCH LABORATORIES

Brunswick - Bailee - Collender G>.,

The
Carborundum Company, The
Columbia Graphophone Manufac-
turing Company
Dodge Brothers
Falls Rubber Company, The
Firestone Tire & Rubber Company
General Bakelite Company
General Tire & Rubber Co.
Goodrich, B. F., Company, The
Goodyear Tire & Rubber Com-
pany, The
Hood Rubber Company
Manhattan Rubber Mfg. Co., The
Miller Rubber Co., The
Portage Rubber Co., The
Redmanol Chemical Products Co.
Rubber Trade Laboratory, The

Soils and fertilizers (nitrates,
phosphates, potash)

American Agricultural Chemical

Company, The
American Cyanamid Company
American Trona Corporation
Anaconda Copper Mining Co.
Armour Fertilizer Works
Cudahy Packing Co., The
Detroit Testing Laboratory, The
Grasselli Chemical Company
Maynard, T. Poole
Meigs, Bassett & Slaughter, Inc.
Metals & Chemicals Extraction

Corporation
Morris & Company
Royster, F. S., Guano Company
Sears, Roebuck and Co.
Swift & Company
United Chemical and Organic

Products Co.
United States Glue Co.
Utah-Idaho Sugar Company
Wiley & Company, Inc.

Sound (acoustics)

Columbia Graphophone Manufac-
turing Company
Hayes, Hammond V.
Riverbank Laboratories

Sabine, Wallace Clement, Labora-
tory

Steam power (boilers, econ-
omizers, engines, turbines.
See also internal combustion
motors)

American Radiator Company
Babcock & Wilcox Co., The
Bethlehem Shipbuilding Corpora-
tion, Ltd.
Cochrane, H. S. B. W., Corpora-
tion
Commercial Testing and Engineer-
ing Co.
Detroit Edison Company, The
Ingersoll-Rand Company
Lunkenheimer Co., The
Minneapolis Steel and Machinery
Ca

Subatomic phenomena and radio-
activity

Radium Company of Colorado,

Inc., The
Radium Limited, U. S. A.
Riverbank Laboratories
Welsbach Company

Sugar (sorghums, syrups)

American Beet Sugar Company
American Diamalt Company
American Sugar Refining Com-
pany, The
Dehls & Stein
Digestive Ferments Co.
Feculose Co. of America
Great Western Sugar Company,

The
Industrial Chemical Institute of

Milwaukee
Industrial Testing Laboratories
New York Sugar Trade Labora-
tory, Inc., The
Nulomoline Company, The
Oliver Continuous Filter Co.
Penick & Ford, Ltd., Incorporated
Schwarz Laboratories
Spreckels Sugar Company
Swenson Evaporator Company

INDUSTRIAL RESEARCH LABORATORIES

119

U. S. Food Products Corp. •
Utah-Idaho Sugar Company
Western Sugar Refinery

Surgical, dental and hospital
equipment and supplies

Caulk, L. D., Company, The
Johnson & Johnson
Lakeview Laboratories
Ransom & Randolph Co., The
Union Carbide and Carbon Re-
search Laboratories, Inc.

Tar and its products

Barrett Company, The

Glysyn Corporation, The

Koppers Company, The

Laucks, I. F., Inc.

Lyster Chemical Company, Inc.

Providence Gas Company, Incor-
porated

Quinn, T. H., & Company

Rubber Trade Laboratory, The

Union Carbide and Carbon Re-
search Laboratories, Inc.

White Tar Company of New Jer-
sey, Inc., The

Zobel, Ernst, Company, Inc.

Textiles, including machinery
(buttons, clothing, cotton and
its products, linen, wool ; water-
proofing)

Abbott, William G., Jr.

Amoskeag Manufacturing Com-
pany

Arlington Mills

Art in Buttons

Barber-Colman Company

Chemical Products Company

Crompton & Knowles Loom Works

Durfee, Winthrop C.

Eastern Finishing Works, Inc.

Eavenson & Levering Co.

Emerson Laboratory

Glysyn Corporation, The

Grosvenor, Wm, M.

Imperial Belting Company

Industrial Chemical Institute of
Milwaukee

Klearflax Linen Rug Company

Little, Arthur D., Inc.

Maynard, T. Poole

Metakloth Co.

Rochester Button Company

Roessler & Hasslacher Chemical

Company, The
Rubber Trade Laboratory, The
Sears, Roebuck and Co.
U. S. Testing Co., Inc.

Water, sewage and sanitation

American Institute of Baking
Babcock & Wilcox Co., The
Banks & Craig

Borromite Co. of America, The
Borrowman, George
Bridgeman-Russell Company
Cams Chemical Company
Cochrane, H. S. B. W., Corpora-
tion
Dearborn Chemical Company
Detroit Testing Laboratory, The
Dorr Company, The
Emerson Laboratory
Fort Worth Laboratories
Great Western Electro-Chemical

Company
Hochstadter Laboratories
Hooker Electrochemical Company
Industrial Chemical Institute of

Milwaukee
Industrial Testing Laboratories
International Filter Co.
Kidde, Walter, & Company, Incor-
porated
Oliver Continuous Filter Co.
Pease Laboratories
Permutit Company, The
Perolin Company of America, The
Souther, Henry, Engineering Co.,

The
Wallace & Tieman Co., Inc.
Wells, Raymond
Weston & Sampson
Wheeler & Woodruff
White Tar Company of New Jer-
sey, Inc., The

120

INDUSTRIAL RESEARCH LABORATORIES

Welding, autogenous, gas, elec-
tric, forge

Bethlehem Shipbuilding G>rpora-

tion, Ltd.
Davis-Boumonville Company
Electrolabs G)mpany, The
Hoskins Manufacturing G>mpany
Union Carbide and Carbon Re-
search Laboratories, Inc.
Western Gas Construction Com-
pany, The

Wire

Belden Manufacturing Company
Hoskins Manufacturing Company
Kokomo Steel and Wire Co.
Scovill Manufacturing Company

Wood products, other than cellu-
lose and paper (see also con-
tainers)

Andrews, A. B.
Babcock Testing Laboratory
Chicago Mill and Lumber Com-
pany
Florida Wood Products Co.
Hercules Powder Co.
Lakeview Laboratories
Nowak Chemical Laboratories
Quinn, T. H., ft Company
Rodman Chemical Company
Wallace, Joseph H., & Co.
Zobel, Ernst, Company, Inc.

INDUSTRIAL RESEARCH LABORATORIES 121

ADDRESS LIST OF DIRECTORS OF RESEARCH

Abbott, W. a, Jr^ Wilton, N. R

Abrams, Duff A., Structural Materials Research Laboratory, Lewis Institute, 1951

W. Madison St, Chicago, 111.
Adams, H. S^ The Naugatuck Chemical Company, Naugatuck, Conn.
Adams, William H., Eastern Finishing Works, Inc., Kenyon, R. I.
Adamsoo, G. P., General Chemical Company, 25 Broad St., New York, N. Y.
Agnew, Theodore M., Physicians and Surgeons' Laboratory, 605 Paxton Blk.,

Omaha, Nebr.
Alexander, Jerome, Uniform Adhesive Company, Incorporated, foot of 3SHh St.,

Brooklyn, N. Y.
Allen, A. S., The Lennox Chemical Co., Euclid, Ohio.

Amend, C G., Eimer & Amend, Third Ave 18th to 19th Sts., New York, N. Y.
Amend, O. P., Eimer ft Amend, Third Ave., 18th to 19th Sts., New York, N. Y.
Anderegg, G. A., Western Electric Company, Incorporated, 463 West St, New

York, N. Y.
Anderson, John F., E R. Squibb ft Sons, New Brunswick, N. J.
Andrews, A. B., Lewiston, Me.

Angelli Chester M., Vesta Battery Corporation, 2100 Indiana Ave., Chicago, 111.
Anglemyer, Wilbur J., Kellogg Switchboard and Supply Co., Adams and Aberdeen

Sts., Chicago, IlL
Anthony, Olney P., Geo. H. Morrill Co., Norwood, Mass.
Appelbaum, A. I., Thac Industrial Products Corp., 58 Middle Rose St., Trenton,

N.J.
Arms, E W., W. ft L. E Gurley, 514 Fulton St., Trpy, N. Y.
Armstrong, P. A. E, Ludlum Steel Company, Watervliet, N. Y.
Arnold, H. D., Western Electric Company, Incorporated, 463 West St., New York,

N. Y.
Aston, James, A. M. Byers Company, Pittsburgh, Pa.

Atkinson, F. C, American Hominy Company, 1857 Gent Ave., Indianapolis, Ind.
Austin, Frederick J., William R, Warner ft Company, Incorporated, 113 W. 18th

St, New York, N. Y.
Austin, H., Ernest Scott & Company, Fall River, Mass.
Avstreih, L M., Avri Drug & Chemical Company, Inc., 421 Johnston Ave., Jersey

Gty, N. J.
Babcock, S. C, Babcock Testing Laboratory, 803 Ridge Road, Lackawanna, N. Y.
Backhaus, A. A., U. S. Industrial Alcohol Company, South Baltimore, Md.
Badger, W. L., Swenson Evaporator Company, Ann Arbor, Mich.
Baekeland, L. H., General Bakelite Company, Perth Amboy, N. J.
Bailey, G. C, National Aniline & Chemical Company, Incorporated, Marcus Hook, Pa.
Bailey, Herbert S., The Southern Cotton Oil Company, Savannah, Ga.
Baker, J. C, Wallace & Tiernan Co., Inc., Box 178, Newark, N. J.
Balke, Qarence W., Fansteel Products Company, Inc., North Chicago, 111.
Banks, H. P., I. F. Laucks, Inc., 99 Marion St., Seattle, Wash.
Banks, Henry W., Banks ft Craig, 51 East 42nd St., New York, N. Y,
Barad, D. N., A. B. Ansbacher ft Company, 310 N. 7th St, Brooklyn, N. Y.
Barnard, Harry E, American Institute of Baking, 1135 FuUerton Ave., Chicago, 111.
Bartholomew, F. J., Charlotte Chemical Laboratories, Inc, 606 Trust Building,

Charlotte, N. C

122 INDUSTRIAL RESEARCH LABORATORIES

Barton, L. E., Titanium Alloy Manufacturing Co., Niagara Falls, N. Y., also Tita-
nium Pigment Co., Inc., Niagara Falls, N. Y.

Base, Daniel, Hynson, Westcott & Dunning, 16 E. Hamilton St., Baltimore, Md.

Bassett, Harry P., Meigs, Bassett ft Slaughter, Inc., Bala, Pa.

Bassett, William H., The American Brass Company, Waterbury, Conn.

Baxter, Florus R., Vacuum Oil Company, Incorporated, Rochester, N. Y.

Baxter, H. A., Tacony Steel Company, Philadelphia, Pa.

Bean, W. R., Eastern Malleable Iron Company, Naugatuck, Conn.

Beaver, A. B., The National Cash Register Company, Dayton, Ohio.

Bebie, Jules, Monsanto Chemical Works, 1800 South 2nd St., St. Louis, Mo.

Beck, Wesley J., The American Rolling Mill Co., Middletown, Ohio.

Beckman, J. W., Beckman and Linden Engineering Corporation, Balboa Building,
San Francisco, Calif.

Beegle, F. M., The Glidden Company, Qeveland, Ohio.

Bell, W. H., The Coleman ft Bell Company, Norwood, Ohio.

Benedict, C. H., Calumet & Hecla Mining Company, Lake Linden, Mich.

Benger, E. B., E. I. du Pont, de Nemours ft Company, Parlin, N. J.

Bengis, Robert O., Heyden Chemical Company of America, Inc., Garfield, N. J.

Bennetts, B. H., Bennetts' Chemical Laboratory, 1142 Market St., Tacoma, Wash.

Berry, C. W., Laclede-Christy Clay Products Company, 4600 S. Kingshighway, St.
Louis, Mo.

Bierbauer, C. F., Hercules Powder Co., Kenvil, N. J.

Bierbaum, C. H., Lumen Bearing Company, Buffalo, N. Y.

Bigelow, W. D., National Canners Association, 1739 H St. N. W., Washington, D. C.

Bitting, A. W., Glass Container Association of America, 3344 Michigan Ave.,
Chicago, 111.

Black; C. A., The Cleveland Testing Laboratory Co., 511 Superior Building, Qeve-
land, Ohio.

Black, Robert S., Special Chemicals Company, Highland Park, 111.

Blanc, Charles, Cosmos Chemical Co., Inc., 709 Berckman St., Plainfield, N. J.

Bloede, Victor G., Victor G. Bloede Co., Station D, Baltimore, Md.

Boeck, P. A., Celite Products Company, Lompoc, Calif.

Bolton, J. W., The Niles Tool Works Company, 545 North Third St, Hamilton,
Ohia

Bond, William G., Bond Manufacturing Corporation, Monroe and Fifth Sts., Wil-
mington, Del.

Bonnett, F., Jr., Atlas Powder Co., Landing, N. J.

Booth, H. T., Curtiss Aeroplane ft Motor Corporation, Garden City, L. I., N. Y.

Borror, W. A., Pure Oil Company, Belle, W. Va.

Borrowman, George, 130 N. Wells St., Chicago, 111.

Bovard, W. M., Package Paper and Supply Corporation, 150 Birnie Ave., Spring-
field, Mass.

Bowman, Jay, United Chemical and Organic Products Co., W. Hammond, 111.

Boyer, A. D., Boyer Chemical Laboratory Company, 940 N. Clark St., Chicago, 111.

Bradley, Lynde, Allen-Bradley Co., 286 Greenfield Ave., Milwaukee, Wis.

Brady, Edward J., The United Gas Improvement Co., 3101 Passyunk Ave., Phila-
delphia, Pa.

Braude, Felix, Palatine Aniline and Chemical Corporation, 81 N. Water St, Pough-
keepsie, N. Y.

Brenner, R. F., H. C. Fry Glass Company, Rochetser, Pa.

Brewer, J. Ed., The Chemical Service Laboratories, Inc., W. Conshohocken, Pa.

INDUSTRIAL RESEARCH LABORATORIES 123

Breyer, F. G^ The New Jersey Zinc Company, 160 Front St., New York, N. Y.
Bridgman, J. A-, The WUbur White Chemical Co., 62 Temple St., Owego, N. Y.
Briggs, C. H., The Howard Wheat and Flour Testing Laboratory, Old Colony

Building, Minneapolis, Minn.
Brill, A., The Bninswick-Balke-CoUender Ca, Muskegon, Mich.
Brock, F. P., Redmanol Chemical Products Co., 636 W. 22nd St., Chicago, 111.
Brockway, C. P., Industrial Research Corporation, 1025 Front St., Toledo, Ohio.
Brown, M. J., The Roessler & Hasslacher Chemical Company, Perth Amboy, N. J.
Browne, C. A., The New York Sugar Trade Laboratory, Inc., 79 Wall St., New

York, N. Y.
Brownlee, W. K., Buckeye Qay Pot Co., Bassett and Ontario Sts., Toledo, Ohio.
Brunjes, W. G., Dicks David Company, Incorporated, 22nd St. and Stewart Ave.,

Chicago Heights, 111.
Bryson, T. A., Tolhurst Machine Works, Troy, N. Y.
Buchanan, A. J., M. B. Chemical Co., Inc, Johnson City, Tenn.
Bullard, Walter Gould, United Shoe Machinery Corporation, Beverly, Mass.
Burdett, J. B., Burdett Manufacturing Company, St. Johns Court at Fulton Street,

Chicago, 111.
Burdick, A. S., The Abbott Laboratories, Chicago, 111.

Burrage, A. C, Jr., Atlantic Dyestuff Company, 88 Ames Building, Boston, Mass.
Bush, v., American Radio and Research Corporation, Medford, Mass.
Cabot, Samuel, Samuel Cabot, Inc., 141 Milk St., Boston, Mass.
Cady, Francis E., National Lamp Works of General Electric Company, Nela Park,

Qeveland, Ohio.
Calbeck, J. H., The Eagle-Picher Lead Company, 208 S. LaSalle St., Chicago, 111.
Callow, J. M., The General Engineering Company, Incorporated, 159 Pterpont St,

Salt Lake City, Utah.
Campbell, J. H., Robert W. Hunt and Co., 175 W. Jackson Blvd., Chicago, 111.
Campbell, Ross, American Writing Paper Co., Holyoke, Mass.
Carothers, J. N., Federal Phosphorus Company, Anniston, Ala.
Carter, Edgar B., Swan-Myers Company, 219 N. Senate Ave., Indianapolis, Ind.
Carter, F. E., Baker & Co., Inc., Newark, N. J.

Carveth, H. R., The Roessler & Hasslacher Chemical Company, Perth Ambpy, N. J.
Case, Theodore W., Case Research Laboratory, Auburn, N. Y.
Cassady, V. K., The Palmolive Company, Milwaukee, Wis.

Catherman, R. F., C. G. Buchanan Chemical Company, Baker Ave., Norwood, Ohio.
Cheney, G. A., A. P. Munning & Co., Matawan, N. J.

Chittick, J. R., Jaques Manufacturing Company, 16th and Canal Sts., Chicago, 111.
Chormann, O. I., The Pfaudler Co., Rochester, N. Y.
Christie, R. E., Spreckels Sugar Company, 2 Pine St., San Francisco, Calif.
Christison, Hugh, Arlington Mills, Lawrence, Mass.
Qark, Edmund, New England Confectionery Company, 253 Summer St, Boston,

Mass.
Qark, F. C, American Writing Paper Co., Holyoke, Mass.
Clark, J. F., Rochester Button Company, 300 State St., Rochester, N. Y.
Clark, Wm. M., National Lamp Works of General Electric Company, Nela Park,

Geveland, Ohio.
Qements, F. O., General Motors Research Corporation, Box 745, Moraine City,

Dayton, Ohio.
Qevenger, Galen H., United States Smelting, Refining & Mining Company, 55 Con-
gress St., Boston, Mass.

124 INDUSTRIAL RESEARCH LABORATORIES

Qifford, R. K., Kokomo Steel and Wire Co., Kolcomo, Ind.

Qowes, G. H. A., Eli Lilly and Company, Indianapolis, Ind.

Codwise, P. W., Byron Weston Co., Dalton, Mass.

Coleman, A. B., The Coleman & Bell Company, Norwood, Ohio.

Collins, T. R., Pittsburgh Plate Glass Co., Newark, N. J.

Colpitis, E. H., Western Electric Company, Incorporated, 463 West Street, New

York, N. Y.
Comstock, Daniel F., Kalmus, Comstock & Wescott, Inc., 110 Brookline Ave.,

Boston, Mass.
Comstock, G. F., Titanium Alloy Manufacturing Co., Niagara Falls, N. Y., and

Lumen Bearing Company, Buffalo, N. Y.
Condit, P. H., Dicks David Company, Incorporated, 22nd St. and Stewart Ave.,

Chicago Heights, 111.
Conwell, E. L., E. L. Conwell & Co., Inc., 2024 Arch St, Philadelphia, Pa.
Comelison, R. W., Peerless Color Company, Bound Brook, N. J.
Costa, Charles, William R. Warner & Company, Incorporated, 113 W. 18th St.,

New York, N. Y.
Craft, E. B., Western Electric Company, Incorporated, 4(i3 West Street, New York,

N. Y.
Craver, H. H., PiUshurgh Testing Laboratory, 616 Grant St, Pittsburgh, Pa.
Crossley, M. L., The Calco Chemical Company, Bound Brook, N. J.
Cruser, Frederick Van Dyke, The Diamond Match Co., Oswego, N. Y.
Currier, Edward E., T. H. Quinn & Company, E. Smethport, Pa.
Cushman, Allerton S., The Institute of Industrial Research, 19th and B Sts. N. W.,

Washington, D. C
Dahlberg, H. W., The Great Western Sugar Company, Sugar Building, Denver, Colo.
Dale, J. K., U. S. Food Products Corp., Peoria, 111.

Dannerth, Frederic, The Rubber Trade Laboratory, 96 Academy St., Newark, N. J.
Davis, Qarke E., National Biscuit Company, 409 W. Fifteenth St, New York, N. Y.
Dean, J. Atlee, Dean Laboratories, Inc., 48th St and Walton Ave., Philadelphia, Pa.
Delbridge, T. G., The Atlantic Refining Company, 3144 Passyunk Ave., Philadel-
phia, Pa.
Del Mar, William A., Habirshaw Electric Cable Company, Inc., Yonkers, N. Y.
Dengler, F. Peter, Industrial Research Laboratories, 190 N. State St, Chicago, IlL
Dennis, Harold, The Martin Dennis Company, 859 Summer Ave., Newark, N. J.
D'Eustachio, G., Standard Underground Cable Company, 26 Washington St, Perth

Amboy, N. J.
Dewey, Bradley, Dewey & Almy Chemical Company, Harvey St., Cambridge, Mass.
Dicken, C O., E. J. Brach and Sons, 215 W. Ohio St, Chicago, 111.
Dickey, C. B., Corona Chemical Division, Pittsburgh Plate Glass Co., Milwaukee,

Wis.
Dickson, J. C, Inland Steel Company, Indiana Harbor, Ind.
Dixon, A. F., Western Electric Company, Incorporated, 463 West Street, New York,

N. Y.
Doane, S. E., National Lamp Works of General Electric Company, Nela Park,

Qeveland, Ohio.
Dorsey, Frank M., National Lamp Works of General Electric Company, Nela Park,

. Geveland, Ohia
Dotterer, David R., Gibbs Preserving Company, 2303 Bostcxi St., Baltimore, Md.
Downs, C R., The Barrett Company, Edgewater, N. J.

INDUSTRIAL RESEARCH LABORATORIES 125

Drogin, David, The Gray Industrial Laboratories, 961 Frelinghttysen Ave., Newark,
N.J.

Duggan, T. R., The Permutit Company, 440 Fourth Ave., New York, N. Y.

Dunham, Henry G., Digestive Ferments Co., Detroit, Mich.

Dunham, H. V., 50 E. 41st St, New York, N. Y.

Dupont, F. M., Industrial Chemical Institute of Milwaukee, 200 Pleasant St., Mil-
waukee, Wb.

Durfee, Winthrop C, 516 Atlantic Ave., Boston, Mass.

Duschak, L. H., Metals^ & Chemicals Extraction Corporation, 1014 Hohart Bldg.,
San Francisco, Calif.

Edison, Thos. A., Thomas A. Edison Laboratory, Orange, N. J.

Edwards, W. F., U. S. Testing Co., Inc., 316 Hudson St, New York, N. Y.

Eichinger, Benjamin F., Bridgeman-Russell Company, 1100 W. Superior St, Duluth,
Minn.

Eldred, Frank R., Eli Lilly and Company, Indianapolis, Ind.

Elliott, George K., The Lunkenheimer Co., Cincinnati, Ohio.

Ellis, Carleton, Ellis-Foster Company, 92 Greenwood Ave., Montclair, N. J.

Emerson, H. C, Emerson Laboratory, 145 Chestnut St, Springfield, Mass.

Emmons, Frank W., Washburn-Crosby Co., Minneapolis, Minn.

Enfield, W. L., National Lamp Works of General Electric Company, Nela Park,
Qeveland, Ohio.

Engelhardt, Herman, Sharpe & Dohme, Baltimore, Md.

Espenhahn, E. V., The Gray Industrial Laboratories, 961 Frelinghuysen Ave.,
Newark, N. J.

Esselen, Gustavus J., Jr., Skinner, Sherman & Esselen, Incorporated, 248 Boylston
St, Boston 17, Mass.

Eustis, F. A., 131 State St, Boston, Mass.

Fahy, Frank P., 50 Church St, New York, N. Y.

Faile, E. H., The Dorite Manufacturing Company, 116 Utah St, San Francisco^ Calif.

Fash, R. H., Fort Worth Laboratories, Box 1008, Fort Worth, Texas.

Fenn, Herbert B., Metakloth Co., Lodi, N. J.

Ferguson, Louis A., Commonwealth Edison Company, 72 West Adams St., Chi-
cago, 111.

Fippin, E. O., National Lime Association, 918 G St N. W., Washington, D. C.

Fisher, J. P., Doherty Research Company, Empire Division, Bartlesville, Okla.

Fiske, Augustus H, Rumford Chemical Works, Providence, R. I.

FitzGerald, F. A. J., The FiuGerald Laboratories, Inc., Niagara Falls, N. Y.

Fitzgerald, F. F., American Can Company, 120 Broadway, New York, N. Y. .

Fitzgerald, Wm. P., J. T. Baker Chemical Ca, Phillipsburg, N. J.

Flagg, F. P., Waltham Watch Company, Waltham, Mass.

Fleming, R. S., Merrell-Soule Laboratory, Syracuse, N. Y.

Fogh, Carl S., Wedge Mechanical Furnace Company, Greenwich Point, Philadel-
phia, Pa.

Ford, Allen P., Crane Co., South Ave., Bridgeport, Conn.

Forman, L. P., American Window Glass Co., Factory No. 1, Arnold, Pa.

Forrest, Charles N., The Barber Asphalt Paving Company, Philadelphia, Pa.

Fox, H. W., The Krebs Pigment and Chemical Co., Newport, Del.

Francis, Charles K., Cosden & Company, Tulsa, Okla.

Francis, J. M., Parke, Davis & Company, Detroit, Mich.

Frary, Francis C, Aluminum Company of America, New Kensington, Pa.

Frees, Herman E., The H. E. Frees Co., 2528 W. 48th Pkce, Chicago, 111.

126 INDUSTRIAL RESEARCH LABORATORIES

French, D. K., Dearborn Chemical Company, McCormick Building, Chicago, 111.
Frick, F. F., Anaconda Copper Mining Co., Anaconda, Mont
Frickstad, £. T., California Ink Company, Inc., Camelia and 4th Sts., Berkeley, Calif.
Frohring, W. O., The Telling-Belle Vernon Company, 3825 Cedar Ave., Cleveland,

Ohio.
Fry, K J., Davis Chemical Products, Inc., Springfield, N. J.
Fuller, A. D., Dextro Products, Inc., 25 Illinois St., Buffalo, N. Y.
Fuller, A. V., The American Sugar Refining Company, 117 Wall St., New York,

N. Y.
Gage, R. M., The Portage Rubber Co., Barberton, Ohio.

Gane, E. H., McKesson & Robbins, Incorporated, 97 Fulton St., New York, N. Y.
Gardner, H. F., The Beaver Board Companies, Beaver Road, Buffalo, N. Y.
Gatward, W. A., Hoskins Manufacturing Company, Lawton Ave. at Buchanan,

Detroit, Mich.
Geer, W. C, The B. F. Goodrich Company, Akron, Ohio.
Gegenheimer, R. E., The Mathieson Alkali Works (Inc.), Niagara Falls, N. Y.
George, Harry, Chase Metal Works, Waterbury, Conn.
Gcrstle, John, The Electro Chemical Company, Dayton, Ohio.
Gcssler, A. E., Ultro Chemical Corporation, 236 46th St., Brooklyn, N. Y.
Gibbons, John T., Feculose Co. of America, Ayer, Mass.
Gill, James P., The Vanadium-Alloys Steel Co., Latrobe, Pa.
Gilligan, F. P., The Henry Souther Engineering Co., 11 Laurel St., Hartford, Conn.
Ginsburg, S., National Gum & Mica Co., 12 West End Ave., New York, N. Y.
Given, G. C, Atlas Powder Co., Stamford, Conn.
Glancy, Warren E., Hood Rubber Company, Watertown, Mass.
Gnadinger, C. B., McLaughlin Gormley King Co., 1715 Fifth St. S. £., Minneapolis,

Minn.
Goldstein, William, Radiant Dye & Color Works, 2837 W. 21st St., Brooklyn, N. Y.
Goldthwait, Charles F., Klearflax Linen Rug Company, 63rd and Grand Aves., West

Duluth, Minn.
Goodale, Frank, Pure Oil Company, York and McLean Aves., Cincinnati, Ohio.
Graber, Howard T., Digestive Ferments Co., Detroit, Mich.

Gravell, J. H., American Chemical Paint Company, 1126 S. 11th St., Philadelphia, Pa.
Gravely, J. S., Winchester Repeating Arms Co., New Haven, Conn.
Gray, Arthur W., The L. D. Caulk Company, Milford, Del.
Gray, Thomas T., The Gray Industrial Laboratories, 961 Frelinghuysen Ave.,

Newark, N. J.
Greenwood, F. E., Joseph H. Wallace & Co., Webbs Hill, Stamford, Conn.,

R. F. D. 29.
Greenwood, H. D., United States Metals Refining Co., Chrome, N. J.
Grondahl, L. O., Union Switch & Signal Company, Swissvale, Pa.
Gross, E. L., The Pcrolin Company of America, 1112 W. 37th St., Chicago, 111.
Grosvenor, Wm. M., 50 E. 41st St., New York, N. Y.
Grotts, F. W., The Holt Manufacturing Company, Peoria, 111.
Grunenberg, Hubert, Newark Industrial Laboratories, 96 Academy St, Newark, N. J.
Gundlach, H. R., The Interocean Oil Company, East Brooklsm, Baltimore, Md.
Haldenstein, A. A., National Gum & Mica Co., 12 West End Ave., New York, N. Y.
Hale, J. E., Firestone Tire & Rubber Company, Akron, Ohio.
Halley, Clifford D., Acme White Lead & Color Works, Detroit, Mich.
Hamilton, Herbert W., The White Tar Company of New Jersey, Inc., Newark, N. J.
Handy, Jas. O., Pittsburgh Testing Laboratory, 616 Grant St., Pittsburgh, Pa.

INDUSTRIAL RESEARCH LABORATORIES 127

Hanson, H. H., Eastern Manufacturing Company, Bangor, Me.

Hargrove, G. C, The Gray Industrial Laboratories, 961 Frelinghuysen Ave., Newark,

N.J.
Harlow, J. B., Western Electric Company, Incorporated, 463 West Street, New

York, N. Y.
Harris, C. P., Tower Manufacturing Co., Inc., 85 Doremus Ave., Newark, N. J.
Harris, J. W., Western Electric Company, Incorporated, 463 West Street, New

York, N. Y.
Hartmann, M. L., The Carborundum Company, Niagara Falls, N. Y.
Hartong, R. C, The Goodyear Tire & Rubber Company, Akron, Ohio.
Hayes, Hammond V., 84 State St., Boston, Mass.

Heim, F. D., Charles M. Childs & Co., Inc., 41 Summit St., Brooklyn, N. Y.
Heinrich, £. O., Heinrich Laboratories of Applied Chemistry, 1001 Oxford St.,

Berkeley, Calif.
Hendry, W. F., Western Electric Company, Incorporated, 463 West Street, New

York, N. Y.
Hentus, Max, Wahl-Henius Institute, Incorporated, 1135 Fullerton Ave., Chicago, 111.
Heyl, Frederick W., The Upjohn Company, Kalamazoo, Mich.
Higgins, C. H., Sears, Roebuck and Co., Chicago, 111.
Higley, H. V., Ansul Chemical Company, Marinette, Wis.
Hill, R. L., Atlas Powder Co., Reynolds, Pa.

Hillman, V. E., Crompton & Knowles Loom Works, Worcester, Mass.
Hilton, Robert W., The Ault & Wiborg Company, Cincinnati, Ohio.
Hinck, C, Lehn & Fink, Inc., 192 Bloomfield Ave., Bloomfield, N. J.
Hirsch, Alcan, The Hirsch Laboratories, Inc, 593 Irving Ave., Brooklyn, N. Y.
Hirschfield, C. F., The Detroit Edison Company, Detroit, Mich.
Hitchins, Alfred B., Ansco Company, Binghamton, N. Y.
Hobbs, G. M., Sears, Roebuck and Co., Chicago, 111.

Hochstadter, Irving, Hochstadter Laboratories, 227 Frcxit St., New York, N. Y.
Hocker, Ivan S., The National Laboratories, 1313 H St N. W., Washington, D. C.
Holmes, Fletcher B., E. I. du Pont, de Nemours & Company, Box 525, Wilmington,

Del.
Holmes, M. E., National Lime Association, 918 G St. N. W., Washington, D. C.
Holtz, F. C, Sangamo Electric Company, Springfield, 111.

Holz, Robert, The Richardson Company, 26th and Lake Sts., Melrose Park, 111.
Hooker, A. H., Hooker Electrochemical Company, Niagara Falls, N. Y.
Hooper, C. W., H. A. Metz Laboratories, Inc., 642 Pacific St., Brooklyn, N. Y.
Houghton, A. C, Semet-Solvay Co., Syracuse, N. Y.
Houghton, E. M., Parke, Davis & Company, Detroit, Mich.
Houseman, Percy A., MacAndrews & Forbes Company, 3rd St and Jefferson Ave.,

Camden, N. J.
Howard, Frank A., Standard Oil Company, 26 Broadway, New York, N. Y.
Howard, Henry, Grasselli Chemical Company, 130p Guardian Bldg., Cleveland, Ohio.
Howell, Frank B., American Radiator Company, Buffalo, N. Y.
Hoyt, L. F., Larkin Co., 680 Seneca St, Buffalo, N. Y.
Hudson, R. M., The Holt Manufacturing Company, Peoria, 111.
Humble, Joseph M., American Diamalt Company, 419 Plum St., Cincinnati, Ohio.
Hutchinson, W. T., Condensite Company of America, Bloomfield, N. J.
Hyde, Edward P., National Lamp Works of General Electric Company, Nela Park,

Qeveland, Ohio.
Isaacs, A. S., The Northwestern Chemical Co., Marietta, Ohio.

128 INDUSTRIAL RESEARCH LABORATORIES

Jackson, R. P., Westinghouse Electric & Manufacturing Company, East Pitts-
burgh, Pa.

Jacobs, B. R., National Cereal Products Laboratories, 1731 H St. N. W., Wash-
ington, D. C.

James, U. S., James Ore Concentrator Co., 35 Runyon St, Newark, N. J.

Janney, Thomas A., Utah Copper Company, Garfield, Utah.

Jarvis, Ernest G., McNab & Harlin Manufacturing Co., 440 Straight St, Paterson,
N.J.

Jefferson, H. F., Kilboume & Clark Manufacturing Company, Seattle, Wash.

Jenkins, L. A., The Kolynos Co., New Haven, Conn.

Jcwett, F. B., Western Electric Company, Incorporated, 463 West Street, New
York. N. Y.

Johns, C. O., Standard Oil Company, Linden, N. J.

Johnson, Charles Morris, Crucible Steel Company of America, Pittsburgh, Pa.

Jones, Minor C K., Consolidated Gas, Electric Light and Power Company of Balti-
more, Spring Gardens Plant, Baltimore, Md.

Jones, R. L., Western Electric Company, Incorporated, 463 West Street, New
York, N. Y.

Josephson, Edgar, The Pantasote Leather Company, Passaic, N. J.

Judd, C W., Chemical Economy Company, 1640 N. Spring St, Los Angeles, Calif.

Jurrissen, A. W., Martinez Refinery, Shell Ca of California, Martinez, Calif.

Kalmus, Herbert T., Kalmus, Comstock & Wescott, Inc., 110 Brookline Ave., Boston,
Mass.

Kamm, Oliver, Parke, Davis & Company, Detroit, Mich.

Kaplan, Philip, Reliance Aniline & Chemical Co., Incorporated, Poughkeepsie, N. Y.

Kasley, A. T., Westinghouse Electric & Manufacturing Company, Essington, Pa.

Keller, L., Western Electric Company, Incorporated, 463 West Street, New York,
N. Y.

Kellner, Hermann, Bausch & Lomb Optical Company, Rochester, N. Y.

Kersey, K. S., The P. W. Drackett & Sons Co., Cincinnati, Ohio.

Kettering, C F., General Motors Research Corporation, Box 745, Moraine City,
Dayton, Ohio.

Keuffel, Carl, Keuffel & Esser Co., Hoboken, N. J.

Kiefer, H. E., Monroe Drug Company, Bottom Road, Quincy, 111.

Kilbom, K. B., The Goodyear Tire & Rubber Company, Akron, Ohio.

Kilmer, Fred B., Johnson & Johnson, New Brunswick, N. J.

King, W. E., Beebe Laboratories, Inc., 161 3rd St, St Paul, Minn.

Kingsbury, H. P. D., Redlands Fruit Products Company, Redlands, Calif.

Kleimenhagen, Karl, Cants Chemical Company, La Salle, 111.

Kleinfeldt, H. F., Abb6 Engineering Company, 230 Java St, Brooklyn, N. Y.

Klopsteg, Paul E., Central Scientific Company, 460 East Ohio St, Chicago, 111.

Koch, George T., The Ohio Fuel Supply Company, Utica, Ohio.

Kohout, Jerome F., Commercial Testing and Engineering Co., 1785 Old Colony
Bldg., Chicago, 111.

Kolb, Frank P., Bausch & Lomb Optical Company, Rochester, N. Y.

Kraeger, J. F., The Federal Products Company, 7818 Lockland Ave., Cincinnati, Ohio.

Kratz, G. D., The Falls Rubber Company, Cuyahoga Falls, Ohio.

Kraus, Charles E., Kraus Research Laboratories, Inc, 130 Pearl St., New York, N. Y.

Lacy, B. S., The Roessler & Hasslacher Chemical Company, Perth Amboy, N. J.

Lamar, William R., Lyster Chemical Company, Inc., Passaic Junction, N. J.

Landis, W. S., American Cyanamid Company, 511 Fifth Ave., New York, N. Y.

INDUSTRIAL RESEARCH LABORATORIES 129

Landman, Everett S., United States Bronze Powder Works, Inc., Closter, N. J.

Langfeld, Millard, The Cudahy Packing Co., South Side Station, Omaha, Nebr.

Langston, R. E., Wasme Oil Tank and Pump Co., Ft. Wasme, Ind.

Laucks, I. P., I. F. Laucks, Inc., 99 Marion St., Seattle, Wash.

Lavett, Charles, Buffalo Foundry and Machine Co., 1543 Fillmore Ave., Buffalo, N. Y.

Lee, O. Ivan, T. M. & G. Chemical Co., 517 Cortlandt St, Belleville, N. J.

LeTellier, A. M., The Peerless Drawn Steel Company, Massillon, Ohio.

Levi, Louis R, Pfister & Vogel Leather Co., 447 Virginia St., Milwaukee, Wis.

Levin, I. H., The Electrolabs Company, 2635 Penn Ave., Pittsburgh, Pa.

Lewis, Charles H., W. H. Long & Co., Inc., 244 Canal St., New York, N. Y.

Liddell, Donald M., Weld and Liddell, 961 Frelinghuysen Ave., Newark, N. J.

Linch, H. A., The Dorr Comptoy, Westport Mill, Westport, Conn.

Lincoln, E. S., E. S. Lincoln, Inc., 534 Congress St., Portland, Me.

Linden, H. E., Beckman and Linden Engineering Corporation, Balboa Building,
San Francisco, Calif.

Littlefield, E. E., Littlefield Laboratories Co., Seattle, Wash.

Locke, Charles E., Richards & Locke, 69 Massachusetts Ave., Cambridge 39, Mass.

Lockhart, L. B., LocHhart Laboratories, 33^ Auburn Ave., Atlanta, Ga.

Long, C. P., The Globe Soap Company, St. Bernard, Ohio.

Loomis, N. E., Standard Oil Company, Linden, N. J.

Loudenbeck, H. C, Union Switch & Signal Company, Swissvale, Pa.

Luckiesh, M., National Lamp Works of General Electric Company, Nela Park,
Qeveland, Ohio.

Lunn, Charles A., Consolidated Gas Company of New York, Lawrence Point,
Astoria, N. Y.

Lyng, J. J., Western Electric Company, Incorporated, 463 West Street, New York,
N. Y.

Lyon, P. S., H. S. B. W. Cochrane Corporation, 17th and Allegheny Ave., Phila-
delphia, Pa.

Lyster, T. L. B., Hooker Electrochemical Company, Niagara Falls, N. Y.

Maas, Arthur R., A. R. Maas Chemical Company, 306 E. 8th St, Los Angeles, Calif.

Macgregor, Robert W., Ernest Scott & Company, Fall River, Mass.

Magruder, E. W., F. S. Royster Guano Company, Norfolk, Va.

Mailey, R. D., Cooper Hewitt Electric Company, 730 Grand St, Hoboken, N. J.

Malmstrom, A., Wilckes, Martin, Wilckes Company, head of Pine St., Camden, N. J.

Marcus, M. M., Rhode Island Malleable Iron Works, Hillsgrove, R. I.

Markush, Eugene A., Pharma-Chemical Corporation, Baycmne, N. J.

Marsh, W. J., Hooker Electrochemical Company, Niagara Falls, N. Y.

Marshall, A. E., The Davison Chemical Company, Baltimore, Md.

Marx, Ernest A., Pyro-Non Paint Co., Inc, 505 Driggs Ave., Brooklyn, N. Y.

Mathias, L. D., Victor Chemical Works, Fisher Building, Chicago, 111.

May, M. S., Speer Carbon Company, St. Marys, Pa.

May, Otto B., May Chemical Works, 204 Niagara St, Newark, N. J.

Maynard, T. Poole, Atlanta, Ga.

McQave, James M., Western Research Corporation, Incorporated, 514 18th St,
Denver, Colo.

McCleary, F. K, Dodge Brothers, Detroit, Mich.

McCoy, H. N., Lindsay Light Company, 161 E. Grand Ave., Chicago, 111.

McDougal, T. G., Champion Ignition Company, Flint, Mich.

Mcllhiney, Parker C, 50 E. 41st St, New York, N. Y.

McKee, C. R., United States Glue Co., Milwaukee, Wis.

130 INDUSTRIAL RESEARCH LABORATORIES

Mees, C. E. K., Eastman Kodak G>mpany, Rochester, N. Y.

Meredith, S. C, Western Sugar Refinery, foot 23rd St, San Francisco, Calif.

Merka, Paul D., The International Nickel Company, Bayonne, N. J.

Merrill, Edward C, United Drug Company, Boston, Mass.

Merrill, W. H., Underwriters' Laboratories, 207 E. Ohio St., Chicago, 111.

Meston, A. F., The DeLaval Separator Co., 165 Broadway, New York, N. Y.

Metz, G. P., H. A. Metz Laboratories, Inc., 642 Pacific St., Brooklyn, N. Y.

Meyer, A. H., Providence Gas Company, Incorporated, Providence, R. I.

Milbnm, Lessiter C, The Glen L. Martin Company, 16800 St. Clair Ave., Cleveland,
Ohio.

Miles, E. J., The Studebaker Corporation, Detroit, Mich.

Miller, A. H., Midvale Steel and Ordnance Company, Nicetown Works, Philadel-
phia, Pa.

Miller, J., The Pierce-Arrow Motor Car Company, Elmwood Ave., Buffalo, N. Y.

Millner, James A., Imperial Belting Company, 400 N. Lincoln St., Chicago, 111.

Miner, C. S., The Miner Laboratories, 9 S. Qinton St, Chicago, 111.

Miner, Harlan S., Welsbach Company, Gloucester, N. J.

Mitchell, Frank H., Dill & Collins Co., Richmond and Tioga Sts., Philadelphia, Pa.

Mitchell-Roberts, J. F., Oliver Continuous Filter Co., No. 9 Red Lion Passage,
Holbom, London, W. C. L, England.

Mojonnier, J. J., Mojonnier Bros. Co., 73^ W. Jackson Boulevard, Chicago, 111.

Mojonnier, Timothy, Mojonnier Bros. Co., 739 W. Jackson Boulevard, Chicago, III.

Montgomery, John A., The Borromite Co. of America, 54 E. 18th St, Chicago, 111.

Montgomery, John K., Theodore Meyer, 213 S. 10th St, Philadelphia, Pa.

Moody, C. S., Minneapolis Steel and Machinery Co., 2854 Minnehaha Ave., Minne-
apolis, Minn.

Moore, Hugh K., Brown Company, Berlin, N. H.

Moore, Thomas E., The Ransom & Randolph Co., 518 Jefferson Ave., Toledo, Ohio.

Morgan, R. H., Industrial Works, Bay City, Mich.

Mork, H. S., Chemical Products Company, 44 K St., South Boston, Mass.

Morrison, H. J., The Procter & Gamble Co., Ivorydale, Ohio.

Morse, H. E., The Goodyear Tire & Rubber Company, Akron, Ohio.

Morton, H. A., The Miller Rubber Co., Akron, Ohio.

Mothwurf, Arthur F. F., Garfield Aniline Works, Inc., Garfield, N. J.

Mowry, C. W., Factory Mutual Laboratories, 31 Milk St., Boston, Mass.

Mullin, Chas. E., Eavenson*& Levering Co., cor. 3rd and Jackson Sts., Camden, N.J.

Mumford, R. W., American Trona Corporation, Trona, Calif.

Munn, W. Faitoute, 518 Main St, E. Orange, N. J.

Murphy, W. B., F. J. Lewis Manufacturing Co., 2513 S. Robey St, Chicago, III.

Myers, C. N., H. A. Metz Laboratories, Inc., 642 Pacific St, Brooklyn, N. Y.

Myers, R. K, Westinghouse Lamp Ca, Bloomfield, N. J.

Napolitan, Frank J., Davis- Bournonville Company, Jersey City, N. J.

Newlands, J. A., The Henry Souther Engineering Co., 11 Laurel St, Hartford, Conn.

Nichols, B., Schaeffer Brothers & Powell Manufacturing Company, 102 Barton St.,
St. Louis, Mo.

Norman, G. M., Hercules Powder Co., Wilmington, Del.

Northrup, H. B., Diamond Chain & Manufacturing Company, 502 Kentucky Ave.,
Indianapolis, Ind.

Nowak, C. A., Nowak Chemical Laboratories, 518 Chemical Building, St. Louis, Mo.

Oldham, K W., Firestone Tire & Rubber Company, Akron, Ohio.

Oliver, E. L., Oliver Continuous Filter Co., 503 Market St, San Francisco, Calif.

INDUSTRIAL RESEARCH LABORATORIES 131

O'Ncfl, F. W., Ingwsoll-Rand Company, 11 Broadway, New York, N. Y.

Ott, Harry G., Spencer Lens Company, Buffalo, N. Y.

Pack, Charles, Doehler Die-Casting Co., Court, Ninth and Huntington Streets,
Brooklyn, N. Y.

Page, Carl M., Riverbank Laboratories, Geneva, 111.

Palmer, R. C, The Newport Company, Pensacola, Fla.

Palmer, W. R., Columbia Graphophone Manufacturing Company, Bridgeport, Conn.

Pastemack, Richard, Chas. Pfizer & Co., Inc., 11 Bartlett St., Brooklyn, N. Y.

Pease, H. D., Pease Laboratories, 39 West 38th St., New York, N. Y.

Pettee, C L. W.,' Laboratories of Charles L. W. Pettee, 112 High St., Hartford, Conn.

Pfanstiehl, Carl, Special Chemicals Company, Highland Park, 111.

Philipp, H., Dicks David Company, Incorporated, 22nd St. and Stewart Ave., Chi-
cago Heights, 111.

Phillips, P. M., Frank S. Betz Company, Henry and Hoffman Sts., Hammond, Ind.

Phillips, R. O., New York Quebracho Extract Company, Incorporated, Greene and
West Sts., Greenpoint, Brooklyn, N. Y.

Poetschke, Paul, The L. D. Caulk Company, Mil ford, Del.

Porro, Thomas J., Porro Biological Laboratories, 625 Puget Sound Bank Bldg.,
Tacoma, Wash.

Porst, Christian E. G., Com Products Refining Company, Edgewater, N. J.

Porter, F. B., Fort Worth Laboratories, Box 1008, Fort Worth, Texas.

Porter, Horace C, 1833 Chestnut St., Philadelphia, Pa.

Potter, Paul D., Sprague, Warner & Company, 600 West Erie St., Chicago, 111.

Powell, J. R., Armour Glue Works, 31st Place imd Benson St., Chicago, 111.

Pratt, Lester A., Merrimac Chemical Company, North Wobum, Mass.

Pressell, George W., E. F. Houghton & Co., 240 W. Somerset St, Philadelphia, Pa.

Prochazka, John, Central Dyestuff and Chemical Co., Plum Point Lane, Newark, N. J.

Pushee, H. B., General Tire & Rubber Co., Akron, Ohio.

Putnam, W. P., The Detroit Testing Laboratory, 3726 Woodward Ave., Detroit, Mich.

Quinn, Don L., Chicago Mill and Lumber Company, Conway Bldg., Chieago, lit

Ramsdell, Bartlett, Babcock Testing Laboratory, 803 Ridge Road, Lackawanna, N. Y.

Randall, J. E., National Lamp Works of General Electric Company, Nela Park,
Geveland, Ohio.

Redman, L. V., Redmanol Chemical Products Co., 636 W. 22nd St, Chicago, 111.

Reese, Charles L, E. I. du Pont, de Nemours & Company, Wilmington, Del.

Reese, W. J., Peet Bros. Mfg. Co., Kansas City, Kans.

Reichel, John, H. K. Mulford Company, Glenolden, Pa.

Rentschler, H. C, Westinghouse Lamp Co., Bloomfield, N. J.

Rhael, Edward W., Foster-Heaton Company, 27 Badger Ave., Newark, N. J.

Rice, F. E., Nestl^'s Food Company, Incorporated, Ithaca, N. Y.

Richards, Robert H., Richards & Locke, 69 Massachusetts Ave., Cambridge 39, Mass.

Richardson, William D., Swift & Company, Chicago, 111.

Riddle, Frank H., Champion Porcelain Company, Detroit, Mich.

Riker, A., Jr., Butterworth-Judson Corporation, Newark, N. J.

Riley, O. B., Westinghouse Electric & Manufacturing Company, East Pittsburgh, Pa.

Risley, R. R, Stockham Pipe & Fittings Co., Birmingham, Ala.

Robbins, William K., Amoskeag Manufacturing Company, Manchester, N. H.

Roberts, L. E., American Writing Paper Co., Holyoke, Mass.

Robinson, C. I., Standard Oil Company, Linden, N. J.

Rodman, Hugh, Rodman Chemical Company, Verona, Pa.

Roeg, Louis M., Musher and Company, Incorporated, Baltimore, Md.

132 INDUSTRIAL RESEARCH LABORATORIES

Rogers, Allen, Hyco Fuel Products Corporation, Edgewater, N. J.
Rogers, F. H., The William Cramp & Sons Ship & Engine Buildmg Co., Philadel-
phia, Pa.
Rogers, F. M., Standard Oil Company of Indiana, Whiting, Ind.
Rogers, J. S., International Shoe Co., Morganton, N. C.
Romer, J. B., The Babcock & Wilcox Co., Bayonne, N. J.
Rosengarten, George D., The Powers- Weightman-Rosengarten Company, 916 Parrish

St., Philadelphia, Pa.
Rosenstein, Ludwig, Great Western Electro-Chemical Company, 9 Main St., San

Francisco, Calif.
Ross, F. W., Art in Buttons, Incorporated, Rochester, N. Y.
Rother, Willard, Buffalo Foundry and Machine Co., 1543 Fillmore Ave., Buffalo,

N. Y.
Ruddiman, Edsel A., John T. Milliken and Co., 217 Cedar St, St. Louis, Mo.
Ruppel, Henry E. K., Gillette Safety Razor Co., 47 W. 1st St, Boston, Mass.
Rykenboer, E. A., The Roessler & Hasslacher Chemical Company, Perth Amboy, N. J.
Sabine, Paul E., Wallace Clement Sabine Laboratory, Riverbank, Geneva, 111.
Saklatwalla, B. D., Vanadium Corporation of America, Bridgeville, Pa.
Salathe, F., The Western Gas Construction Company, 1429 Buchanan St, Ft Wayne,

Ind.
Sammet, C. Frank, Crane & Co., Dalton, Mass.
Sanborn, Justus C, Strathmore Paper Company, Mittineague, Mass.
Saums, H. L., Pyrolectric Instrument Company, 636 E. State St., Trenton, N. J.
Saunders, Harold F., The Glysyn Corporation, Bound Brook, N. J.
Schaefer, George L., The New York Quinine & Chemical Works, Incorporated,

135 William St, New York, N. Y.
Schenck, P. D., The Duriron Company, Inc., N. Findlay St, Dayton, Ohio.
Schlesinger, W. A., The Radium Company of Colorado, Inc., 18th and Blake Sts.,

Denver, Colo.
Schlichting, Emil, Industrial Testing Laboratories, 402 West 23rd St., New York,

N. Y.
Schmid, M. H., United Alloy Steel Corporation, Canton, Ohio.
Schmidt, A. H., Universal Aniline Dyes and Chemical Ca, 11th and Davis Sts.,

S. Milwaukee, Wis.
Schneller, M. A., The Nulomoline Company, 111 Wall St, New York, N. Y.
Schwartz, H. A., The National Malleable Castings Company, 10600 Quincey Ave.,

Qeveland, Ohio.
Schwarz, Robert, Schwarz Laboratories, 113 Hudson St, New York, N. Y.
Schwenk, N. H., The William Cramp & Sons Ship & Engine Building Co., Phila-
delphia, Pa.
Scott, A. A., Nestl6's Food Company, Incorporated, 130 William St, New York,

N. Y.
Scott, John G., Porro Biological Laboratories, 625 Puget Sound Bank Bldg., Tacoma,

Wash.
Seabury, R. W., Boonton Rubber Manufacturing Company, Boonton, N. J.
Seibert, F. M., Gulf Pipe Line Company, Houston, Texas.
Selke, George H., The Milwaukee Coke & Gas Company, 1st National Bank Bldg.,

Milwaukee, Wis.
Seydel, Paul, Seydel Manufactunng Company, Jersey City, N. J.
Sharp, Clayton H., Electrical Testing Laboratories, 80th St and East End Ave.,

New York. N. Y.

INDUSTRIAL RESEARCH LABORATORIES 133

Sharp, Donald £., Spencer Lens G>nipany, Hamburg, N. Y.

Shcard, Charles, American Optical G)mpany, Southbridge, Mass.

Sbemdal, A. E., H. A. Metz Laboratories, Inc., 642 Pacific St., Brooklyn, N. Y.

Sherwood, C. M., Hercules Powder Co., Brunswick, Ga.

Shively, W. R, The Goodyear Tire & Rubber Company, Akron, Ohio.

Shoeld, M., Armour Fertilizer Works, 209 W. Jackson Blvd., Chicago, 111.

Shrceve, H. E., Western Electric Company, Incorporated, 463 West St., New York,

N. Y.
Simon, Arthur, The Cutler-Hammer Mfg. Co., Milwaukee, Wis.
Simon, C. K., Dye Products & Chemical Company, Inc., 200 5th Ave., New York,

N. Y.
Simons, John P., Saginaw Salt Products Co., Saginaw, Mich.
Singer, Henry H., Radium Limited, U. S. A., 2 W. 45th St., New York, N. Y.
Singmaster, J. A., The New Jersey Zinc Company, 160 Front St, New York, N. Y.
Skidgell, Chas. E., International Silver Company, Meriden, Conn.
Skinner, C. K, Westinghouse Electric & Manufacturing Company, East Pitts-
burgh, Pa.
Skowronski, S., Raritan Copper Works, Perth Amboy, N. J.
Sladek, George E., Beaver Falls Art Tile Company, Beaver Falls, Pa.
Slaght, W., The Pierce-Arrow Motor Car Company, Elmwood Ave., Buffalo, N. Y.
Smith, E. B., Florida Wood Products Co., Jacksonville, Fla.

Smith, Irving B., Leeds & Northrup Company, 4901 Stenton Ave., Philadelphia, Pa.
Smith, R. B., Hercules Powder Co., Emporium, Pa.
Smith, W. C, United States Metals Refinuig Co., Chrome, N. J.
Snell, H. Sterling, William Heap & Sons, Grand Haven, Mich.
Snook, H. C, Western Electric Company, Incorporated, 463 West St., New York,

N.Y.
Speller, F. N., National Tube Company, Frick Building, Pittsburgh, Pa.
Sperr, F. W., Jr., The Koppers Company, Pittsburgh, Pa.
Sperry, D. R., D. R. Sperry & Co., Batavia, 111.
Spring, L. W., Crane Co., 836 South Michigan Ave., Chicago, 111.
Squier, C W., The Harrison Mfg. Co., 55 Union St, Rahway, N. J.
Stanforth, Richard, Art in Buttons, Incorporated, Rochester, N. Y.
Stein, L., Dehls & Stein, 237 South St, Newark, N. J.
Stevens, A. L., Lakeview Laboratories, 2 Jersey St, Buffalo, N. Y.
Stevenson, Earl P., Arthur D. Little, Inc., 30 Charles River Road, Cambridge 39,

Mass.
Stoddard, W. B., Hochstadter Laboratories, 227 Front St., New York, N. Y.
Strong, W. W., The Scientific Instrument and Electrical Machine Company, 500 S.

York St., Mechanicsburg, Pa.
Stull, W. N., Mallinckrodt Chemical Works, St Louis, Mo.
Stupp, C. G., The Barrett Company, Edgewater, N. J.
Sturtevant, W. L., The Manhattan Rubber Mfg. Ca, Passaic, N. J.
Styri, Haakon, S. K. F. Industries, Inc., Front St and Erie Ave., Philadelphia, Pa.
Sullivan, E. C, Coming Glass Works, Coming, N. Y.
Sundstrom, Carl, The Solvay Process Company, Syracuse, N. Y.
Sutermeister, E., Cumberland Mills, Cumberland Mills, Me.
Swart, W. G., Mesabi Iron Company, Babbitt, Minn.
Taber, Harry P., American Chemical and Manufacturing Corporation, Cranford,

N.J.

134 INDUSTRIAL RESEARCH LABORATORIES

Taggart, Arthur F., Taggart and Yerxa, 165 Divisicxi St., New Haven, Conn.

Takamine, Jokichi, Takamine Laboratory, Inc., Gifton, N. J.

Tanberg, A. P., E. I. du Pont, de Nemours & Company, Henry Qay, Del.

Taub, Joel, The Utility Color ft. Chemical Co.. 395 Frelinghuysen Ave., Newark, N. J.

Teeple, John E., SO E. 41st St.. New York, N. Y.

Temple, Sterling, The Roessler ft Hasslacher Chemical Company, Perth Amboy. N. J.

Thomas. John F., Berry Brothers, Inc., Detroit, Mich.

Thompson, Firman, Bowker Insecticide Company, Everett. Mass.

Thompson, Gustave W., National Lead Company, 129 York St., Brooklyn, N. Y.

Thorburn, A. D., Swan-Myers Company, 219 N. Senate Ave., Indianapolis, Ind.

Thurston, S. R., Bethlehem Shipbuilding Corporation, Ltd., Union Plant, San Fran-
cisco, Calif.

Titus, E. G., Utah-Idaho Sugar Company, Salt Lake City, Utah.

Toch. Maximilian, Toch Brothers, 320 Fifth Ave., New York, N. Y.

Tolman, L. M., Wilson & Co., Chicago, 111.

Uhlig, E. C, The Brooklyn Union Gas Company, 176 Remsen St, Brooklyn, N. Y.

Unger, J. S., Carnegie Steel Company, 1054 Frick Annex Building, Pittsburgh, Pa.

Vail, James G., Philadelphia Quartz Company, Philadelphia, Pa.

Van Buskirk, J. V., Belden Manufacturing Company, 23rd St. and Western Ave.,
Chicago, 111.

Van Marie, D. J., Buffalo Foundry and Machine Co., 1543 Fillmore Ave., Buffalo,
N. Y.

Van Schaack, R. H., Jr., Van Schaack Brothers Chemical Works, Inc., 3358 Avon-
dale Ave., Chicago, 111.

Vollertsen, J. J., Morris ft Company, Union Stock Yards, Chicago, 111.

Vosburgh, Warren C, The Eppley Laboratory, 12 Sheffield Ave., Newport, R. I.

Walker, R. Gordon, Oliver Continuous Filter Co., 226 E. 41st St., New York, N. Y.

Wallace, C. F., Wallace ft Tiernan Co., Inc., Box 178, Newark, N. J.

Walters, A. L., Eli Lilly and Company, Indianapolis, Ind.

Watkins, J. A., American Blower Company, 6004 Russell St., Detroit, Mich.

Webster, W. R., Bridgeport Brass Company, Bridgeport, Conn.

Weir, J. W., Ventura Refining Company, Fillmore, Calif.

Weirick, Elizabeth, Sears, Roebuck and Co., Chicago, 111.

Weiss, J. M., The Barrett Company, 40 Rector St., New York, N. Y.

Weith, A. J., Redmanol Chemical Products Co., 636 W. 22nd St., Chicago, 111.

Welch, H. v., Western Precipitation Company, 1016 W. Ninth St., Los Angeles,
aiif.

Wells, Raymond, Homer, N. Y.

Wescott, E. W., Kalmus, Comstock ft Wescott, Inc., 110 Brookline Ave., Boston,
Mass.

Westoo, Robert S., Weston & Sampson, 14 Beacon St., Bpston, Mass.

Wheat, J. C, Industrial Works, Bay City, Mich.

Wheeler, H. J., The American Agricultural Chemical Company, Carteret, N. J.

Wheeler, T. L., Wheeler ft Woodruff, 280 Madison Ave., New York, N. Y.

Whitney, Willis R., General Electric Company, Schenectady, N. Y.

WhiUen, G. R., The J. O. Whitten Company, Cross St., Winchester, Mass.

Whittington, F. G., Stewart- Warner Speedometer Corporation, Chicago, 111.

Wiley, Samuel W., Wiley ft Company, Inc., 904 N. Calvert St., Baltimore, Md.

Wille, H. v.. The Baldwin Locomotive Works, Philadelphia, Pa.

Williamson, A. M., Acheson Graphite Company, Niagara Falls, N. Y.

Wilson, C P., California Fruit Growers Exchange, Box 518, Corona, Calif.

INDUSTRIAL RESEARCH LABORATORIES 135

Wilson, £. A., E. I. du Pont, de Nemours & Company, Arlington, N. J.

Wilson, Fred D., The National Association of Corrugated and Fibre Box Manufac-
turers, 1821 Republic Building, Chicago, 111.

Wilson, John Arthur, A. F. Gallun & Sons Co., Milwaukee, Wis.

Winther, H., Industrial testing Laboratories, 402 West 23rd St, New York, N. Y.

Wisdom, Roy H., Stamford Dyewood Company, Stamford, Conn.

Witt, J. C, Structural Materials Research Laboratory, Lewis Institute, 1951 W.
Madison St., Chicago, 111.

Woiski, B., Lumen Bearing Company, Buffalo, N. Y.

Wolfe, Wm. S., The Goodyear Tire & Rubber Company, Akron, Ohio.

Wolgemuth, L. E., Sears, Roebuck and Co., Chicago, 111.

Woltersdorf, A. H., Pittsburgh Plate Glass Co., Milwaukee, Wis.

Woodbury, C. A., E. I. du Pont, de Nemours & Company, Box 424, Chester, Pa.

Woolson, L. M., Packard Motor Car Company, Detroit, Mich.

Worth, Barzillai G., Walter Kidde & Company, Incorporated, 140 Cedar St., New
York, N. Y.

Yerxa, R. B., Taggart and Yerxa, 165 Division St., New Haven, Conn.

Youngman, R. H., Harbison-Walker Refractories Company, Farmers Bank Bldg.,
Pittoburfl^, Pa.

Zerban, F. W., Penick & Ford, Ltd., Incorporated, Marrero, La.

Zimmerman, K C, Firestone Tire & Rubber Company, Akron, Ohio.

Zimmerman, R. £., American Sheet and Tin Plate Company, 210 Semple St., Pitts-
burgh, Pa.

Zimmermann, F., Baker & Co., Inc., Newark, N. J.

Zinsser, J. S., Zinsser & Ca, Hastings-on-Hudson, N. Y.

Zobel, F. C, Ernst Zobel Company, Inc., 104 2nd Ave., Brooklyn, N. Y.

Zurbrigg, D. Anton, The L. D. Caulk Company, Milford, Del.

Bulletin of the National Research Council

Volume 1

Number 1. The national importance of scientific and industrial research. By
George Ellery Hale and others. October, 1919. Pages 43. Price 50 cents.

Nninber 2. Research laboratories in industrial establishments of the United
States of America. Compiled by Alfred D. Flinn. March, 1920. Pages 85.
Price $1.00.

Number 3. Periodical bibliographies and abstracts for the scientific and tech-
nological journals of the world. Compiled by R. Cobb. June, 1920. Pages
24. Price 40 cents.

Number 4. North American forest research. Compiled by the Committee on
American Forest Research, Society of American Foresters. August, 1920.
Pages 146. Price $2.00.

Number 5. The quantum theory. By Edwin P. Adams. October, 1920. Pages
81. Price $1.00.

Number 6. Data relating to X-ray spectra. By William Duane. November,
1920. Pages 26. Price SO cents.

Number 7. Intensity of emission of X-rays and their reflection from crystals.
By Bergen Davis. Problems of X-ray emission. By David L. Webster.
December, 1920. Pages 47. Price 60 cents.

Number 8. Intellectual and educational status of the medical profession as
represented in the United States Army. By Margaret V. Cobb and Robert
M. Yerkcs. February, 1921. Pages 76. Price $1.00.

Volume 2

Number 9. Funds available in 1920 in the United States of America for the en-
couragement of scientific research. Compiled by Callie Hull. March, 1921.
Pages 81. Price $1.00.

Number 10. Report on photo-electricity including ionizing and radiating poten-
tials and related effects. By Arthur Llewelyn Hughes. April, 1921. Pages
87. Price $1.00.

Number 11. The scale of the universe. Part I by Harlow Shapley. Part II by
Heber D. Curtis. May, 1921. Pages 47. Price 60 cents.

Number 12. Cooperative experiments upon the protein requirements for the
growth of cattle. First report of the Subcommittee on Protein Metabolism
in Animal Feeding. By Henry Prentiss Armsby, Chairman. June, 1921.
Pages 70. Price $1.00.

Number 13. The research activities of departments of the state government of
California in relation to the movement for reorganization. By James R.
Douglas. June, 1921. Pages 46. Price 60 cents.

Number 14. A general survey of the present status of the atomic structure
problem. Report of the Committee on Atomic Structure of the National
Research Council. By David L. Webster and Leigh Page. July, 1921.
Pages 61. Price 75 cents.

Number 15. A list of seismological stations of the world. Compiled by Harry
O. Wood. July, 1921. Pages 142. Price $2.00.

Volume 3

Number 16. Research laboratories in industrial establishments of the United
States, including consulting research laboratories. Originally compiled by
Alfred D. Flinn; revised and enlarged by Ruth Cobb. December, 1921.
Page 135. Price $2.00.

and Circular Series of the National
Research Council

Number 1. Report of the Patent Committee of the National Research Council.
Presented for the Committee by L. H. Baekeland, Acting Chairman.
February, 1919, Pages 24. Price 30 cents.

Number 2. Report of the Psychology Committee of the National Research
Council. Presented for the Committee by Robert M. Yerkes, Chairman.
March, 1919. Pages 51. Price 60 cents.

Number 3. Refractory materials as a field for research. By Edward W. Wash-
burn. January, 1919. Pages 24. Price 30 cents.

Number 4. Industrial research. By F. B. Jewett. 1918. Pages 16. Price 25
cents.

Number 5. Some problems of sidereal astronomy. By Henry N. RusselL
October, 1919. Pages 26. Price 30 cents.

Number 6. The development of research in the United States. By James
Rowland AngelL November, 1919. Pages 13. Price 25 cents.

Number 7. The larger opportunities for research on the relations of solar and
terrestrial radiation. By C. G. Abbot. February, 1920. Pages 14. Price
20 cents.

Number 8. Science and the industries. By John J. Carty. February, 1920.
Pages 16. Price 25 cents.

Number 9. A reading list on scientific and industrial research and the service of
the chemist to industry. By Clarence Jay West. April, 1920. Pages 45.
Price 50 cents.

Number 10. Report on organization of the International Astronomical Union.
Presented for the American Section, International Astronomical Union, by
W. W. Campbell, Chairman, and Joel Stebbins, Secretary. June, 1920.
Pages 48. Price 50 cents.

Number 11. A survey of research problems in geophysics. Prepared by Chair-
men of Sections of the American Geophysical Union. October, 1920.
Pages 57. Price 60 cents.

Number 12. Doctorates conferred in the sciences by American universities in

1920. Compiled by Callie Hull. November, 1920. Pages 9. Price 20 cents.

Number 13. Research problems in colloid chemistry. By Wilder D. Bancroft.
January-April, 1921. Pages 54. Price 50 cents.

Number 14. The relation of pure science to industrial research. By John J.
Carty. October, 1916. Pages 16. Price 20 cents.

Number 15. Researches on modern brisant nitro explosives. By C. F. van
Duin and B. C. Roeters van Lennep. Translated by Charles E. Munroe.
February, 1920. Pages 35. Price 50 cents.

Number 16. The reserves of the Chemical Warfare Service. By Charles H.
Herty. February, 1921. Pages 17. Price 25 cents.

Number 17. Geology and geography in the United States. By Edward B.
Mathews and Homer P. Little. April, 1921. Pages 22. Price 20 cents.

Number 18. Industrial benefits of research. By Charles L. Reese and A. J.
Wadhams. February, 1921. Pages 14. Price 25 cents.

Number 19. The university and research. By Vernon Kellogg. June, 1921.
Pages 10. Price 15 cents.

Number 20. Libraries in the District of Columbia. Compiled by W. I. Swan-
ton. June, 1921. Pages 19. Price 25 cents.

Number 21. Scientific abstracting. By Gordon S. Fulcher. September, 1921.
Pages 15. Price 20 cents.

Number 22. The National Research Council. Its services for mining and
metallurgy. By Alfred D. Flinn. October, 1921. Pages 7. Price 20 cents.

Number 23. American research chemicals. By Clarence J. West September,

1921. Pages 28. Price 50 cents.

r.vv i.. .., 'A \ « ^ * » ' I « O

Vol.3. Part 2 MARCH, 1922 Number 17

Bulletin

OF THE

National Research

Council

SaENTIFIC papers PRESENTED

BEFORE THE AMERICAN GEOPHYSICAL UNION

AT ITS SECOND ANNUAL MEETING

Published by The National Rbseasch G)uncil

OF

The National Academy of Sciences

Washington, D. C.

1922

Announcement Concerning Publications

of the

National Researcli Council

The Proceedings of the National Academy of Sciences

has been designated as the official organ of the National
Research Council for the publication of accounts of research,
committee and other reports, and minutes.

Subscription rate for the "Proceedings" is $5 per year.
Business address: Home Secretary, National Academy of
Sciences, Smithsonian Institution, Washington, D. C.

The Bulletin of the National Research

presents contributions from the National Research Council,
other than proceedings, for which hitherto no appropriate
agencies of publication have existed.

The "Bulletin" is published at irregular intervals. The
subscription price, postpaid, is $5 per volume of approxi-
mately 500 pages. Numbers of the "Bulletin" are sold
separately at prices based upon the cost of manufacture (for
list of bulletins see third cover page).

The Reprint and Circular Series of the National

Research Council

renders available for purchase, at prices dependent upon the
cost of manufacture, papers published or printed by or for
the National Research Council (for list of reprints and circu-
lars see fourth cover page).

Orders for the "Bulletin" or the "Reprints and Circulars"
of the National Research Council, accompanied by remit-
tance, should be addressed: Publication Office, National
Research Council, 1701 Massachusetts Avenue, Washington,
D. C

BULLETIN

OF THE

NATIONAL RESEARCH COUNCIL

Vcrf. 3, Part 2 MARCH, 1922 ' Number 17

SCIENTIFIC PAPERS

PRESENTED BEFORE THE AMERICAN GEOPHYSICAL

UNION AT ITS SECOND ANNUAL MEETING

CONTENTS
Section of Geodesy

Measurement of gravity at sea. A review. By Lyman J. Briggs 3

Isostasy. By John F. Hayford 11

The earth-tide experiment. By Henry G. Gale 16

The Eotvos balance. By W. D. Lambert 17

The problem of the earth tides. By W. D. Lambert 18

Section of Meteorology

Solar radiation and terrestrial phenomena. By C. G. Abbot 27

Relations between solar activity and its various aspects, and the phenomena

of terrestrial weather. By C. F. Marvin 31

Daily meteorological charts of the world. By Edward H. Bowie 36

World aerology. By Willis Ray Gregg 41

World digest of meteorological data. By W. J. Humphreys 49

General adoption of the centesimal system of angular measurement with

application to anemometers and nephoscopes. By Alexander McAdie... 50
Section of Terrestrial Magnetism and Electricity

A sine galvanometer for determining in absolute measure the horizontal

intensity of the earth's magnetic field. By S. J. Barnett 54

Activity of the earth's magnetism in 1915. By D. L Hazard 55

On measures of the earth's magnetic and electric activity and correlations

with solar activity. By Louis A. Bauer 59

The penetrating radiation and its bearing upon the earth's electric field. By

W. F. G. Swann 65

Recent results derived from the diurnal-variation observations of the

atmospheric-electric potential-gradient on board the Carnegie. By

S. J. Mauchly 73

Section of Physical Oceanography

Suggestions relative to the application of mathematical methods to certain

basic problems in dynamic oceanography. By G. F. McEwen 78

State of progress in continuous recording oceanographical instruments. By

Albert L. Thuras 82

1

CONTENTS

Present status of researches on marine sediments in the United States. By
Thomas Wayland Vaughan 85

The intervab that should obtain between deep-sea soundings to disclose the
orography of the ocean basins. By G. W. Litdehales 90

New methods of observing winds at flying levels over the ocean. By Alex-
ander McAdie 94

The steering line of hurricanes. By Alexander McAdie 102

^

These papers were presented at the second annual meeting of the
American Geophysical Union held at the National Research Council,
Washington, D. C, April 18, 19 and 20, 1921. The names of three of
the sections (seismology, volcanology and geophysical chemistry) do not
appear in the table of contents, as these meetings were devoted to discus-
sion and no scientific papers were presented.

The American Geophysical Union is the Committee on Geophysics of
the National Research Council and is the National Committee for the
United States of the International Geodetic and Geophysical Union.

MEASUREMENT OF GRAVITY AT SEA

A Review

r

' By Lyman J. Buggs

METHODS OF DETERMINING GRAVITY AT SEA

The measurement of the acceleration of gravity over the oceans is a
matter of interest to the geodesist in the determination of the figure of th<^
geoid and in investigations relating to isostatic compensation. For the
requirements of these problems it is desirable that the probable error of
the gravity determinations should not exceed one part in 50,000. At first
sight this accuracy does not seem to be unattainable, particularly to those
who are familiar with the remarkable work that has been done with inva-
riable pendulums on land stations, where the probable error has been
reduced to. two or three parts in a million. But when we consider that in
measurements at sea the vertical acceleration of the ship is imposed upon
the gravitational acceleration which we are trying to measure, the difficulty
of the problem becomes apparent. For example, in the case of a ship
which rises and falls through a height of a meter during a period of 10
seconds, the average vertical acceleration without regard to sign is about
0.004g, or 200 times the permissible probable error of the measurement.
Such vertical accelerations of the ship are not uncommon, although the
sea is sometimes so smooth that an index point may be set in grazing
contact with the surface of a mercury column with almost laboratory
precisioiL

The oscillations of the sensitive element due to the vertical accelerations
of the ship may be controlled by damping, but it is doubtful whether the
damped system gives the same result as when vertical accelerations are
absent. This question cannot be answered from observations at sea, and
necessitates laboratory methods of testing gravity apparatus which will
be referred to later.

Virtually all the methods that have actually been used for gravity deter-
minations at sea involve and depend upon observations of the length of a
mercurial column supported by gas pressure. The pressure p of the gas
may be equated to p g h where p and h refer to the density and height of
the mercury column. Knowing p, p and h, the acceleration of gravity g
can at once be evaluated. The various methods differ in the manner in
which p and h are determined.

Siemens' method: The first actual measurements of the variation of
gravity at sea appear to have been made by Sir Wm. Siemens* in 1875.
He believed that at sea the value of g was diminished by an amount very
nearly proportional to the depth of the ocean, and his primary purpose

^C. Wm. Siemens. On determining the depth of the sea without the use of a
sounding line. PM, Trans,, 167, 1877, 671-692.

3

4 GEODESY

was to develop a sounding apparatus on this principle as an aid to navi-
gation.

His first instrument consisted of a barometer with a large sealed-off
air chamber to eliminate the effect of variations in atmospheric pressure
(fig. 1). The barometric column included three liquids — mercury, diluted
alcohol and juniper oil. The vertical column was expanded into a bulb
b, at the level where the mercury and alcohol surfaces were in contact,
with a second evacuated bulb d at the top, which contained the free
juniper-oil surface. The readings were made on the position of the
'alcohol-oil surface c in the constricted tube joining the two bulbs. With
this device the scale deflection for a given change in g was 300 times that
of a simple mercurial column.

This instrument proved unsatisfactory and was abandoned in favor of
a second "bathometer," which consisted of a steel spring balance of a
peculiar type. A vertical steel tube was fitted with reservoirs at top and
bottom, the floor of the lower reservoir consisting of a thin corrugated
steel diaphragm. This system contained mercury, the free surface of the
mercury being in the upper reservoir, which during measurements was
open to the atmosphere. The load on the diaphragm was carried by two
long steel spiral springs connected to a yoke beneath the diaphragm and
suspended from the upper reservoir. The mercury constituted the load
on the balance and at the same time served to damp the oscillations
through the action of a constriction in the tube just below the upper
reservoir. The change in the load due to a change in g resulted in a
vertical displacement of the yoke and was measured by means of a mi-
crometer screw supported from the lower reservoir. The observations
required corrections for temperature and atmospheric buoyancy. The
instrument was not checked by testing it at two land stations where the
relative value of g was accurately known. His results show, however,
a remarkable correlation with the depth as obtained by direct soundings,
which were made immediately after the bathometer readings. In a series
of about 30 observations involving depths up to about 2,500 fathoms, the
discrepancy was seldom more than 10 per cent. No corrections were
applied for variation in latitude, which ranged from 45* to 49* N.

Hecker's fnethod: The appearance of Helmert's equation for the varia-
tion of g with latitude, based on land stations, led Hecker^ to undertake
the task of providing data to test its validity for sea stations as well. In
1901 he began an extended series of gravity measurements at sea which
eventually included systematic observations in the Atlantic, Pacific and
Indian oceans and in the Black Sea.

In using a barometric column for measuring g, two procedures are
available : ( 1 ) the air chamber at the base of the column may be sealed,

^ For a description of the apparatus empioyed by Hecker and a summary of hit
ocean measurements, see Hecker, C, Bestimmune der Schwerkraft auf dem
Schwarzen Meere und, an dessen Kiiste sowie neue Ausgleichung der Schwerkraft-
messungen auf dem Atlantischen, Indischen, und Groszen Ozean, Zentralbur. Inter-
not. Erdmessung^, Veroiientlichungen Berlin, N. F., Nr. 20 (1910).

b nkf

t

Fic.1. Sfcroen.* F«g. 2. DuffieWs

first apparatus. apparatus.

^.

Fic 3. Briggs' apparatus.

6 GEODESY

as in Siemens' first instrument, in which event we eliminate variations in
pressure due to atmospheric changes, but are left with an instrument
which is very sensitive to temperature changes, as it becomes in effect a
gas thermometer; or (2) we may leave the air-chamber open to the
atmosphere ; in this case the temperature effects are greatly reduced, but
it is necessary to measure the air-pressure p by some independent means.

Hecker in his pioneer investigations chose the latter procedure and
determined the pressure from boiling-point measurements of water re-
ferred to vapor-pressure tables. The vapor-pressure of water increases
very rapidly with temperature in the neighborhood of the boiling-point, —
approximately one twenty-eighth of an atmosphere per degree. There-
fore, in order to determine the pressure to 1 part in 50,000 it is necessary
to know the temperature interval from freezing-point to boiling-point with
an error not greater than 0^.0006 C.

Hecker used mercury thermometers in his boiling-point determinations.
He had no means of checking the fundamental intervals of these ther-
mometers at sea, for this determination depends upon accurate barometric
pressure measurements which can be obtained only at land stations where
the value of g is known. Consequently any departure of the fundamental
interval from that determined at land stations enters directly into the
boiling-point determinations at sea as a systematic error. Furthermore,
the best of mercurial thermometers exhibit variations in the fundamental
interval. For example, Waidner and Dickinson^ found that the funda-
mental interval of the primary mercurial standards of the Bureau of
Standards varied through a range of 0^.015 C. during a ten-day period,
which would correspond to a variation of more than one part in 2,000
in the value of g. The probable error of the fundamental interval deter-
minations in Waidner and Dickinson's measurements under favorable
laboratory conditions was dbO^'.OOS C, which may be taken as a measure
of the maximum refinement obtainable in barometer-hypsometer measure-
ments aboard ship ; and this corresponds to a probable error in the value
of g of more than one part in 10,000.

If hypsometer determinations are to be made, resistance thermometry
would be preferable to mercurial thermometry, since the resistance ther-
mometer is more sensitive and shows less variation from day to day in
the fundamental interval. The steam point of a resistance thermometer
can be readily determined under laboratory conditions to 0^.002 C, but
whether this accuracy could be obtained on board ship with the galvano-
meter on an unstable base is questionable.

In my opinion, the barometer-hypsometer method is not the most prom-
ising way of attacking the problem of measuring g at sea, because (1) the
method involves two operations; (2) the temperature errors in hypsome-
try lead to errors in the derived value of g which are ten times as great
as those produced by equal temperature errors in a closed system; and
(3) the motion of the ship relative to the air produces a change in baro-

^Bull Bur. Standards, 266^, Washington, D. C (1907).

GEODESY 7

metric pressure below deck. If this relative motion is unsteady from any
cause, as for example variable winds, errors may result unless the barom-
eter and hypsometer are read simultaneously. Dufiield (1921) on board
a destroyer observed pressure effects of this kind as large as one millibar.

The use of a sensitive aneroid barometer has been proposed as a substi-
tute for hypsometric measurements, but here again a double operation is
involved. If a spring system could be devised which would be sufficiently
sensitive and reliable to measure pressure to the required degree of accu-
racy, it would be better to employ it directly as a force balance to measure
the change in g than to equate the observed pressure to the observed
length of a mercury column, for the double operation serves only to
increase the probable error of the final result.

Duffield's method: Duffield^ in 1914 employed the apparatus shown in
figure 2 in some preliminary measurements of g during a voyage from
Australia to England. The apparatus, which is of the sealed gas-chamber
type, possesses a unique and valuable temperature-compensation feature.
A constant volume of air is maintained in the bulb B by keeping the
mercury always up to the electrical contact at C. The air in the bulb B
is under reduced pressure in order to reduce the length of the apparatus.
The barometer tube is bent so that H is vertically above C, the length of
the column HC being approximately 20 centimeters. The mercury level is
kept at C by raising or lowering the mercury in the index tube D. This
operation is effected by slowly exhausting or admitting air through F as
required. The index tube D is of fine bore and the value of g is calculated
from readings upon the level of the mercury in this tube when contact is
made with the pointer at C. The side tube E is used only for the purpose
of making initial adjustments, and to permit the apparatus to be used for
various ranges of temperature.

The reservoir of mercury R is introduced for the purpose of tempera-
ture compensation and when the dimensions of the apparatus are suitably
chosen the increased pressure of air in the bulb B due to a given rise of
temperature is automatically counterbalanced by the rise of the level H,
occasioned by the expulsion of mercury from the reservoir. The com-
pensation is perfect at only one temperature, but for small departures
the error is small. At sea the apparatus was immersed in a water bath
which was hung by cords from the ceiling of the refrigerator room of the
ship, and readings on the index tube were taken through a window in
the side.

Duffield's apparatus as reconstructed in Australia was over compen-
sated, an increase in temperature of 1** C. necessitating the removal of a
thread of mercury 60 mm. long. In other words, a temperature change
of this amount resulted in a change in the reading of the index scale
corresponding to the computed change in g in going from the equator to
54® N. Lat. Trouble was also experienced from a break in the capillary

* W. G. Diiffield. Apparatus for the determination of gravity at sea. Proc, Roy.
Soc. Land, (A), 112, 1916, 505-517.

8 GEODESY

mercurial column in the barometer tube. Owing to these difficulties actual
observations were limited to a series of 25 preliminary measurements in
the Indian Ocean from Lat. 0° to 16** N. These results show an average
deviation with regard to sign of H-O-Ol cm./sec* from values computed
from Helmert's equation; in other words, a chance distribution of the
observations about the line representing Helmert's computed values is
indicated. The average deviation of the observations without regard to
sign is 0.14 cm./sec*. The anomalies are, however, not known. If we
assume that there are no anomalies, this corresponds to an average error
of 1.4x10""^. It is interesting to note in this connection that Schuster
computes from the dimensions of Duffield's apparatus a maximum error
due to pumping of about 3.6X 10~< for vertical motions of the ship of one
meter amplitude.

In a recent article Duffield^ has given a brief description of tests made
with apparatus other than his own during his voyage to Australia in 1914.
This included instruments of the sealed-cistern barometer type constructed
by Prof. Hecker and an aneroid barometer supplied by Sir Horace Dar>
win. No quantitative results are given.

Schuster* has contributed a valuable analysis of the effects of forced
vibrations which may be imposed on the mercury in gravity apparatus by
the vertical acceleration of the ship. This includes a discussion of ( 1 ) a
single constricted barometer tube, (2) the oscillations in a complex inter-
connected system of three tubes as in Duffield's apparatus, and (3) the
experimental errors in the latter apparatus as affected by the relative
dimensions of the various parts. He emphasizes the importance of the
condition that the flow of mercury in the barometer tube and contact tube
be such that the difference in level is always that of hydrostatic equi-
librium. This condition is fulfilled if the cross-sections of the capillaries
are equal and the lengths of the capillaries are inversely as the reduced'
cross-sections of the tubes at the free surfaces of the mercury.

Briggs' method: Briggs* employed apparatus similar to that shown
diagramatically in figure 3 for gravity observations during a voyage from
Sydney to San Francisco in 1914 and again from New York to San Fran-
cisco via Panama in 1915. This apparatus is of the closed-barometer

• The investigation of gravity at sea. Nature, 106, 1921, 732-734.

"Arthur Schuster. On the determination of gravity at sea (Note on Dr. Duffield's
paper). Proc, Roy. Soc. Lond. (A) 112, 1916, 517-528.

• The reduced cross-section a represents the actual cross-section a corrected for
the effect of the pressure and volume of the air above the surface of the mercury.
Let

^=:the original volume of air.
P=thc original air pressure.
A=the height of the mercury column equivalent to P,

Then — -« h it

Qi a V

• L. J. Briggs. A new method of measuring the acceleration of gravity at sea.
Proc. Nat, Acad, Set. 2, 1916, 399-407.

GEODESY 9

type. The mercurial column is contained in the capillary c (bore 0.6-0.7
mm.), the lower end of which opens beneath mercury in the bottom
of the gas chamber d. This capillary is sealed to the wall of the gas-
chamber where it passes through the upper end. The upper part of the
capillary is bent into a flexible zigzag and ends in the spherical bulb h
((tiameter 2 cm.). The bulb contains a fixed iron point p sealed to the
inside of the bulb by means of an inserted platinum wire and extending
vertically downward, so that the point is approximately at the center of
the bulb. The length of the mercurial colunm is about 74 cm.

The flexible capillary permits a slight vertical movement of the observ-
ing bulb with respect to the gas chamber. This movement is determined
by a micrometer screw of 1 mm. pitch which controls the motion of a
carriage in which the observing bulb is rigidly mounted. The carriage
slides on parallel rods mounted on a base which is rigidly cemented to
the neck n of the gas chamber, so that the position of the bulb relative
to the gas chamber is definitely determined by the screw.

The apparatus is protected by a close-fitting metallic jacket, and is kept
at a constant temperature in a bath of melting ice. It is necessary to
determine only the position of the upper end of the barometric column.
The design of the instrument is such that in setting the index in contact
with the- mercury surface the enclosed gas is automatically reduced to a
constant volume ; and since the temperature is constant, all measurements
are made at constant pressure. The relative value of g at two stations
is thus inversely proportional to the observed length of the column at
these stations.

The contact of the index point with the mercury surface can be deter-
mined either electrically or by direct observation. Both methods were
used. In the latter case the fixed point was observed through a glass tube
introduced through the ice, the tube containing a low-power lens. The
point was illuminated through a similar tube on the opposite side. If the
sea was so rough as to cause pumping of the column, the point was so
adjusted that it was in contact half the time as nearly as possible. Since
the motion of the ship is not strictly periodic, there is considerable uncer-
tainty connected with such settings.

This instrument possesses the following features which experience has
shown are desirable in gravity apparatus: (1) The glass part of the appa-
ratus is hermetically sealed and can be made really gas-tight. There are
no stop-cocks, ground joints, or mercury seals. (2) It is necessary to
make settings only at the upper end of the barometric column. (3) This
permits the complete immersion of the apparatus in an ice-bath, which is
the most dependable source of constant temperature for use on shipboard.
(4) The apparatus is portable, since at room temperatures the pressure
is sufiident to fill the observing bulb with mercury.

Gravity determinations which were made in 1914 on board ship in
Wellington harbor and Sydney harbor using observations in San Francisco
harbor as a base station, and in 1915 in San Francisco harbor using New

10 GEODESY

York harbor as a base, show an average departure from pendulum obser-
vations of about one part in 50,000. These observations were made under
favorable conditions and serve to show the degree of accuracy with which
the instruments held their adjustments during the long voyages rather
than to provide any indications of the accuracy of the sea observations.
In fact, the publication of the sea measurements, which show some rather
large anomalies, has been withheld in the hope that apparatus similar to
that described in the last section of this paper might be available to deter-
mine the errors of the instruments under oscillations approximating sea
conditions.

CORRECTIONS FOR THE COURSE AND SPEED OF THE SHIP

Eotvos* has shown the necessity of applying a correction for the east-
erly or westerly motion of the ship, due to the fact that the ship's motion
modifies the angular velocity of the apparatus about the earth's axis. The
centrifugal force acting on the mercurial column when on board a ship
moving east or west is therefore not the same as when the ship is at rest
or moving north or south. The correction may be as great as 1 part in
10,000, but can be accurately computed if the course, speed, and approxi-
mate latitude of the ship are known. Dufiield (1921) observed a change
equivalent to 0.1 millibar when the course was altered from east to west
when steaming at 22 knots, which corresponds to an apparent change of
1 part in 10,000 in g. He does not state where his experiments were
made. For Lat 55**, which represents approximately the mean latitude
of the North Sea, the computed change is about 0.19 cm./sec.*, or 2 parts
in 10,000 in the value of g.

LABORATORY TESTING OF GRAVITY APPARATUS

There seems to be no practical way of determining the accuracy of
gravity measurements directly from sea observations. With a ship at
one's command, repeated traverses could be made of the same sea station
under varying sea and weather conditions ; but while this would provide
a measure of the accidental errors, it would tell us nothing regarding
systematic errors, for the exact value of g at the station would not be
known. The value of ^ in a long, narrow bay could be closely approxi-
mated from pendulum observations on both shores. To satisfy other
requirements the bay would have to be sufficiently open and windswept
to represent the conditions prevailing in a moderate sea. Such a test
seems precluded without the enlistment of government aid.

We can, however, simulate sea conditions at a land station where g is
accurately known and I wish to emphasize the importance of sudi tests
for all gravity apparatus. Suppose we construct a platform capable of
independent reciprocating horizontal translations in two directions at right
angles, corresponding in period to the roll and pitch of the ship. Let us
mount upon this platform a second one arranged for vertical oscillations,

* Sec Hclmert, loc. cit.

GEODESY 11

large enough to carry the apparatus and the observer. All the recipro-
cating motions are to be capable of a continuous change in amplitude if
desired in order to secure the conditions which arise when the period of
the ship differs from that impressed by the waves. We have omitted
the motions corresponding to the angular motions of the deck, for struc-
tural reasons ; for these angular motions would only increase the rotation
of the apparatus in its gimbals, and there will always be sufficient move-
ment of the gimbals arising from the horizontal accelerations to simulate
disturbances due to friction. With this apparatus it will be possible to
investigate the effects of horizontal and vertical accelerations, singly and
combined, under conditions where g is accurately known and thus obtain
a measure of the accidental and systematic errors of the method. Inde-
pendent rolling and pitching oscillations are probably not required for
gravity apparatus, but the equipment proposed has other useful applica-
tions, as for example in testing gyroscopic compasses and gyroscopic
stabilizers, in which both rolling and pitching accelerations should be
included. Plans for such an equipment are now being prepared, and it
is hoped that arrangements can soon be made for its construction and
installation at the Bureau of Standards, where it will be available for
testing all gravity apparatus.

Bureau of Standards,

Washington, D. C.

ISOSTASY
By John F. Hayfobd

I assume that I am expected to bring forward whatever ideas will, in my
opinion, help most at the present time in the development of a study of
isostasy by stimulating further thinking along that line. From among
the many topics which might be treated, I am selecting three :

(1) I shall make some remarks on the recent paper entitled "The
Chemistry of the Earth's Crust," by H. S. Washington.

(2) I shall try to emphasize the desirability of an intensive study of
two small areas by observations of gravity and deflections of the vertical.

(3) I shall offer some considerations that lead me to believe that the
undertow involved in isostatic readjustment is above, rather than below,
the depth of compensation.

CHEMISTRY OF THE EARTH'S CRUST

Dr. Washington, in his paper entitled "The Chemistry of the Earth's
Crust," * has set forth a piece of research of much importance to those
who are studying isostasy. He has set forth the evidence derived from
studies of the densities of igneous rocks. These densities are determined
from chemical analyses. In the latter part of the paper he sets forth the
correlation which is observable between the density of igneous rocks, on

' Journal of the Franklin Institute, 190, December, 1920, 757-815.

12 GEODESY

the one hand, and, on the other hand, the elevation of that part of the
earth's surface under which the rocks lie.

The evidence given in the paper seems to be conclusive in its general
features on three points, namely :

(1) That igneous rocks under the oceans are denser than those under
the continents ;

(2) That igneous rocks under the various continents are less dense
the greater the mean elevation of the continent ; and

(3) That igneous rocks under different parts of any one continent are
less dense the higher is that part of the continent.

All three of these conclusions are in accord with the theory of isostasy
and, in Dr. Washington's words, constitute "almost a conclusive proof of
the general validity of the theory of isostasy."

Dr. Washington seems to interpret the relations pointed out as being
due to original, or early, segregation of the material. I do not question
this interpretation so far as ( 1 ) is concerned. The relative positions of
the oceans and continents are permanent or semi-permanent. It may well
be that the oceans are now in their present position because material of
such a nature as to form dense rocks was placed early in the positions
now occupied by the oceans.

The same interpretation may, however, be questioned, in part, in so far
as (2) and (3) are concerned. Are the differences in densities between
the different continents referred to in (2) and between different parts of
any one continent referred to in (3) due to original or early segregation?
Or are these differences due, in part at least, to some response of the
material to a change of pressure in such a manner as to bring about a
change of density ? The continents probably have not had the same rela-
tive elevations throughout geologic time as they now have. Certainly,
the different parts of any one continent have not, in general, the relative
elevations now that they had at various times during the geologic history
of the continents. For example, the Appalachian region, referred to by
Dr. Washington, in the eastern part of the United States, is now much
lower than Utah, Colorado, and Nevada, and the density of the material,
as measured by him, is considerably greater under the Appalachian region
than under the states named. There was, however, a time during the
geologic history of the North American continent when the Utah-
Colorado-Nevada region was much lower than the Appalachian region was
at that time. Have the relative densities in the two regions changed
between that period and the present?

This comment on Dr. Washington's paper is not intended at all as a
criticism ; it is intended to supplement the paper and to stimulate further
thought based on it. All who are interested in isostasy should be very
keen to follow the further develpoments along the line indicated by Dr.
Washington's paper. I understand that he, himself, will develop the evi-
dence much more fully. The more complete evidence should be examined
very carefully, with a view to determining the bearing of this evidence on

GEODBSY 13

any theories which may be held in regard to isostasy and the isostatic
readjustment.

PROPOSED INTENSIVE STUDY OF SMALL AREAS

We shall take a short time to consider the possible benefits which would
follow from an intensive study of two small areas, let us say about 100
miles square, one on rather flat country, such as Louisiana, and another
in hilly or mountainous country. The intensive study would be based on
closely spaced stations of two kinds, stations at which deflections of the
vertical are determined and gravity stations. Let us suppose that obser-
vations of these two kinds were made in each of two such areas and then
that, by the proper office methods and by the combination of the two kinds
of observations, the distribution of the densities beneath each of the two
areas was determined with as great accuracy as is possible. What benefits
would follow from two such studies? Briefly, I believe that the benefits
would be as indicated in the four paragraphs which follow :

1. The studies would bring out the actual advantages and limitations
of this line of attack on the problem of determining the distribution of
densities beneath the earth's surface. The relative strength or weakness
of this line of attack, as compared with the usual line of attack, is not
now well known. The usual line of attack is to use deflections of the
vertical stations or gravity stations which are widely distributed somewhat
uniformly over a very large area.

2. The studies would probably furnish a considerable amount of evi-
dence on the point which has been ably brought out by Dr. William Bowie
that the small anomalies in gravity and in deflections of the vertical which
remain after correcting for topography and isostatic compensation are
closely related in many regions, if not as a rule, to the surface geology
of those regions. It is important that the extent to which this is true
should be determined. There seems to be no doubt that it is true, in a
general way, for large areas. The question is, to what extent is it true for
small areas?

3. The studies would give valuable indirect evidence as to the extent
to which the present conclusions from the evidence now available are
vitiated by local effects which* are at present unavoidably assumed, for
want of more exact information, to extend half way to the next station,
so to speak.

4. The two intensive studies would help to determine whether oil and
salt may be located by geodetic measurements. This suggestion has been
made at various times by Dr. David White, Chief Geologist of the U. S.
Geological Survey. It is important to know whether gravity stations and
deflections of the vertical may be used as a divining rod for that purpose
with sufficient accuracy. If the accuracy of such a divining rod is so low
that its indications are likely to be misleading, then it is not feasible to
try to use the observations in this way. On the other hand, if two such
intensive studies as are indicated here show that sufficient accuracy is

14 GEODESY

possible, then certainly the method should be used, in combination possibly
with other methods. The two intensive studies might possibly be made
in regions in which salt or oil are believed to exist, or are known to exist,
in large quantities.

The suggestion that one study should be in rather fiat country and the
other in rather hilly country is based on two considerations. In very flat
country the conclusions reached will be vitiated to a much smaller extent
by an error in the assumed surface density of the material than in rough
country. On the other hand, in hilly or mountainous country the depth
at which any material of abnormal density lies may be determined with
greater accuracy than in flat country,

DEPTH OF UNDERTOW

Let us now turn to the question, is the undertow involved in isostatic
readjustment above or below the depth of compensation?

Assume that, at some time in the remote geologic past, the North Amer-
ican continent and its various major parts have been almost completely
compensated in the isostatic sense. Assume that, in later geologic past
time and up to the present time, there have been large amounts of erosion
from large portions of the continent and corresponding large amounts of
deposition in other parts. Assume that in that period and up to the
present, readjustment toward isostatic conditions has been in progress by
horizontal transfer of material from beneath the regions of deposition
towards the regions of erosion.

It is reasonably certain that, on the whole, these three assumptions are
true. The question on which it is desired to concentrate attention is, has
the horizontal transfer taken place below or above the depth of com-
pensation?

It is important to secure the correct conclusion on this point, whatever
it is, because the apparent correlation of geodetic and geologic evidence,
or the apparent contradictions between two lines of evidence, probably
depend somewhat intimately upon the conclusion reached. I believe that
the horizontal transfer has taken place above the depth of compensation,
say within 100 kilometers of the surface, rather than below that depth.
I propose to state, very briefly, some of the lines of thought that have led
me to that conclusion.

If it is assumed that, under an area of deposition, the material down
to the depth of compensation all sinks under the added load and that the
horizontal transfer of material occurs below that depth, the case is similar
to that of an ice floe. Under each elevation on the upper surface, in this
case, there must develop a much larger bump on the lower surface of the
floating mass. The conception is that of a crust floating on a relatively
plastic stratum. The level of compensation, in this case, is at the lower
side of the floating crust. As in this case there must be extensions of
the crust downward, below the mountains, the depth of compensation will
be variable, being great under high areas and small under low areas. The

GEODESY 15

geodetic evidence, as far as I have been able to examine it, does not seem
to be conclusive that there is any such relation between the depth of com-
pensation and the elevation of the surface. This leads me to be skeptical
of such theory, which involves a horizontal flow limited mainly to those
portions of the earth that are below the depth of compensation.

In general, I am skeptical of any explanations of isostatic readjustment,
or of other phenomena in the earth, which involve a relatively plastic
stratum in contrast to more rigid material above. In each of the cases
in which I have been able to follow the mechanics of the problem to itiy
own satisfaction, I have not been convinced that the resort to the device of
introducing a plastic layer into the concept is necessary. So, in the prob-
lem now under consideration, which is that of the isostatic readjustment,
I do not find it necessary to assume any stratum to be more plastic than
the one above, in order to harmonize the observed facts of various kinds.

There is abundant geological evidence of horizontal stresses and strains
in the earth's crust. This evidence seems to me to be conclusive. The
geologic evidence seems to me to indicate a horizontal transfer of material
during isostatic readjustment relatively near the surface, rather than at
great depths. If the horizontal transfer involved in the undertow were
in a plastic stratum more than 100 kilometers below the surface, certainly
the horizontal stresses set up in the surface material would be much less
than if the same transfer occurred in less plastic material nearer the
surface.

In attempting to determine the mechanics of the isostatic readjustment
which apparently takes place when great loads are removed from a region
of erosion and equally great loads are added to other regions as deposited
material, it is extremely important to keep in mind that material deforms
under relatively small differences in the two principal stresses. A very
large increase in both the principal stresses is necessary in order to pro-
duce deformation if the increases are equal. In an elementary cube of
the material, let p^ be the pressure on the upper and lower faces of the
cube. Let p^ be the pressure on theside faces of the cube. If pt=pi,
the material is under isostatic conditions. Under these conditions both
^3 and Pi may be increased very largely before appreciable compression is
produced in such material as constitutes the earth's crust. On the other
hand, if p^ is increased without changing p^, then deformation of the cube
will be produced for a relatively small difference ^j — Pu corresponding
to an added load of a few thousand feet only of materisil. The cube will
be deformed in the sense in which the vertical dimension is decreased
and the horizontal dimensions increased. Consider the movements which
will take place if the many elementary cubes under a large loaded area are
so distorted. Evidently the motions of the material under the margins
of such a load will have a horizontal component. This line of thought,
followed through to its logical conclusions, and made more definite by
careful analysis, will indicate that the horizontal transfer of material
occurs largely at moderate depths, certainly at less than 100 kilometers
as a rule.

16 GEODESY

I am perfectly aware^that the presentation of the considerations set
forth in the last few paragraphs has been too brief for clearness or con-
clusiveness. The paragraphs have been written as suggestions rather
than as demonstrations. They indicate lines of thought which should be
followed up carefully if one wishes to reach true conclusions.

I desire to reiterate my opinion/ based on such thought as I have been
able to give to the subject, that the undertow involved in isostatic re-
adjustment is above the depth of compensation.

THE EARTH-TIDE EXPERIMENT
By Henry G. Gale

I understood Professor Hayford to say that he did not believe that a
fluid layer exists beneath the solid crust of the earth. The same conclu-
sion may be drawn from the earth tide experiment which was conducted
on the grounds of Yerkes Observatory by Professor Michelson and my-
self. It seems pretty certain that the earth tides are the same as they
would be if the earth were a highly elastic homogeneous solid, both with
respect to the phase and amplitude of the earth tides.

The experiment was entirely successful from a physicist's standpoint.
The interferometers gave very little trouble. One of them did not require
readjustment during the entire year. Two were readjusted to change the
width of the fringes, and on one interferometer it was necessary to re-
silver one of the mirrors. As a source of light we used commercial alter-
nating Cooper-Hewitt lamps, and they proved to be entirely satisfactory.
The only serious interruptions were caused by breaks in the electric service
due to storms, and occasional short shut-downs by the electric light com-
pany which supplied the current. The experiment at Yerkes Observatory
was continued for just one year. This is probably long enough to give
values of the semi-diurnal and diurnal tides, accurate to a few tenths of
one percent. For tides of longer period the experiment should be con-
tinued for, say, five years, although possibly three years would be long
enough.

The experiment is now being repeated on the grounds of the California
Institute of Technology at Pasadena, California. One additional station
should be installed, preferably on an island in the Pacific, far from the
continental borders. If the three stations should give results in agreement
on both the phase and amplitude of the earth tides, I should feel that the
problem had been solved. If the three were not in agreement, at least
one-half dozen additional stations would be desirable. They should be
well scattered in latitude, and with reference to tidal coasts.

The cost of installing a station is not excessive, and one man can easily
keep a station in operation and reduce the observations. It would proba-
bly be worth while to look for a correlation between the slight changes of

^ As early as 1911 I had reached the above conclusion, as indicated by one of the
diagrams in an article published in Science, 23, No. 841, 199-206, Feb. 10, 1911.
entitled "The relations of isostasy to geodesy, geophysics and geology."

GEODESY 17

level often shown by the apparatus and the approach or passage of the
intense barometric lows and highs of large area. It would certainly be
worth while to install a specially designed apparatus, similar to that used
for the earth tides, to detect and measure the rate of tilting in the surface
layers of the earth's strata at especially favorable places. A relation be-
tween such rates of tiltii^ and earthquakes might be detected.

THE EOTVOS BALANCE
By W. ST. Lambkkt

There is one instrument of use in the study of terrestrial gravity which
has not yet been employed in the western hemisphere, though it has to a
limited extent in Europe. This instrument is the Eotvos balance. It
should be tried, I believe, in making the proposed minute investigation of
gravity in a level region of limited extent.* To judge by accounts of work
done with the balance in Europe it would certainly supplement the pendu-
lum advantageously in the proposed gravity survey and might largely
supersede it. This does not by any means signify, however, that the
balance is always and everywhere a substitute for the pendulum. This
is not the occasion for an exposition of the principles' of the instrument,
but it may be said that the Eotvos balance connects the results at adjacent
points in a limited region with one another in a way that the pendulum
cannot well do.

In Europe they evidently believe in its possibilities as an indicator of
the existence and location of concealed irregularities and discontinuities in
density. Some of these discontinuities may mean strata of geological or
commercial interest. At least three attempts have been or are being made
to locate such strata by the use of the balance ( 1 ) by Dr. Schumann to
locate lignite deposits in Austria,' (2) by Professor Schweydar for geo-
logical purposes in the region about Hamburg, Germany,' and (3) an
attempt to locate salt deposits in Poland.^ It will be of interest to know
how successful these attempts prove to be.

The Eotvos balance determines certain second derivatives of the gravity
potential function. It does not, however, determine the second derivative
in the vertical direction. This quantity has to be determined theoretiqilly.
This is the most serious deficiency of the instrument and explains why it
is of comparatively little use in rough country. Attempts are being made
by Berroth, Helmert's assistant, in his last researches on gravity, to devise
a means of determining experimentally the second derivative in the verti-
cal direction.'

' See p. 13 of this bulletin.

'Akademie der Wissenschaften m IVien: math. phys. Klasse. Sttzung vom 8
Jaimer, 1920. Reported in the Akademische Anseiger. Nr. 1.

' '*Rapport sur les Travaux du Bureau Central de 1 Association Geodesique Inter-
nationale en 1920,** p. 4.

* 1 have seen nothing in print about this third attempt I have word of it person-
ally from the man proposing to make the observations. [Added in proof: This
attempt has not yet been made on account of the lack of instrument]

' Same reference as in second preceding footnote.

18 GEODESY

THE PROBLEM OF THE EARTH TIDES

By W. D. Lambekt

There are two methods of attacking the problem of the elastic proper-
ties of the earth, ( 1 ) the study of the seismological evidence, on which I
shall not touch at all, and (2) the study of earth tides. Even after the
best available observational evidence from the earth tides has been ob-
tained, a good deal of h3rpothesis and interpretation is required before we
can say : "The elastic constants of the earth are thus and so." It is not,
however, of this that I wish to speak, but rather of the problem of obtain-
ing the true values of the earth tides. This subject is connected with
several other geophysical questions and this paper, therefore, falls natu-
rally into three divisions: (1) Earth tides and the long-period oceanic
tides; (2) earth tides and the short-period oceanic tides; (3) earth tides
and the variation of latitude.

EARTH TIDES AND THE LONG-PERIOD OCEANIC TIDES

The first quantitative evaluation of the earth tides and hence of the
elastic properties of the earth came from a discussion of the so-called
long-period oceanic tides. The suggestion which initiated the work ap-
pears to have come from Lord Kelvin and the method and results are
given in Thomson and Tait's "Natural Philosophy,"* but the actual dis-
cussion of the observations was made by Darwin.^

The discussion is based on the assumption that the long-period oceanic
tides for an ocean on a perfectly rigid globe may be calculated on the
equilibrium theory. This means that it is assumed that the disturbance
of equilibrium caused by the tide-producing forces of long period can
travel so rapidly through the water and the forces themselves change so
slowly that the ocean has time to adjust itself to the forces and that at
any given instant the surface of the ocean forms an equipotential surface
for the instantaneous field of force.* The observed oceanic tide would
be the difference between the oceanic tide for a rigid body and the earth
tide, and when the first two are known the earth tide may be inferred.
From 33 years of observation on both the monthly and the fortnightly
tides at 14 different ports Darwin deduced that the observed tides were
about two-thirds as large as they would be on a perfectly rigid globe. The
earth tide corresponding to the other third implies an effective rigidity of
the earth about equal to that of steel.* Later methods of attacking the

'Second Edition (1883), | 848.

*G. H. Darwin: Scientific Papers," I, 340. This contains a reproduction of the
passage in Thomson and Tait with some changes of notation and unimportant
omissions.

'This assumption is very evidently incorrect as re^rds the diurnal and semi-
diurnal tides. It was supposed to be at least approximately true for the monthb^
and fortnightly tides.

*The tides at Indian ports, which are more consistent with one another than the
others, gave a much higher rigidity.

GEODESY 19

problem have given results somewhat similar, though in general tending
towards somewhat higher values of the rigidity.

Only a few years passed before this result was questioned because of
its dependence on the equilibrium theory, and Darwin himself, apparently,
was the first to question it.^ In treating the problem of tides on a globe
covered with water — ^a problem first formulated by Laplace — ^he discov-
ered solutions which for depths anything like these of our actual oceans
gave tidal oscillations of perhaps half the amount deduced from the equi-
librium theory. He concludes his article by saying: '*Thus it does not
seem likely that it will ever be possible to evaluate the effective rigidity
of the earth by means of tidal observation.'^ The mathematical treatment
of the problem has since been developed by Hough* and Goldsborough."
The latter has extended the treatment to include basins bounded by two
parallels of latitude, a polar basin or one covering the entire globe being
special cases. The general result is that for oceanic depths such as we
know the monthly and fortnightly tides in such basins differ considerably
from what the equilibrium theory gives, being in general considerably
smaller.

These results are decidedly puzzling when compared with observation,
for the observed tides would then be larger than the computed tides of
Darwin, Hough and Goldsborough instead of being smaller by an amount
representing the yielding of the earth to the tide-producing forces. Dar-
win's solution was re-examined critically by the late Lord Rayleigh.* He
reaches the conclusion that the solution is a very special one that applies,
of course, to the ideal water-covered globe postulated, but which has little
relation to our actual oceans, interrupted as they are by continental bar-
riers. He says : "H this conclusion be admitted, the theoretical fortnightly
tide will not differ materially from its equilibrium value, and Darwin's
former calculation as to the earth's rigidity will regain its significance."
After a word of caution about possible exceptional conditions he con-
cludes : "In any case I think that observations and reductions of the fort-
nightly tide should be pursued. Observation is competent to determine
not merely the general magnitude of the tide but the law as dependent
upon latitude and longitude. Should the observed law conform to that
of the equilibrium theory, it would go a long way to verify d posteriori
the applicability of this theory to the circumstances of the case."

Rayleigh's belief in the legitimacy of calculating the fortnightly tide —
and A fortiori the monthly tide — from the equilibrium is supported by the
opinion of Love' and there is still another reason for accepting this idea,

^Proceedings of the Royal Society of London, 40 (1886), 337, or "Scientific
Papers," I, 366.

* Philosophical Transactions of the Royal Society of Lofidon, 189, 1897, 201, and
191, 1898, 139. Some account of Hough's work is given in Darwin's article on
''Tides" in the eleventh edition of the Encyclopaedia Britannica, parts of which are
given also id Darwin's "Scientific Papers," I, 347.

* Proceedings of the London Mathematical Society, Vol. for 1914-15, 31 and 207.

* London, Edinburgh and Dublin Philosophical Magazine, 5, 1903, 136.
'"Some Problems in Geodynamics" (Cambridge, England, 1911), 51.

20 GEODESY

namely, the eflFect of friction. Hough^ evaluates the modulus of decay
under friction for various types of oscillation. The modulus of decay is a
time of the same order of magnitude as the period of an oscillation that
is slow enough to conform approximately to the equilibrium law. For
oscillations most nearly corresponding to the long-period tides and with
the laboratory value of the coefficient of viscosity of water he found
moduli of the order of magnitude of ten years. But this use of the
laboratory value of the viscosity seems to be fallacious. The laboratory
value applies to the so-called laminar motion, while the motion in the
actual ocean is turbulent. When we call the motion turbulent, we say in
effect that we do not understand it very well ; but it is known that if we
attempt to represent turbulent fluid motion by equations of the same form
as are used for laminar motion, then the coefficient of viscosity in the
latter, which is the laboratory coefficient, must be replaced by a coefficient
of virtual viscosity many times greater. Ekman found for water that the
coefficient of virtual viscosity was 15,500 times' greater ; for air, Taylor
fotmd the ratio of the virtual viscosity to the laboratory value to be
between 6,000 and 50,000.' It is not to be supposed tliat any one ratio of
virtual coefficient to laboratory coefficient would apply under all condi-
tions, but if any such virtual coefficient as is here suggested were used
instead of the laboratory value, Hough's modulus of decay would be
greatly reduced and the equilibrium theory would apply to tides of com-
paratively short period, even apart from the effect of continental barriers.

It seems to me, then, that it is pretty safe to assume that the monthly
and fortnightly tides conform quite closely to the equilibrium theory and
that it would be well worth while to resume the study of these tides for
the light that they may throw on the rigidity of the earth. The only
discussion of the observations since Darwin's original one is due to
Schweydar,^ in which are discussed observations at 43 ports covering 194
years. There must be available among all the harmonic analyses of tides
that have been made since that time a great deal of material still unutilized.

In regard to the utilization of this material and the procurement of new
material there are three suggestions I should like to make.

First: In securing the observations care should be taken that the tide
gauge has a firm foundation. Tide gauges. are frequently located on docks
supported by piles, or on made ground. If the cost could be afforded,
it would be desirable to place them, if necessary, at some little distance
from the water in order to get a good foundation, and to connect the
gauge with the water by a large pipe. Incidentally, if a firm foundation

^Proceedings of the London Mathefnaiical Society, 28, 1896, 264.

•These values arc taken from a paper by McEwcii: Ocean Temperatures, their
relation to Solar Radiation and Oceanic Circulation ; miscellaneous Studies of Agri-
culture, Biology, Semicentennial Publications of the University of California. Other
examples of very wide diflferences between laboratory and field values are also given
by McEwen. The work of Ekman and Taylor is found respectively in the Arkiv
for Matematik, Astronomi och Fysik 2, 190S, 1. and in the Philosophical Transac-
tions of the Royal Society of London, 215, Ser. A, 1915, 1.

• Beitrage zur Geophysik, 9, 1908, 64.

GEODESY 21

were assured, the tide records would be valuable in studying the secular
rising and sinking of the coast. Too many long series of tidal observa-
tions that would have been valuable for this purpose in the past have been
rendered useless for lack of connection with well-established permanent
marks. Schweydar deduced from his discussion of the tides an eflFective
rigidity of the earth rather less than that of steel and suggests that the
discrepancy between this result and the higher rigidity obtained from the
variation of latitude may be due to a plastic stratum beneath the crust.
He seems to have abandoned later his belief in this plastic stratum. Per-
haps such a stratum exists, after all, but it may consist simply of the
alluvial ground, mud or made land on which some of the tide gauges are
situated.

Second : In reducing a year's tidal observations at a place to obtain the
long-period tides it is found that the tides are much entangled with one
another. To separate them Darwin gives a rather tedious process of
successive approximations. It is possible to dispense, with these repeti-
tions at the price of a rather heavy piece of preliminary computation that
is done once for all series of a given length, such as a year. If many
long-period tides are to be reduced, it would be well worth while to do
this preliminary work. Furthermore, Harris has suggested that if four
consecutive years be taken together, the several long-period tides will
separate satisfactorily from one another and the computation will be
simpler.

Third : In the final discussion we seek to obtain the earth tides from
the difference between the theoretical tides on a rigid globe and the ob-
served tides. Now, even accepting the comparatively simple equilibrium
theory, which we are at present supposing to be adequate, the tides on
a rigid globe have not been evaluated as accurately as could be vrished.
The difficulties are the attraction of the water on itself and the presence
of the continents. Either difficulty by itself is readily overcome. If there
were no continents and the globe were covered with water, a simple factor
derived from the fundamental principles of spherical harmonics would
take care of the self -attraction of the water. On the other hand, if we
disr^ard the self-attraction of the water, the influence of the continents
can be allowed for as Darwin and Turner^ have done in following Sir
William Thomson's (Lord Kelvin's) suggestion.

It is the combination of the two circumstances, neither of which is
troublesome by itself, that makes trouble, for the resultant of the two
cannot be had by simply superposing two corrections. The total effect
cannot be very large, but the modulus of rigidity is rather sensitive to
changes in the ratio of the observed tide to the theoretical tide, especially
if this ratio be near to unity, so that it seems perhaps quite possible that
the modulus of rigidity may be changed as much as thirty or forty percent
by the application of the corrections above mentioned.

• Proceedings of the Royal Society of London, 40, 1886, 303, or Darwin's "Scien-
tific Papers/' I, 328.

22 GEODESY

The evaluation of this correction of the combined eflFect of the conti-
nents and the self-attraction of the water seems to me a problem worthy
of study. No advance seems to have been made since Poincare's paper/
in which he works out a solution very elegant in conception, but one
which, as Poincare himself says, would lead to calculations far too com-
plicated to be practicable even if the shore line of the continents were
arbitrarily simplified into a rude approximation to its actual form. I
believe, however, that it may be possible to solve the problem numerically
by a laborious process of successive approximations, involving the prepa-
ration of maps showing the equilibrium tide corrected for the continents
alone, and the reading of these maps for many points on the earth, much
as contour maps are read to obtain the deflections of the vertical. The
labor would be more than an individual investigator would care to under-
take, but might well be within the means of an institution. Perhaps the
theory of integral equations in its recent developments might afford means
of lightening the labor.

EARTH TIDES AND THE SHORT-PERIOD OCEANIC TIDES

In discussing observations of earth tides as a means for obtaining the
elastic constants of the earth — observations made either with horizontal
pendulums or with Michelson's tube and interferometer — it is important
to have a knowledge of the oceanic tides. This statement applies to tides
of all periods, but these remarks apply more particularly to tides of short
period, i. e., to the diurnal and semidiurnal tides. Tides of longer period
have just been discussed.

The shifting mass of tidal water exerts a direct gravitational effect on
the horizontal pendulum or on the liquid in the tube, and the direct
effect is reinforced by the tilting of the earth's crust under the shifting
load of tidal water. The periods of these effects are precisely the periods
of the tide-producing forces ; hence it is impossible to make an adequate
estimate of the yielding to the tidal forces of the solid earth as a whole
until the direct and indirect effects of the oceanic tides have been allowed
for.

Probably the most satisfactory determination of the earth tides is that
of Michelson and Gale,* and it is noteworthy that their observations were
made at Williams Bay, Wisconsin, which is some 800 miles distance from
the ocean. The yielding of the earth deduced from the north-and-south
displacements is nearly the same as that deduced from the east-and-west
ones and the yielding deduced from the diurnal declinational tide, O^, is
nearly the same as deduced from the principal semidiurnal tide, M,.

This satisfactory state of affairs no longer obtains with some of the
observations taken elsewhere with the horizontal pendulum. Keeker's

^Journal de Mathimatiques Pures et Appliquies, 2, 1896, 57.

• Journal of Geology, 22, 1914, or the identical article in the Astrophysical Journal
for March, 1914. An important correction is given in Science, 50, 1919, Z27, Defi-
nitive results will be found in the Astrophysical Journal for December, 1919.

GEODESY 23

results at Potsdam,^ which seemed to show a greater yielding of the earth
in the meridian than in the prime vertical, have been a standing puzzle;
attempts to explain this peculiarity as due to the rotation of the earth
have been unsatisfactory.'

More recent observations by Schweydar with horizontal pendulumB at
Freiberg in Saxony* show anomalies also. Perhaps these anomalies are
due to the inferior accuracy of the horizontal pendulum as compared with
Michelson and Gale's apparatus, but it seems to me probable that the un-
eliminated effects of the oceanic tides may also play a part.^ This effect
has never been calculated, so far as I know, and has been assumed to be
so small as to be practically negligible, but, as we shall shortly see, there
are reasons for questioning this assumption.

The only serious attempt to allow for the oceanic tides, so far as I
know, is due to Prof. Shida" of Kyoto University, Japan. He observed
with horizontal pendulums for a year at Kamigamo Geophysical Observa-
tory, near Kyoto. In Japan it is, of course, impossible to get far from
tide water, and though the tides in the surrounding waters are not par-
ticularly large, but rather the contrary, still considerable effect is to be
looked for. Harris's cotidal maps* were used for the M, component.
They are based on actual observation for the coast, but necessarily on
theory and inference for the open sea. For the Oi component special
maps were drawn from the somewhat meager data available. From the
maps the gravitational effects of the tidal water was read off just as the
topographic deflection of the vertical may be read. The effect due to the
yielding of the earth's crust was also computed on the most plausiSle
assumptions practicable. The actual computation is not unlike that of
the deflection of the vertical ; the result comes out that the deflection due
to the tilting under the tidal load is about twice the direct gravitational pull
of the load itself, and the two corrections combined were of the same
order of magnitude as the earth tides proper. For example, the observed
Oi tide on the pendulum swinging northwest-southeast came out

+0:00525 cos ^-hOlWOSg sin t

while the total deflection in the same direction due to the ocean tides was
computed as

+0:00410 cos *— 0.:00250 sin /

^ "Beobachtungen an Horizontalpendeln tiber die Deformation des Erdkorpers
tinter dem Einflusz von Sonne tind Mond/' Veroffentlicfaungen des Konigl. Preuss-
ztschen Geodatischen Institutes n. f. no. 32, Berlin, 1907, and Heft 2, n. f. no. 49,
Berlin, 1911.

'See Love: "Problems in Geodynamics" (Cambridge, England, 1911), 75.

' "Bericht uber die Tatigkeit des Zentralbureaus der Intemationalen Erdmessung
im Jahre 1920," 6.

^OrlofTs observations at Dorpat (reported in Astronotnische Nachrichten, 1S6,
1910, 81) show a peculiarity similar to Hecker's but to a less degree. Dorpat is
farUier removed than Potsdam from the influence of the large tides in the Atlantic

' "Memoirs of the College of Science and Engineering," Kyoto Imperial University,
IV, no. 1 (Nov., 1912).

• Manual of Tides. Part IV B (Report of the U. S. Coast and Geodetic Survey
for 1904, appendix 5).

24 GEODESY

where / is the hour angle of the fictitious O^ tide-producing body of the
harmonic analysis. Evidently any inference drawn from the uncorrected
earth tides would be quite wide of the mark. The computation was ex-
tended to a distance of 40° of great circle (nearly 2,800 statute miles)
from the station. This was considered sufficient in view of the meager-
ness of the tidal data and the approximate nature of the work, but it
should be remarked that the zone between SO"* and 40° gave in some cases
a result equal to more than one-tenth of the whole correction, suggesting
the desirability of extending the calculation to even greater distances.

When we consider the great areas of ocean that lie within 40° of any
one of the European horizontal-pendulum stations, it seems rash to assume
without careful calculation that we may neglect the effects of the oceanic
tides on the observed earth tides, and it may even appear desirable to see
whether Michelson and Gale's result may not be susceptible of improve-
ment by appl3ring the correction for oceanic tides.

The primary difficulty with calculations of this sort is our lack of knowl-
edge of tides in the open sea. We have Harris's cotidal lines, and it may
be said that when these were used for reducing the Kyoto observations
the results obtained seem, quite satisfactory, thus verifying to a certain
extent the theories on which the lines were based. But at best this is
theory rather than observation, and Harris himself was as keenly aware
as anyone else of the incompleteness of' his work and the necessity of
verifying it by observation.

The direct observation of tides at sea is a problem beset with difficulties.
To observe tides by means of soundings repeated every hour or so at the
same point seems impracticable on account of the great depth to be
sounded, rendering an accuracy of a foot or less impracticable, and on
account of the difficulty of recovering the same point. Pressure gauges
in one form or another have been suggested, but the instrument that will
sustain the load of a thousand fathoms of water and at the same time be
sensitive to variations in that load of a foot or so has not yet been
devised.*

It has occurred to me, however, that the question of tidal oscillations
at sea could be approached somewhat differently, namely, by a study of
the horizontal oscillations, that is, the tidal currents. In the open sea
these tidal currents would, of course, be small, but not always too small
to be detected and studied. Given a good knowledge of the tidal currents,
the tidal rise and fall could be inferred with fair certainty. The relatively
large tidal currents are to be looked for near the nodal lines of the sta-
tionary tidal oscillations, and Harris's theory will indicate plausible places
in which to look for such nodal lines.

What has chiefly impressed me with the possibility of measuring tidal
currents at sea was the reduction that I made for the late Dr. Harris of

^ A recording tide gauge for work at sea invented by M. Fave, a French hydro-
graphic engineer, that is said to have given good results at Dover, England, and in
ftie Thames Estuary, is mentioned in the Observatory, 43, Aug., 1920, 2/9.

GEODESY 25

observations taken some thirty-five years ago by Lieut. Pillsbury/ as he
was then, later Rear Admiral Pillsbury. They were not made with the
study of tidal currents chiefly in view, but for the exploration of the
Gulf Stream. Dr. Harris had them worked over again by more modem
methods to see what information about tidal currents could be extracted.^

The series were all short, a few days at the most, and some of them
did not put in evidence an unquestionable tidal current, but a number of
them did. A plotting of these latter showed that the results of the approx-
imate harmonic analyses that were made could not be far from the truth.
The velocities found ranged in general from 0.05 to 0.3 knot. It may be
of interest to remark that the times and directions of the current were
in general agreement with Harris's theory of stationary tidal oscillations.
The observations, of course, were made before this theory was formu-
lated, but they were not reduced for tidal purposes till some time after
the theory was published, so that they serve as a partial confirmation of it.

H an expedition were sent out to determine tidal currents at sea in
somewhat the way here suggested, it would have to occupy one spot for
several days, or preferably longer. While the vessel remained on the
spot for current observations there would be an excellent opportunity for
other kinds of scientific observation, magnetic, geophysical and biological.
The intensity of gravity at sea is a great desideratum in geophysics and
as soon as adequate apparatus is devised for the purpose observations of
gravity should certainly be made in connection with observations of the
currents.

EARTH TIDES AND THE VARIATION OF LATITUDE

Just as we may make a harmonic analysis of the readings of a hori-
zontal pendulum or a Michelson tube in order to evaluate the earth
tides, so for the same purpose we may make a harmonic analysis of
latitude observations. In all three cases we are observing the direction
of the vertical or plumb line. In the first two cases the vertical is referred
to some mean position determined on the instrument itself and connected
with the ground immediately around it, and thus shifting its position as the
ground tilts under the influence of the tide-producing forces in the earth
and under the load of the oceanic tides. In observations of the latitude
the direction of the vertical is referred not to the ground round about the
instrument, but to the direction of the earth's axis. The tilting of the
ground is allowed for when the level readings are taken and the proper
corrections for them applied. This absence of the tilting enables us to
get a hold on the problem of the earth tides somewhat different from
that afforded by observation with the horizontal pendulum or with the
tube and interferometer.

* Appendices to Reports of the U. S. Coast and Geodetic Survey for 1885, 1886,
1887, 1889 and 1890. More especially that for 1890, which contains an account of
the apparatus used.

' For the tidal data deduced see Harris's Manual of Tides, Part V (U. S. Coast
and Geodetic Survey, Report for 1907, Appendix 6), 409-13.

26 GEODESY

The latitude observations most obviously appropriate for such a har-
monic analysis are those of the international latitude service and I believe
that they should be systematically discussed in this manner. Some pre-
liminary work of this sort has been done by Shida and Matsuyama/ and
Shida has proposed to the International Geodetic Association that it
undertake the work. Further work on the subject has been done by
Przbyllok, whose work has perhaps already been published, though with-
out coming generally to the attention of scientists on this side of the
water' on account of still unsettled international conditions. I believe
the International Geodetic and Geophysical Union should plan to continue
the work.

SUMMARY

The observation and reduction of the long-period oceanic tides should
not be neglected. The equilibrium theory has not been as fully developed
as is desirable and an attempt should be made to allow both for the self-
attraction of the water and for the presence of the continents. When this
has been done and observation compared with theory, it seems probable
that a good value for the long-period earth tides will result. To get good
results for the short-period earth tides the oceanic tides of like period
must be known and their effects allowed for. One promising means of
getting this knowledge seems to be a study of the oceanic tidal currents,
which appears to be more feasible than the direct observations of the tides
themselves. While the currents were being observed other scientific obser-
vations could be made. The earth tides affect the plumb line and their
effects must therefore be present in the observations of the International
Latitude Service. It is desirable to continue the discussion of these obser-
vations in order to throw light on the earth tides.

* "Memoirs of the College of Science and Engineering/' Kyoto Imperial University,
IV, no. 1, 1912. 277.

' "M. le. Prof. Przybyllok a tach^ de d^terminpr plus exactement les constantes de
qaetques termes p^riodiques dont les observations du service international des lati-
tudes avaient fait connaitre Texistence; il s'est surtout occupe a deduire des obser-
vations astronomiques les constantes de la maree M| dans le moovement de la
verticale de la terre consider^e comme corps 61astique. Les resultats seront public
sous peu dans les Astron. Nachrichten.** — From Rapport sur les Travaux du Bureau
Central de I'Association G^d^ique Internationale en 1920, p. 3 (dated Jan., 1921).

SOLAR RADIATION AND TERRESTRIAL PHENOMENA

By C G. Abbot

SOLAR RADIATION
Its Vabiabiuty and Its Relations to the Atmosphere

For more than fifteen years the Astrophysical Observatory of the
Smithsonian Institution has been engaged in making measurements of
the radiation of the sun. These measurements have indicated that the
sun's emission is variable. The Institution now maintains two stations —
one in Arizona and the other in Chile — for observing the solar variability.
Telegraphic reports of the results obtained in Chile have been forwarded
to Buenos Aires and Rio de Janeiro for the use of the meteorological serv-
ices of Argentina and Brazil.

The variation of the solar emission is of two kinds — one, of long
period, associated with variations in the visible solar phenomena like sun-
spots, faculae, prominences, and the like; the other, of short irregular
period, apparently depending upon inequalities of radiation in different
directions, which, rotating with the sun, produce at the earth the variation
just mentioned. This hypothesis is confirmed by the photo-electric cell
observations of Guthnick, who found variations of Saturn occurring
earlier or later than corresponding ones in solar radiation observed in
Chile, depending on the heliocentric longitudes of the earth and Saturn.
Higher values of the solar emission occur when the sunspots are most
numerous, which gives rise to a paradox, because the temperature of
most meteorological stations is lower at sunspot maximum. This negative
correlation between solar emission and terrestrial temperature may be due,
however, to a variation in terrestrial cloudiness or to other variations in
the composition of the terrestrial atmosphere. It has long been known
that a close correlation exists between the sunspot numbers and the varia-
tions of terrestrial magnetism. The intermediary mechanism producing
this relation is not known. However, it has long been suspected to be
due to the bombardment of the earth by ions shot out from the sun. If
this is the case, these ions may assist in the production of cloudiness and
have also influence in the production of ozone in the higher atmosphere,
and thus in one or both of these ways operate on terrestrial temperatures.

Observed variations of the sun have hitherto lain within the maximum
range of about 12 percent. It is rare that fluctuations exceeding 3 percent
occur within a single week or fortnight. Such studies as have been made,
notably those of Mr. Clayton of Buenos Aires, indicate that corresponding
to these small fluctuations of the sun there may be variations of several
degrees in the mean temperature of meteorological stations. Accordingly
it is highly desirable to be able to detect with certainty fluctuations of the
solar emission of the order of 1 percent. This is a hard requirement,
and it is only within the last year that the establishment of the Arizona

27

28 METEOROLOGY

•

and Chile stations has warranted the hope that it can be met. Prior to
that time, errors of 2 or more percent were probably not infrequent in
solar radiation determinations at Mt. Wilson, which imtil 1918 was the
only station in the world where the solar constant observations were being
made. Hence we must be required to wait for another decade of years
before having a thoroughly satisfactory series of solar radiation measure-
ments to compare with meteorolc^cal observations. It would be a very
great advantage if two additional solar stations could be equipped in
Northern and Southern Africa in the most cloudless and favorable con-
ditions, so that there would be four stations operating under a homogene-
ous scheme for determining the variability of the sun.

The question arises whether observations of the visible phenomena upon
the sun's surface, such as sunspots, faculx, prominences, or the like, or
the observation of terrestrial magnetism, which, as has been said, is closely
related to solar phenomena, may furnish some index to solar conditions
as valuable as the difficultly obtained determinations of solar radiation.
Many statistical comparisons have been published on relations of sunspots
and terrestrial phenomena, and to a less degree the other solar appearances
have also been correlated thereto. It must be confessed, however, that
the result of this enormous amount of work has not been as favorable as
would have been hoped. In almost all instances, relations which appeared
to hold for a few months or years are reversed in other months or years.
Except for the well-known correlation of terrestrial magnetism with the
sunspot numbers, there is hardly any other pair of phenomena which
would be universally accepted as related. Whether a similar disappoint-
ment will attend the proposed studies of solar radiation can not yet be
foretold.

Of variable terrestrial influences, the most profoundly active on the
solar radiation are the water vapor of the earth's atmosphere and the
clouds and haze which are formed from it. Water vapor itself produces
powerful absorption bands in the red and infra-red spectrum. Associated
with dust, water vapor produces haze which is effective throughout the
visible spectrum, and more effective the shorter the wave-length. A great
mass of observations of these things has been made by the Smithsonian
Institution in connection with its studies of the solar constant of radiation.
My colleague, Mr. Fowle, has published a number of papers covering the
results of these studies of the subject.

Water vapor itself removes from the direct solar beam which encircles
the earth somewhere from 10 to 20 percent of its intensity, depending
upon the humidity of the air and other circumstances. The haze may
readily produce far greater reduction to the intensity of the solar beam.
In addition, we have the clouds. My colleague, Mr. Aldrich, took advan-
tage of the presence of the balloon school near Mt. Wilson to observe
from a balloon the reflecting power of the upper surface of the vast oceans
of fog which come in from the Pacific and cover the San Gabriel Valley.
He found that a continuous sheet of level cloud would reflect away ap-

METEOROLOGY 29

proximately 77 percent of the solar rays. As the earth is on the average
about 50 percent cloudy^ the great effect of these factors on terrestrial

temperatures is obvious.

Cloud measurements are among the most unsatisfactory which are re-
corded by meteorologists. They depend largely on the personal equation
and indeed no really adequate statistical study of them has hitherto been
available. The preparation of proper automatic observing apparatus and
the study of observations of clouds are highly desirable.

TERRESTRIAL RADIATION
Its Relations to the Atmospheie

When we take up the question of the terrestrial radiation, we deal with
another Tegion of wave-lengths from that which is covered by the prin-
cipal incoming solar rays. The direct rays of the sun and the skylight
are almost altogether confined to the region of wave-lengths extending
from 0.3 micron to 3 microns. The region of the terrestrial radiation
extends from 5 microns to 50 microns. Spectrum measurements have
been made through a part of this region by my colleague, Mr. Fowle,
who used an artificial source of light and a very long column of air of
known humidity and carbon dioxide content. In this way he determined
the influence of terrestrial humidity upon the rays as far as 17 microns.
Beyond that, from 17 to 50 microns, no adequate studies have been made,
and indeed the difficulty of making them is immense. Apparently the
.water vapor existing in a column of air a quarter of a mile long cuts off
all terrestrial rays except in the region from 8 microns to 13 microns.
In this region, water vapor is almost perfectly transmissible and in this
region occurs, therefore, almost all of the terrestrial radiation which,
rising from the earth's surface, escapes to space and tends to cool the
earth. No constituents of' the air at the earth's surface seem to affect
the transmissibility of rays between 9 and 12 microns in wave-length, but
the matter is quite different in the upper atmosphere. Mr. Fowle found
that a strong band of absorption occurs in the direct solar beam squarely
in the middle of this^very transmissible region. It appears from some
measurements of K. Angstrom that the cause of this band is ozone. Thus,
owing to the accidental position of this powerful absorption band in the
middle of the only region where the other atmospheric constituents are
almost perfectly transmissible, ozone plays an important part in deter-
mining terrestrial temperatures.

A research ought to be undertaken to determine the influence of ozone
in this region of the terrestrial spectrum, the variations of its amount in
the atmosphere, and, in short, the dependence of the terrestrial tempera-
tures on ozone. This research will be very difficult, owing to the long
wave-lengths of the rays involved and owing to the occurrence of ozone
high up in the terrestrial atmosphere. The investigation would involve
the determination of the dependence of the ozone content of the atmos-
phere on solar radiation as well as on influencing terrestrial conditions.

30 METEOROLOGY

Hitherto the measurement of the outgoing terrestrial rays — that is, of
the so-called nocturnal radiation — ^has been very unsatisfactory on account
of the lack of a surface which radiates these rays perfectly. Blackened
flat surfaces have been used in the instruments employed, but the black-
ening by means of smoke, lampblack paint, or platinum black are all
unsatisfactory because these substances are not full radiators and absorb-
ers for the very long wave-lengths involved. Smoke, for instance, is
strongly transmissible beyond 10 microns, and lampblack paint falls off
in its absorption very rapidly beyond 15 microns. It is necessary, in order
to obtain exact knowledge, to employ some radiating and absorbing instru-
ments which are perfectly radiating and absorbing by reason of their
shape ; that is to say, which approximate to the so-called absolutely black
body.

Hitherto, only one such instrument has been developed, an instrument
of which there is yet no published description, namely, the honeycomb
pyranometer, or Melikeron, recently invented by Abbot and Aldrich.
This instrument consists of 200 deep cells made by fluting a ribbon of
thin manganin, the whole presenting a surface comparable to a honey-
comb, in which the rays penetrate deeply and are absorbed by repeated
reflections. The heat produced by the rays of the sky or outgoing to the
sky can be compensated by the introduction of the energy of the electric
current. This instrument is but just past its experimental stage, and only
a few as yet unpublished measurements have been made with it.

Spectrum observations ought also to be undertaken in the region of
wave-lengths from 15 to 50 microns. Rock salt is no longer available in
this region, so that some special optical instrument, either a special grating
or a special prism to be made of potassium iodide, must be employed.

NEEDED INVESTIGATIONS

It will be seen from these remarks that the most outstanding needs in
the investigation of radiation for meteorological purposes are:

First, the continuance of accurate observations of the variation of the
sun. These measurements are now going on under the auspices of the
Smithsonian Institution in Arizona and Chile, but should preferably be
supplemented by the provision of two additional stations, perhaps in North
and South Africa, so that variations of the sun could be adequately
studied every day in the year. Twenty-seven thousand dollars a year
would provide the two stations within two years and maintain them
thereafter perpetually.

Second, the painstaking studies of terrestrial cloudiness, its causes and
its effect on the incoming of solar radiation.

Third, the study of the quantity and variability of ozone in the upper
atmosphere, its dependence on solar and terrestrial conditions, and its
influence on terrestrial temperatures.

Fourth, extensive studies with the "black-body" nocturnal radiation

METEOROLOGY 31

instrument, and if possible the development of new instruments of that
class.

Fifth, an investigation of the effects of terrestrial atmospheric con-
stituents on rays between the wave-lengths of 15 and 50 microns should
be undertaken. This involves the development of special optical means
to take the place of rock-salt prism as a dispersing medium, since rock salt
is non-transmissible to the rays in question.

RELATIONS BETWEEN SOLAR ACTIVITY AND ITS

VARIOUS ASPECTS AND THE PHENOMENA

OF TERRESTRIAL WEATHER

By C F. Hakvin

My contribution to this discussion is an appeal for a more careful and
consistent recognition of the effects and operations of chance in the study
of data which may be employed in investigations of solar and terrestrial
relations, periodicities, etc. Weather conditions, atmospheric transmission
of radiation, magnetic phenomena, sunspots, observed intensities of radia-
tion, and values of like phenomena are subject to large and irregular
accidental variations, due account of which must be taken in reaching

conclusions.

SOLAR RADIATION

My point of view is illustrated in an admirable manner by figure 1,
which serves to show how weak the observational basis still is to justify
the claim that there are important irregular changes from day to day in
the intensity of solar radiation.

Discussion of diagram. — ^The vertical lines of the diagram represent
throughout each group 25 observations, counted from the beginning of
the group. The points on the zigzag lines are simply consecutive obser-
vations without reference to the interval of time between observations.

The chronological sequence of the groups of data is represented by the
numbers 1, 2, 3, etc. Only a few observations were made between 1902
and 1907 at Washington at wide intervals. From 1905 to 1918 (groups
2 to 7, inclusive) the observations were made at Mount Wilson, Calif.
In all these cases the interval between observations is as nearly as
possible one day, although periods of bad weather frequently caused two
or three, or more days sometimes, to intervene. As a rule, observations
were made only from June to November.

Extreme variability is shown in the observations in group 1. Groups
2, 3, 4, 5, and the first portion of 6, show distinctly a lower order of
variability, although occasional extremely high and low values occur in-
frequently.

Beginning at K, group 6 for 1912, great variability again appears in the
consecutive values extending into, although diminishing during, 1913.
The high value at K, group 6, marks the arrival at Mount Wilson, Calif.,
of the dust from the Katmai volcanic eruption.

32

METEOROLOGY

So

Id

'J

S

**«

if .

82

E

tt ♦*

cog

4^ 2k
K^
it

M

*«^ B

C o
of

O viH V^

•C.2B
o c-o

B
ft

B

£5^

METEOROLOGY 33

From 1914 to 1918, well after the atmosphere had cleared itself of the
Katmai dust, the variability became quite similar to observational results
at Mount Wilson prior to the Katmai year, 1912.

Group 8 represents the observations at the station at Calama, Chile.

All observations, including the first portion of group 8, were made by
the holographic method.

Finally, the last group of observations, ending December 31, 1919, were
made by a new empirical method based on the bolc^aphic method but
permitting two or three observations to be made the same day, thus giving
a mean average value for a day of higher accuracy.

This diagram tells a very important story with great force and plain-
ness. Great variations in consecutive values of intensity mark the early
observations in Washington with imperfect equipment and poor atmos-
pheric conditions.

Observations at Mount Wilson from 1905 to 1912 show far more nearly
constant values of radiation until the arrival of the atmospheric dust from
the Katmai volcanic eruption, after which day-to-day or consecutive values
showed great variations. Everyone probably ascribes these increased
variations, not to increased solar activity, but to inaccuracies of measure-
ment due to atmospheric dust. The large variations disappeared with the
dust. Furthermore, some increased accuracy (smaller variations) char-
acterized the observations at the station at Calama, Chile, either because
of the better instrumentation, greater observational experience, or better
observing conditions, or all of these in combination.

Finally, it is most striking that a further marked reduction in day-to-day
variability immediately resulted from the introduction in 1919 of the
pyranometer method of observation.

The percentage probable error of a single value has been carefully
computed for each group of observations and the results are showii
graphically in figure 2.

Entirely terrestrial causes easily explain the great changes and gradual
diminution in variability shown by the observations, the accuracy of which
has been wonderfully increased by improvements in instruments, methods,
and location of stations. What are now regarded as good observations
for a single day's work show a probable error as low as 5 or 6 tenths of
one percent. This is remarkable precision. This analysis of the whole
body of radiation data brings one face to face with the important
question :

Is all of this 5 to 6 tenths of one percent of day-to-day variation in solar
radiation intensities real error of measurement only? Or is part of it
error of mieasurement and part real solar change f If the latter, what are
the respective amounts of each variation f

Some conclusive answer to this question is necessary before inferences
and claims of solar and terrestrial correlations can be set up and justified.

It can never be claimed, of course, that single daily values, however
carefully made, are perfectly accurate. Probably simultaneous observa-

54

METEOROLOGY

tions at several stations is the only answer to this question. Caution is
necessary even here, because mere coincidence of variations due absolutely
to errors only will come in to affect comparisons at two stations. If, for
example, e is the probable variation of, say, a season's work at two per-

mpTi

mrr

H- Mt. Wilson. Cal.

KUlama

Fig. 2. Height of bars shows probable error of an observation for a single day of
the intensity of solar radiation as measured by the Astrophysicat Observatory of the
Smithsonian Institution at Washington, Mount Wilson, Calif., and Catama, Chile.

fectly equal stations, then for pure chance relations between values the
variation c„ of the mean of the two must be

<m

VT

This furnishes an interesting test to apply to the simultaneous obser-
vations at Arizona and Chile when these are released by the Smithsonian
Institution.

TERRESTRIAL DATA

No serious study of any kind dealing with hidden or obscure relations
between data subject to large irregular variations should ever be under-
taken without a careful application of the principles of probabilities and
a consideration of the operation of the elements of chance upon the
phenomena under study.

In the discussion of this portion of the subject the author gave the first

METEOROLOGY

35

public account of the mathematical and graphic device which has been
designated The Periodocrite.

Periodocrite^ is a word coined from Greek roots signifying a critic, a
judge, a decider of periodicities, and is a name applied to a mathematical
and graphic method or device which has been developed to aid in the
conclusive separation of obscure and hidden cycles and periodicities pos-
sessing a real existence from those whose essential features are only such

""7

W^t

^^P

tirfoi

'ilclty

r*""

1

r 1

1 7i

/

4

•

A

1

t
\

•

■<^

^ .

X

X

•

A

f'

•
•
f

T

J9

••

1

•

(

)

/#

V,

•

i

• 9

J

J

f -

^ .i

f J

r A

r .A

F .4

r .;

f .4

f 4

f 4

9

Fig. 3. Rainfall periodocrite : X» Annual cycle five stations in Iowa,
36>year record ;•, Annual cycle Washington rainfall, 50-year record;
-f > Annual cycle Boston, Mass., 103-year record, very feebly defined;
3, A 15-noonth sequence Iowa rainfall; other sequences, 15 months,
16 months, one-ninth the variable sunspot period, like the circles, all
fall in the class of perfect fortuity.

as would result from, and can be explained by, entirely chance combina-
tions of the data employed.

The periodocrite does not disclose or discover the length of suspected
periods or cycles. Other methods, such as the harmonic analysis, Schus-
ter's periodogram, or any of the many methods which have been offered
for this purpose must first be employed to ascertain the proper length of
any suspected cycle.

* Prof. C. F. Talman supplied this name from K^pio&)S, a period -{- Kpinyf ,
a judge, decider, umpire, from Kpcya, to separate, investigate, judge.

36 METEOROLOGY

The theory of the periodocrite depends upon the principle that in an
entirely fortuitous combination of data the standard deviation, vn of a
result made up of n individual observations in combination is given by

the equation <rn ^—7=^ in which fr^\A the standard deviation of the orig-
inal data not in combinations.

Writing y= — ^ and x= -7= we get

^o vn

y=x

which is the equation of a line of perfect fortuity passing through the
origin of coordinates (see figure 3). It can also be shown that for perfect
periodicity y=constant=l.

The full account of this aid to scientific investigation of periodicities
has been published in the Monthly Weather Review for March, 1921, to
which the reader is referred for the development of this idea and certain
related matters dealing with the discussion of meteorological data briefly
outlined in the address.

DAILY METEOROLOGICAL CHARTS OF THE WORLD

By Edwaid H. Bowb

INTRODUCTION

Investigations in the field of general meteorology have been and con-
tinue to be restricted and handicapped by the fact that there are not
available to investigators daily charts of the world's weather. It is neces-
sary in investigations in meteorology and particularly in weather fore-
casting that problems now but imperfectly understood be considered from
a world-wide viewpoint, for there can be no doubt that much that we call
"weather^' is not of local origin, but has its inception in the general actions
and reactions that involve at times the atmosphere over an entire hemi-
sphere and possibly both hemispheres. Hence, investigation based on a
study of daily synoptic charts for a limited area, such as Europe, the
United States of North America, or of India, can lead to but an imperfect
understanding of the general physical processes that are in operation to
produce our day-to-day weather. Moreover, when it is understood how
rapid are the changes in speed, direction of movement and magnitude of

METEOROLOGY 37

areas of high and low barometric pressure, there arises the natural desire
to look into the observable facts over a world-wide area in an attempt
to determine the causes of them. Meteorology without a world-wide
weather map is laboring under difficulties as great, or greater, without
realizing it, than astronomy without its star charts. Hence it is contended
that many of the important problems of meteorology will not and cannot
be solved until there be available daily synoptic charts of the various
meteorological elements of, relatively speaking, the entire world. It would
redound to our credit if the American Geophysical Union should become
instrumental in bringing this about.

NEED FOR WORLD-WIDE CHARTS

There is reason to believe that meteorology has for its goal the making
of accurate forecasts of' weather, temperature and wind for long periods
in advance. Such an attainment would not only mark a distinct advance
in the science of meteorology, but in its practical application would be of
great economic importance to the life of the nation.

The problem of making such forecasts has for years received considera-
tion not only from meteorologists of good repute — scientifically speaking
— but also from others untrained or at least not having a thorough grasp
of the question. Many explanations of the frequently marked deviations
from the normal that occur in the meteorological elements at any given
place have been given publicity. Much attention has been given to the
question of cycles, even more has been given to the question of sunspots,
and recently attention has been focused on the so-called variations in the
solar constant. But all these efforts fail to offer a method that permits
the making of long-range forecasts of a sufficient degree of accuracy to
have a practical application of any importance.

One naturally inquires why nothing really definite has been evolved that
might be useful in long-range forecasting, and the only possible answer
that can be made is that the right combination has not yet been found.

My impression is that the answer will probably be had from a prolonged
and careful study of world-wide meteorological conditions, and this study
is possible only in the event that daily S3moptic charts embracing the major
part of the world are available. In my opinion all marked deviations from
normal weather and temperature are associated with changes in magnitude
and position of the so-called "great centers of action." Any material
addition to our knowledge of these must come from the study of world
charts, for it is only by such a method that we can hope to gain a more
comprehensive knowledge of the general atmospheric circulation and the
resulting changes in the centers of action. In Supplement No. 1, Monthly
Weather Review, 1914 (Bowie and Weightman), it is stated that:

Conspicaously abnormal pressures in the regions of these so-called "centers of
action" are related to marked departures from normal weather and temperature
condttions in the United States. Some authorities assume that these abnormal
distributions of pressure are due to extra-terrestrial and others assert that they are
due to terrestrial causes. If it be true that the solar output is a variable quantity.

38 METBOROLOGY

it is possible that the solar variations are associated with marked changes in pressure
in the "centers of action/' and thus may be found a key for defining for consideraUe
periods in advance the general character of coming weather changes for a given
region. If on the other hand abnormal pressure distributions occur with an unvary-
ing solar radiation, the causes thereof must be traced to a terrestrial source. The
vanring effects of the nearly constant radiation on land and water surfaces and on
the air under different conditions of temperature, water vapor content, dust content,
etc., are sufficient in the minds of some writers to explain these phenomena, L e.,
the changes in the position and magnitude of the "centers of action."

Regardless, however, of the cause of abnormalities in the "centers of action" the
importance of their relation to the character and paths of storms in the United
States is well recognized and therefore should be carefully considered in day-to-day
weather forecasting in the United States. To illustrate: Of the centers of action
that affect the weafiier conditions of the United States east of the Rocky Mountains,
the subpermanent high over the middle latitudes of the North Atlantic Ocean i%
perhaps the most influential. When this is well developed and stable, temperatures
above the seasonal average are to be expected over the great central valleys and the
eastern and southern states, and the areas of high and low barometer crossing the
United States will move in high latitudes and pass on to the ocean by way of the
St Lawrence valley. In fact, all prolonged periods of heat in the regions east of
the Rocky Mountains occur simultaneously with the abnormal development of this
subpermanent high. When, however, it is weak and ill-defined, cool weather prevails
over the eastern half of the countiy.

Again, the variations in the position and magnitude of the elongated subpermanent
area of low pressure that normally extends from southeastern Alaska westward to
Kamchatka, have a decided influence on the characters of, and courses followed by,
storms that cross the United States. If this Aleutian low is north of its normal
position, lows will move along our nortfiem border; whereas, if it is south of its
normal position, lows will move far south of their normal tracks and stormy
weather with great alternations in temperature will occur over the United States.

Perhaps the best examples of unusual winters in the United States are
those of 1917-1918 and 1920-1921. In the former, which was one of
great severity, the pressure was abnormally high over Alaska and the
Aleutian Islands ; while in the latter, which will go down in meteorological
history as one of the mildest known, the pressure was much below the
normal over those regions.

To be able to formulate correct forecasts, a knowledge of the general
circulation is fundamental. It is questionable whether a proper under-
standing of the general circulation can be gained from monthly averages.
It certainly is not as stable as the text books would lead us to believe, for
there are frequently marked changes in both the surface and upper air
flow. It has been customary to think of the general or primary wind
circulations of the two hemispheres as separate and distinct, but this view
is not tenable. A cursory examination of plate 14, "Bartholomew's Phys-
ical Atlas,'' Meteorology, volume III, shows a tremendous seasonal inter-
flow between the northern and southern hemispheres, indicating that the
two systems of general circulation are in a way interlocked. In the winter
of the northern hemisphere the air flows normally from the interior of
Asia southward over the Indian Ocean, eastern Africa and the East
Indian Archipelago on beyond the equator as far south as northern Aus-
tralia; in the summer of the northern hemisphere a return flow takes
place over essentially the same geographical area. No one can say, for
the lack of the necessary data, whether these currents do not bring about

METEOROLOGY 39

profound changes from normal weather and temperature conditions over
large areas outside the regions where these flows and counterflows are in
operation, but it seems logical to suppose that such is the case. Further,
the trades and antitrades are probably not fully understood. Certainly
these wind systems undergo pronounced changes that are independent of
the seasonal changes. In connection with the antitrades, Sir Napier Shaw
in a recent number of Nature remarks :

At the same time I may remark that I find it very difficult to grasp the meaning
that is intended by "anti-trades." The original convection theory suggested that the
anti-trade was the trade returning up aloft above its old patii, but, so far as I can
understand the situation, the track of the wind from the equator must begin from
the east and become southwest by what I will describe as the hurricane track. On
the other hand, a southwest wind may be a part of the westerly circulation diverted ;
the difference of origin of the observed southwesterly wind is of some dynamical
importance.

It seems possible that the antitrade may be a northward extension of the
southeast trade of the southern hemisphere, which on crossing the equator
is turned to the right by the deflective force of the earth's rotation, and
being warmer and of less density, overruns the low-lying northeast
trade. But this is not definitely known and sufficient data are not at hand
to prove or disprove the assertion.

Bjerknes has recently given meteorology the term "polar front," a line
of discontinuity separating the polarward from the equatorward flowing
winds, and he has urged the collection of daily meteorological observations
from larger geographical areas that this "polar front" may be delineated
on the weather charts for the aid of the forecaster. He believes this
essential to forecasting, for the theory he advances places the origin of
cyclones and consequently all marked variations in weather, temperature
and wind changes along this line of discontinuity. His presentation of
the idea of the "polar front" and its attendant phenomena is worthy of
him, and augments the necessity of observations over large geographical
areas in weather forecasting.

It will be seen from the foregoing that meteorology must in the near
future consider the question of securing observations from every accessi-
ble place and assemblit^ them for the construction of daily world-wide
weather charts at one or more great world centers for intensive study.
First, there must be a skeletonized chart based on observations collected
by cable, radio and land lines, and, second, a more nearly perfect and
complete chart based on the same observations supplemented by those
collected from remote land areas in which cable or radio is not available
and from ships at sea. The former chart would serve for day-to-day
forecasting; the latter, for study purposes and eventually for long-range
or seasonal forecasting.

HISTORY OF THE MOVEMENT FOR WORLD-WIDE CHARTS

The need of a daily synoptic survey of the earth's atmosphere was co-
incident no doubt with the beginning of synoptic weather charts, which
was at approximately the middle of the nineteenth century. We learn

40 METBOROLOGY

that at the first meeting of the International Meteorological Congress,
assembled at Vienna in 1873, a proposition was adopted to the effect :

That it is desirable, with a view to their exchange, diat at least one nnifonn
observation, of such a character as to be suitable for the preparation of synoptic
charts, be taken and recorded daily and simnltaneonsly throogfaout the world.

Later, on December 9, 1876, it was announced that the United States,
through the Chief Signal Officer, U. S. A., no doubt inspired by the late
Prof. Qeveland Abbe, then assistant to the Chief Signal Officer, was
undertaking the task of establishing cooperation for the recording and
exchange of simultaneous meteorological observations between the United
States and the following named countries: Algeria, Austria, Belgium,
Great Britain, Denmark, France, Germany, Italy, The Netherlands, Nor-
way, Sweden, Switzerland, Turkey, Greece, Canada, the Hawaiian Islands,
Dutch Guiana and Japan. This cooperation extended to naval and mer-
chant vessels of these nations, and thus were secured the simultaneous
observations of atmospheric changes over much of the northern hemi-
sphere. Thus came about the "Bulletin of the International Meteoro-
logical Observations," published by the Signal Service, U. S. A., for the
years 1877-1887 — incomplete, it is true, as to world-wide weather maps,
but a remarkable contribution which has left its impress on meteorology
even tmtil today.

An effort to accomplish the preparation of northern hemisphere weather
maps by means of daily telegraphic reports for the purpose of extending
the forecast period to cover the general weather of the United States was
undertaken by the U. S. Weather Bureau in 1907, and the area covered
by such reports grew until the outbreak of the Great War in 1914, when
the scheme was unavoidedly interrupted. No other really effective efforts
looking to the preparation and publication of even partial world-wide daily
synoptic charts are known to me.

PRESENT STATUS OF DAILY SYNOPTIC METEOROLOGICAL CHARTS

Nothing approaching a world-wide daily synoptic chart is prepared and
published by the meteorological service of any nation. Instead, every, or
nearly every, national meteorological service decides for itself: (a) the
scale of the map, (b) the units of measurement, and (c) to a greater or
less extent the hours of observation. In addition to the various charts
of land observations, charts are also made of the meteorological conditions
over one or more of the oceans. Daily weather maps for their respective
geographic areas are now prepared and published by the United States,
Canada, Mexico, Argentina, Chile, Brazil, Japan, China (Zi-ka-wei Ob-
servatory), Australia, India, South Africa, Great Britain, France, Portu-
gal, Belgium, The Netherlands, Norway, Sweden, Denmark, Germany,
Austria ( ?), Russia ( ?) and others. There is thus available a tremendous
mass of valuable data awaiting action that will assemble them into one
standard, world-wide weather map which will permit investigations in
meteorology to be carried beyond any point now possible.

METEOROLOGY 41

RECOMMENDATIONS

This matter is believed to be of such importance at the present time that
it is contemplated recommending in appropriate form that some action be
taken by the Geophysical Union for the accomplishment of the desired
objects through international cooperation.

There is no doubt in the minds of those familiar with the present status
of meteorology that the carrying out of this proposal will be well worth
while, not only from a scientific standpoint but also from the standpoint
of service to the general public.

The preparation of the data for charting and the printing or lithograph-
ing of the charts for American use could best be done by the U. S.
Weather Bureau, but to do this, additional funds must be provided
through congressional appropriation.

WORLD AEROLCX5Y
By Wnxis Ray Gbbgg

Aerology may be very simply defined as '*the study of the free air";^
world aerology, as an extension of that study to all parts of the world.
By this we mean not only the continental areas, but the seas as well ; and
not merely sections of a hemisphere, but from pole to pole. It is our
'purpose to review very briefly ( 1 ) what has been and what is being done
toward this end ; (2) more particularly, to outline what can at once and
also what should later from time to time be undertaken.

PAST AND PRESENT

Methods, — ^As early as the middle of the 18th century kites were used,
by William Wilson at Glasgow University and by Benjamin Franklin at
Philadelphia, in making free-air observations. Others followed their ex-
ample, with more or less success, but it was not until about 1890 that the
kite came into general use for this purpose.

So far as known, the first manned balloon ascent in the interests of
science was made by Robertson and Lhoest in 1803. During the next 75
years much interesting information as to free-air conditions was obtained
by means of numerous similar ascents, among the most notable of which
were the classic voyages of Glaisher, Flammarion, de Fonvielle and
Tissandier.' Unlike the kite, however, the manned balloon has in recent
years suffered a decline as a means of aerological exploration, because of
the large expense involved and the impossibility of providing satisfactory
exposure of the instruments. Although the observations made by these
two methods were extremely interesting, yet prior to 1890 they 3delded
comparatively little of value, owing to their fragmentary character and
none too great accuracy.

' See "Meteorological Glossary," British Meteorological Office, M. O. 225 ii, 1918,
p. 16. London.
' "Travels in the Air," edited by James Glaisher, F. R. S., Philadelphia, 1871.

42 METEOROLOGY

Since 1890 rapid strides have been made. The kite has been developed
from a mere toy into a very efficient means of exploration. With it
heights slightly exceeding 7 kilometers have been reached, although the
average daily height is a little under 3 kilometers. Recording instruments
carried by these kites furnish information as to pressure, temperature,
humidity and wind at various heights and their changes from day to day,
season to season, and under various types of weather at the earth's surface.
Since the kite can be flown only when there is appreciable air movement,
its use has in some instances been supplemented by that of a small captive
balloon, and thus we have some records in calm weather. Generally
speaking, however, the captive balloon has proved to be rather unsatis-
factory and its use has been largely discontinued.

For exploring the air to greater heights than can be reached with kites,
so-called "sounding" balloons are used. Made of pure rubber, filled
with hydrogen and carrying self-recording instruments, these balloons
have given us information of great interest and value to heights of
30 kilometers or more. Smaller, so-called "pilot" balloons, because of
their comparative cheapness and convenience in handling, have in recent
years come into general use for observing wind direction and speed. On
clear days, when the wind is not too strong, these balloons can be fol-
lowed by means of theodolites to heights well above 10 kilometers.

All of these methods have been rather extensively employed in Europe,
particularly in England, France and Germany, and in the central and
eastern portions of the United States. Some of them have been used to a
limited extent also in Canada, Australia, Java and Argentina, as well as
on a few expeditions of short duration to different parts of the Atlantic
Ocean. In addition, there should be mentioned the great mass of cloud
observations, some of which, particularly those during the International
Campaign of 1896-97, were accurately and systematically made by means
of nephoscopes and theodolites and furnished information, not only as to
the heights and other characteristics of the clouds themselves, but also as
to wind conditions at various levels.

Results. — Although a considerable amount of data has been gathered by
the methods above outlined, it must be confessed that we know even yet
comparatively little with reference to what is going on in the atmosphere
above the earth's surface. The general state of our knowledge can be
briefly summarized as follows :

(a) For parts of Europe and the United States we have well estab-
lished average monthly, seasonal and annual values of all the meteoro-
logical elements from the surface to about the 5-kilometer level. Pressure
of course always diminishes with altitude ; temperature and humidity do
so on the average, except that in the north-central portions of the United
States there is a temperature inversion in the lower levels during the
winter months; wind velocity increases, sharply in the lowest half kilo-
meter, more gradually above that height ; and wind direction is in the mean
very nearly westerly at all levels, except in the southern part of the United
States, where, during the sununer, it is south to east near the surface.

METEOROLOGY 43

(b) Of conditions between 5 and 25 to 30 kilometers we have rather
limited information. We know that the temperature continues to diminish
at a fairly uniform rate until a height of 8 to 18 kilometers is reached —
this height varying with latitude, season, and sea-level pressure ; above this
limiting height the temperature ceases to diminish and in fact has a
tendency to increase to some extent, at any rate during the summer half
of the year. The boundary plane between the lower region of temperature
decrease, known as the troposphere, and the upper region of little tempera-
ture change, known as the stratosphere, is in general well defined. Clouds
do not occur in the stratosphere and winds generally have lower speeds
here than in the troposphere. There is some evidence that at still greater
heights wind direction changes from westerly to easterly, but data on
this point are not conclusive. Other characteristics of the stratosphere
are the lower temperature and greater height of its base in low than in
high latitudes and during falling than during rising air pressure at the
earth's surface; also, its greater height in summer than in winter.

(c) Of the relations found to exist between surface weather and free-
air conditions, perhaps none is more significant than that between surface
temperature distribution and winds in the upper levels. As is well known,
the winds at and very near the surface conform quite closely to the surface
pressure gradient, but at greater heights they often depart widely from it.
If the temperature is fairly uniform over wide areas, the free-air winds
are very nearly parallel to the surface isobars, and show that anti-cyclones
and cyclones extend as such to great heights. If, on the other hand, the
latitudinal temperature gradients are steep at the surface and also, though
to a less extent, in the higher levels, then the surface pressure systems
lose their identity at a very low altitude, the isobars opening out on the
north side of cyclones and on the south side of anti-cyclones, and the
winds veering or backing from those at the surface in conformity with
the altered pressure distribution at the higher levels. This relation of
free-air winds to surface temperature distribution has not thus far been
accorded the attention it deserves. With the development of aviation and
the resulting demand for accurate free-air wind forecasts, the significance
of this relation must necessarily receive increasing recognition.

The foregoing summary is very sketchy and incomplete, but it will
serve as a basis for the consideration of problems which must be attacked
in the future, if real progress is to be made.

THE FUTURE

The present age, to a greater extent than any in the past, may be called
an "age of projects." More and more mankind is giving heed to Emer-
son's exhortation, ''Hitch your wagon to a star," and perhaps this is an
especially appropriate motto for the aerologist to adopt as his own. Of
the many ambitious plans that we read and hear about, some undoubtedly
will yield negative results only, but it is equally certain that others will
contribute very materially to human welfare. And it is better that some

44 METBOROLOGY

should fail than that none should be tried. The projects to be presented
here are not visionary, but on the other hand very practical ones, and for
the most part they are not difficult to carry out. They will be stated in
the order in which it is believed they can be put into execution.

1. Further study of data already accumulated, — ^There is much mate-
rial now available that has not been sununarized and studied in detail or
properly applied to the problems of aviation and forecasting, and to the
solution of perplexing questions relative to the larger features of atmos-
pheric circulation. One of the first things to be undertaken is the prepa-
ration of such a summary. Very few men, outside of the government
services, are giving the subject any thought. Those in the government
services can devote comparatively little time to it, because of other more
pressing duties. There are needed at once for this purpose half a dozen
well-trained men (well trained both in theory and in field experience)
who can give all of their time to this subject for a period of 3 or 4 years.
This, then, is a comparatively simple project — one requiring only a small
outlay of funds, but giving results of immense value.

2. Development of new methods of observation, — As already stated,
nearly all observing at the present time is done with nephoscopes, kites,
pilot and sounding balloons. All of these methods have well-known limi-
tations, but should be continued. There should be added, if possible,
observations with kite balloons and airplanes. Kite balloons, although
more expensive than kites and pilot balloons, would furnish data of
correspondingly greater value, since they could be used with greater
regularity, irrespective of weather conditions. Indeed, their use would
be limited only by very high winds, and records could thus be obtained
under conditions unfavorable for kites ; in other words, by combining the
two methods, the atmosphere could be explored up to 3 kilometers prac-
tically every day in the year. Like kites, their use would be restricted to
regions not frequented by airplanes, because of the danger of fouling with
the cable. Their use would be further restricted to places where hydrogen
could be obtained. These limitations, however, are no more serious than
others under which we now labor and can be overcome at comparatively
small expense, when we consider the great value of the results obtained.

Development of suitable apparatus for use in airplanes should be pushed
vigorously. As aviation expands, there will necessarily be a large number
of places at which regular daily flights will be made at and above the fields
for purposes of testing the machines and the training of pilots. Obser-
vations during these flights would add little expense and would provide
information not otherwise obtainable, such as the thickness of cloud
layers, etc. Work along this line has been done in England and France,
but thus far to no great extent in this country, because of inadequate
appropriations. It will be taken up as soon as funds for the purpose are
made available.

3. Extension of observation stations to all parts of the world. — This
must be done through international cooperation, but the United States can

METEOROLOGY

45

make a good start by more completely covering its own territory, includ-
ing Alaska, the Hawaiian and Philippine Islands, etc. In making such
an extensi(Mi, and a further extension later by all other countries, two
separate and distinct purposes are to be served: (a) the furnishing of
current information of immediate practical value to aviators; (b) the

m^-^ ^

^ir^PS^B^SMy^

s ^

fC'v^^^^^f^'*

-i

K^K*

.. y

'0 ' 3L

Z-.

-/r~-s. j^^"^ '

k

1 ^ ^ -1

Fig. 4. Average summer values of pressure, temperature, density and resultant wind

at the 3-kilometer level

collection of statistical information required to explain the physical causes
of various phenomena and, as a necessary consequence, to increase the
accuracy of weather forecasting. For the first purpose it is sufficient to
establish observii^ stations having comparativdy simple equipment, by
means of which the atmosphere may be explored to moderate heights only.
It 19 not essential that temperature and humidity be observed, but it is
essential that frequent observations be made of wind, cloudiness and visi-

46 METBOROLOGY

bility, these being the factors of vital interest to aviators. For the second
purpose a much more comprehensive prc^^m is necessary. We should
have accurate values of temperature and moisture as well as of wind.
Observations should extend to as great heights and be as nearly continu-
ous as possible, in order that we may know the diurnal and annual varia-
tions throt^hout the troposphere and much of the stratosphere; the
characteristics of the atmosphere under different types of surface pressure
and temperature distribution ; and latitudinal and longitudinal variations.
It is absolutely necessary that these data be collected and carefully studied.
Otherwise we shall continue to be bombarded by theories and» worse still,
by sweeping conclusions which can hardly stand the test of further light
on the subject, but which (and this is the unfortunate feature), being
advanced by men of recognized standing, find their way into textbooks as
facts and thus start the student upon an entirely wrong track. As in all
other matters, so in meteorology it is regrettably true that ''a little learning
is a dangerous thing." Specific references need not be made, but it may
be remarked that in the past year or so there have been some particularly
glaring instances of the promulgation of theories, based upon incomplete
data, and of the more or less universal acceptance of those theories.

As examples of the kind of information needed, figures 4 and 5 are
shown. They are based upon all observations thus far made with kites
in this country, and give respectively average summer and winter values
of' pressure, temperature, density and resultant wind at the 3-kil<mieter
level. These and similar charts for other levels, also charts showing
relative humidity and vapor pressure, form part of a summary now in
preparation, to be known as "An Aerological Survey of the United
States." Some of the more prominent features shown in figures 4 and 5
are: (a) the close relation between the latitudinal pressure and tempera-
ture gradients; (b) the small latitudinal density gradient, owing to the
counterbalancing effects of pressure and temperature, i. e., density varies
directly with pressure, inversely with temperature; (c) the slight south-
ward trend of lines of equal values of these elements from the interior to
the eastern portions of the country; (d) the close agreement between
computed and observed resultant winds in the winter season. The less
satisfactory agreement in sunmier is due to the greater frequency of days
with winds too light for kite flying (another argument for the use of kite
balloons and airplanes) ; and (e) the small latitudinal difference in resul-
tant wind speeds, due to the fact that these vary directly with the pressure
gradient, but inversely with the sine of the latitude.

It is probably not a coincidence, but rather a matter of considerable
significance, that the average movement of cyclones in the United States
during the winter, as determined by Bowie and Weightman, is 13.4 m.p.s.^
— ^a value in striking agreement with the resultant wind shown in figure 5.
The agreement is less close in summer, apparently indicating that cyclones

*■ Types of storms of the United States and their average movements. Monthly
Weather Review, Sui>plement no. 1, p. 8.

METBOROLOGY

47

extend to a greater height in that season than in winter, and this, as
already pointed out, is undoubtedly the case.

These figures, however, are not shown with the view of discussing them
as such, but rather with that of indicating how important it is that we
obtain similar information for all other parts of the world — ^the sea as

"^ (^

mi

y""^'^'^ (\

JPi^

^^^

■* "^"S"

r a^ii^

h^'*"

f^'^'^'^f

\ SVfHr£Mt

\aeMSfrr /t§/bm

m

Fig. 5. Average winter values of pressure, temperature, density and resultant wind

at the 3-kilonieter level.

well as the land. Obviously, it is impossible to carry out this program
at once in its entirety. We must therefore start with the most pressing
needs, as follows :

It is well known that the type of pressure distribution prevailing in the
region of Alaska exercises a dominating influence on the weather of the
United States. Similar relations are found in other parts of the northern

« METEOROLOGY

hemisphere and emphasize the importance of having a network of stations*
observations from which would make possible the construction of world
weather maps — ^a subject which has already been presented by Major
Bowie (see page 36). As indicated by him, such observations would
enable the forecaster to follow from day to day the eastward march of
the so-called '^polar front." ^ There should be a string of stations as far
north as possible and, in the southern hemisphere, another as far south as
possible. Some of these stations at least should be provided with equip*
ment for free-air exploration, this exploration to include accurate obser-
vations of wind and clouds by means of theodolites and nephoscopes. A
few should, in addition, be equipped for makii^^ measurements of tem-
perature and moisture, as well as wind, to great heights. These few would
necessarily have to be not too far removed from sources of supply, but the
others, if equipped with radio, could well be located as far north as living
conditions would permit.

It has been said that definite meteorological laws will be established
only from observations made at sea. These are difficult, perhaps impos-
sible at the present time, to make, but there are numerous small islands
where the influences of the land upon the atmosphere are negligible. Data
of inestimable value can be obtained by establishing free-air observing
stations in Bermuda, the West Indies, the Azores and the islands of the
Pacific ; also in Central America, where continental effects would be small.
We know none too much about the prevailing westerlies, but our knowl-
edge of them is voluminous compared to that of the antitrades. Such a
network of stations as I have indicated, especially if operated for a con-
siderable period of time and supplemented by observations from ships at
sea, would provide the information now lacldng and, in addition, would
solve the much discussed and still unsettled questions of the exdiange of
air between the equator and the poles, the movements of hurricanes, etc.

Aside from the settlement of these theoretical questions, and perhaps
more important, is the value of such observations for daily use in fore-
casting. With the development of radio communication, reports from
these stations should be capable of speedy transmission to forecast centers,
where they could be charted on upper-air maps, supplementary to the
world weather maps, already discussed. Their value to aviators need
not be argued. Can anyone doubt their even greater value, with further
study, to the forecasting, not only of day-to-day weather, but also of week-
to-week, month-to-month and possibly year-to-year changes in weather?

^ See V. Bjerknes. The meteorology of the temperate zone and the general atmos-
pheric drculation. Nature, June 24, 1920, S22-524.

METEOROLOGY 49

WORLD DIGEST OF METEOROLOGICAL DATA

By W. J. HUMPHBBYS

Meteorological data are gathered f or, and serve, many purposes :

They are abundantly used in forecasting the weather of the morrow,
but obviously used only once, and hence for this purpose need not be
recorded.

They also are collected in the course of special studies, but the comple-
tion of each investigation renders useless the preservation of the particular
material treated. It is the generalization — ^the law — ^that counts, and not
the isolated values from which it happened to be deduced.

Finally, they are essential to many studies of interrelations between
meteorological elements; to a knowledge of the relation of the weather
in one part of the world to that occurring either previously, simultaneously,
or subsequently, in others ; and to all accurate knowledge of climates and
their changes. For each of these purposes it is necessary that meteoro-
logical data be indefinitely accumulated, and equally necessary that they
be put in manageable form and made widely available.

Now, although fully three-fourths of the surface of the earth is a
meteorological blank, the mass of data already accumulated from the
remaining one-fourth is so vast and heterogeneous as to be beyond the
power of any individual to analyze and study in detail. Furthermore,
even approximately complete sets of these data have been assembled in
very few places. Hence much of the information contained in this
meteorological material certainly is not only unknown, but even beyond
the power of individual effort to know.

Therefore it is suggested that a comprehensive digest of all existing
meteorological data be made and published. A possible way of accom-
plishing this great labor is as follows :

1. Let the data to be published (monthly and annual normals and
departures therefrom, special phenomena, and what not), the units to be
used, the form of publication, and all other details of this kind, be agreed
to internationally.

2. Let each country furnish the digest of its own data.

3. Let the digest for each country consist of the individual digests for,
and made at, its several meteorological stations.

4. Let some one agency, supplied with adequate funds and personnel,
be charged with the duty of assembling sporadic data from countries that
have no official meteorological organization ; and with the further impor-
tant duty of editing the entire work.

In this way the proposed vast labor would be divided up between sev-
eral countries, and further subdivided among many individuals in each
country, and the product of the combined effort of the many workers —
the digest of all the world's meteorological data — soon made available to
every institution that needs it and to every individual who wishes to
study it.

50 MBTBOROLOGY

There then could be students of world meteorological data» and would
be ; now there is none — and can not be.

It will be recognized of course that the plea here is for a greater
"Reseau Mondial/' one covering a larger number of meteorological ele-
ments than does that splendid publication, and also extending back to the
banning of meteorological observations. It would both include and
supplement the data contained in the present Reseau Mondial, but would
not take the latter's place as a convenient annual summary of the more
important elements of the world's weather at selected places.

GENERAL ADOPTION OF THE CENTESIMAL SYSTEM OF

ANGULAR MEASUREMENT— WITH APPLICATION

TO ANEMOMETERS AND NEPHOSCOPES

By ALBXANim McAon

Reviewing an article on "Uniformity in Aerographic Notation/' ^ Sir
Napier Shaw ' calls attention to the common usage of the capital letters
N.E.S.W. for wind directions, and the established usage in Physics of
N for Avogadro's constant, E for Energy, S for Entropy and W for
internal work.

The criticism is constructive and suggestive. The question arises : Is it
not desirable to follow the lead of navigator and magnetidan and use
degrees instead of letters to indicate direction of air flow? There are
some distinct gains from such a usage for the aerographer or chart maker
of the winds. Official weather bureaus record direction on a 45-degree
basis ; that is, eight directions are given. It has long been felt that such
records were not sufficiently detailed. Precision, detail and convenience
are gained by the use of the degree.

There is no mechanical difficulty in getting continuous records of wind
direction for the entire circle. Many forms of anemoscope give such
records. Figure 6 gives such a record sheet based on one used at Blue
Hill for 35 years. The eight cardinal directions are noted ; but instead of
32 points of the compass, as heretofore, the intervals are at 10 degrees,
and thus 36 divisional lines appear instead of the old compass point 11.25.
For convenience in computation there is also introduced Greenwich Mean
Civil Time, beginnii^ at an hour appropriate for changing records on
this coast, noon 75th meridian time being 17 hours Greenwich time. The
sheet, however, is adaptable to any station meridian time, by inserting
the proper hour in the S.M.T. column. If, however, the centesimal system
is to be used, the number of divisional lines is increased to 40, and since
there are 400 grads in the circle, each division represents 10 grads or
9 degrees.

* H. A. 83-4, pp. 16^180.

* Nature, Nov. 4, 1919.

MBTBOROLOGY

51

WJND DiRecTION CHART

AiH tTATlON AT

4^

:

at

r

S^

Mi*

mtjii

wm

r^

i^

L^

sac.

A

k^

aATf

m

kd!

^

'J jaL

<M ; rn nri; r

/« Ji0 ,30 fpA €0 f tp§a . 4 ^

Ml

Liii

7

4

I

w

::

iiiiiM'i Iqi >'^i?''«^ fnViV>i 111111

llXlil

[■P f» *• « •* ''^ *

1^

hJL

fifi

III 1,1.1. 1.1 !

i-l

irr

n

u

tfea

/i

J3L

n

-£l

.a

J3_

a

J»L

Ct.pt.T- MiS»M«n<w
•.lft.7 «7WritfW -rime

Fig. 6. Wind-direction chart.

METBOROLOGY

By the use of such charts, the words westerly, easterly, and other like
terms disappear and the flow is more definitely described. Wind vanes,
unfortunately, do not fly with the wind, but against the wind, the arrow-
head pointing into the wind. On the other hand, in all charts of air flow,
the arrows fly with the stream.

Fig. 7. Wind protractor for use with McAdie nephoscope.

The direction of flow is read to the right, starting from zero, at the
north ; and thus an east wind is defihitely recorded as 100 grads (or 90
degrees), and a south wind as 200 grads (or 180).

The value of the natural sine of 100 gp^ds is 1. The following abridged
table gives the sines, cosines, tangents and cotangents of every 10 grads :

METEOROLOGY 53

grads sine cosine tangent cotangent

100... 1.00000.. 0.0000 oe 0.0000

90... .9877... 0.1546.... 6.3138.... 0.1584

80... .9511... 0.3090.... 3.0777.... 0.3249

70... .8910... 0.4540.... 1.9626.... 0.5095

60... .8090... 0.5878 1.3764 0.7265

50... .7071... 0.7071.... 1.0000.... 1.0000

40... .5878... 0.8090.... 0.7265.... 1.3764

30... .4540... 0.8910.... 0.5095.... 1.9626

20... .3090... 0.9511.... 0.3249.... 3.0777

10... .1546... 0.9877.... 0.1584.... 6.3138

0... .0000... 1.0000.... 0.0000 oc

During the World War those of us who were engaged in aerogra^hic
work in France found it necessary to use the centesimal system. Since
the war» the method has been adopted by the Scandinavian countries.

In nepho$copic determinations, the method has been used with success
at Blue Hill. A comparative dial of the compass, the magnetic and the
centesimal values is given in figure 7.

A SINE GALVANOMETER FOR DETERMINING IN ABSOLUTE

MEASURE THE HORIZONTAL INTENSITY OF

THE EARTH'S MAGNETIC FIELD ^

By S. J. Baknrt

A brief historical statement was made with reference to the measure-
ment of the horizontal intensity of the earth's magnetic field by electrical
methods, and a general description of sine and tangent galvanometers was
given, with the suggestion of an improvement in the latter. Then fol-
lowed a detailed description of a new sine galvanometer, constructed, with
certain exceptions mentioned below, in the workshop of the Depart-
ment of Terrestrial Magnetism.

The base of the instrument, including the tripod, circles, etc., was taken
from one of Wild's theodolites, constructed by Edelmann, and was much
improved by the substitution of non-magnetic parts for parts too mag-
netic, and by the substitution of electrical illumination of the precision
circle for daylight illumination by mirrors.

The magnetometer-box is of pure copper, the damping being chiefly
electro-magnetic. The magnet-mirror is a fine disc of chrome steel with
optically flat and parallel surfaces, being in fact one of the gages made
by the Bureau of Standards. The torsion tube and head are similar to
those of the C. I. W. magnetometers. A suspension of phosphor-bronze
strip with torsional constant about 0.001 is generally used. The telescope
is small but powerful ; the scale is ruled to thirds of mm., on white pyralin,
with all necessary adjustments. The period of the magnet and the damp-
ing, which is adjustable, are such that readings require only a few
seconds.

The arrangement of coils is approximately that due to Helmholtz. The
spool was machined from white Carrara marble impregnated with parafiin
at a temperature near its boiling point. The coils were wound under
tension in a single layer in spiral grooves cut with a carbon diamond tool.
The wire is pure copper, especially prepared in the research laboratory
of the General Electric Co. Each coil is wound in two halves and contains
10 turns with a diameter of approximately 30 cm. and a pitch of approxi-
mately 2 mm. The two halves start from the same horizontal plane 180
degrees apart, so that the distance between centers of adjacent wires is
approximately 1 mm. The axial distance between the centers of the
two coils, or the distance between corresponding turns of the spirals, is
approximately 15 cm. The insulation resistance between adjacent wires is
very high.

The methods of measuring the diameters and axial distances of the
spirals were briefly described and some of the results were given in tables

^Abstract of the report presented before the American Geophysical Union, Wash-
ington, D. C, April 18, 1921.

54

TERRESTRIAL MAGNETISM AND ELECTRICITY 55

and curves, projected on the screen. The magnetic tests, of three kinds,
proving the materials to be satisfactory, were also described.

The theory of the instrument, the method of using it, and the calcula-
tion of the error in the constant of the coils due to construction, as well
as of the other errors introduced in the measurement of the horizontal
intensity, were briefly presented.

It was shown that the errors in reading the circle and the telescope scale
when sufficiently large angles are used, and the error in the constant of
the coil, were quite negligible ; and that the only other error necessary to
consider, viz, that introduced in the measurement of the current travers-
ing the coils, can also be made entirely negligible. In consequence, the
horizontal intensity of the earth's magnetic field can be determined with
an error less than 1 part in 10,000, which more than fulfills all necessary
requirements.

The instrumental work, done in the shop of the Department, chiefly by
Mr. G. H. Jung, instrument-maker, is highly satisfactory.

The report was closed with acknowledgments.^

Department of Terrestrial Magnetism,

Cam^e Institution of Washington.

ACTIVITY OF THE EARTH'S MAGNETISM IN 1915

By D. L. Hazabd

At the meeting of the International Commission for Terrestrial Mag-
netism held at Innsbruck in 1905 a resolution was adopted recommending
that magnetic observatories classify each day according to its magnetic
character as quiet, moderately disturbed, or severely disturbed, using the
notation 0, 1, and 2 for this purpose. This recommendation has been
adopted by different observatories, one after another, so that now nearly
all of the prominent observatories are sending quarterly reports of the
magnetic character of days to the Netherlands Meteorological Institute
and that institution is publishing them, thus making the data available for
all. While this method of characterization is necessarily rough and influ-
enced by the personal equation of the observer, yet the mean of a large
number of estimations (between 35 and 40 at the present time) gives a
very good idea of the relative magnetic condition of the whole earth from
day to day. It does not, however, give an absolute measure of the daily
fluctuations of the earth's magnetism nor does it permit a comparison of
conditions in different parts of the earth.

In order to determine quantitatively as well as qualitatively the varia-

* Since the presentation of this report, the constant of the coil has been redeter-
mined by the use of many additional linear measurements, and two series of simul-
taneous determinations of the horizontal intensity with the sine galvanometer and the
C. I. W. standard magnetometer No. 3 have been made, Messrs. Fleming. Fisk,
Peters, Ives, and Bamett participating. The results obtained showed a satisfactory
agreement between the two different types of instrument A complete account of
the instrument is given in Vol. IV of the "Researches of the Department of Ter-
restrial Magnetism."

56 TERRESTRIAL MAGNETISM AND ELECTRICITY

bility of the earth's magnetism as a whole, the late Doctor Bidlingmaier
devised a method which takes as a measure of the activity of the earth's
magnetism its departure from moment to moment from its normal or
undisturbed condition. As we have no means of determining as yet what
the normal magnetic condition of the earth is, he adopted as the basis for
his computations the mean value for the period under discussion ; that is,
the activity for a day is based on the momentary departures from the mean
value for the day. He found that in determining the activity for the day
with reference to the mean, the computation could be separated into two
parts, first the departures of the hourly mean from the mean for the day,
and second the departures of the individual values in each hour from the
mean value for that hour. If it later should become desirable to refer
to a base value other than the daily mean it would only be necessary to
add a third term, depending on the difference between the daily means
and the new base value. In each step the mean of the squares of the
departures from the base value is computed and this quantity expressed
in y' must be divided by 8ir to get the activity expressed in terms of the
unit 10"" erg/cm' ; that is, the energy per unit volume.

The regular observatory tabulations contain the data for computing the
first part provided the mean ordinate for each hour is tabulated, as is now
the established practice. Computation of the second part would ordinarily
involve the reading of ordinates at frequent intervals for each hour and
the computation of the mean of the squares of the differences from the
mean value for the hour, this being the so-called hour-integral used in
Bidlingmaier's formula in determining that portion of the activity. The
amount of work involved would, of course, be prohibitive and he accord-
ingly simplified the process by reading the ordinates for a limited number
of hours and using the relation between the hour-integral and the hourly
range as a basis for determining the hour-integral for the remaining hours
from the hourly range. When the results were plotted with amplitude
(half range) as abscissa and hour-integral as ordinate, it was found that
the line joining the plotted points formed a smooth curve of parabolic
form. While it was found that the values of hour-integral corresponding
to a given hourly range differed considerably among themselves, as would
naturally be the case because of the varying character of the fluctuation
within the period of an hour, yet it was believed that for most purposes,
where the results would be combined to obtain mean values, the relation
between hourly range and hour-integral derived from a limited number
of hours could safely be used in determining the hour-integral for a long
period, as for a year.

The parabolic form of the curve representing the relation between
amplitude and hour-integral suggested the probability that a linear relation
would be found to exist between the square of the amplitude (or range)
and the hour-integral. In fact, this must necessarily be the case if the
value of activity is to be independent of the sensitivity of the instruments.

At the request of the International Commission for Terrestrial Mag-

TERRESTRIAL MAGNETISM AND ELECTRICITY 57

netism a number of observatories undertook to compute the activity of
the earth's magnetism according to this method for each day of the year
1915, the Coast and Geodetic Survey carrying out the work for its mag-
netic observatory at Cheltenham, Maryland. It was suggested by the
International Commission that other observatories might safely accept the
relation between hourly range and hour-integral as determined by Bid-
lingmaier for Wilhelmshaven for the year 1911. It was thought best,
however, by several observatories, to re-determine this relation in order to
be assured that it did not change from place to place. The results show
that while for Wilhelmshaven the hour-integral was equal to 11^4 percent
of the square of the range, for Cheltenham the factor was 10 percent
and for Seddin, near Potsdam, 8^ percent, and an investigation by Chree,
which included the study of the records of the British Antarctic Expedi-
tion, showed that while for ordinary latitudes the variation in the factor
was not great, conditions were quite different in very high magnetic
latitudes.

When the preparation of this paper was undertaken it was expected
that it would be possible to compare the results from several observa-
tories, but it was found that only the Seddin results, in addition to those
for Cheltenham, were available in printed form.

The geographic positions and mean values of the magnetic elements
for 1915 for these two stations are as follows:

Observatory Cheltenham Seddin <

Latitude 38** 44' N 52^ 23' N

Longitude 76 50W 13 04 E

Declination 6 04W 8 17W

Dip 70 47 N 66 25 N

Horizontal intensity 19417y 18726y

Vertical intensity 55694/ 4289&y

Total intensity 58982y 46806y

It will be seen that the two observatories differ very nearly 90® in
longitude, and while Seddin is much farther north than Cheltenham, the
magnetic dip and intensity are much greater for Cheltenham, the latter
station being nearer the magnetic pole.

The activity has been computed for each hour for D, H, and Z at
Cheltenham and for X, Y, and Z at Seddin, and these three are combined
to get the total activity. The quantities published are the mean value for
each day of the year and the hourly means for each month.

In discussing the results of this method of determining the activity three
things must be kept in mind. First, since the activity is based on the
square of the departure from the mean value, a day of very large disturb-
ance will have an overpowering effect on mean values in which it enters.
For example, the activity at Cheltenham for June 17 was 1477 and the
total for the other 29 days was only 666. Second, Bidlingmaier's concep-
tion of activity is different from the usual idea of activity as represented

58 TERRESTRIAL MAGNETISM AND ELECTRICITY

by a magnetic disturbance. In the latter case we think only of abnormal
variations, whereas he includes all variations, whether systematic (as
diurnal variation) or abnormal. Third, as part of the variation of the
earth's magnetism is a function of local mean time and part is a function
of absolute time, the results for different observatories are not strictly
homogeneous. For this reason a more satisfactory agreement between
the results for different places may be expected if only that part of the
activity is considered which is derived from the hour*integral, as this is
to a greater extent independent of local mean time. Even then, how-
ever, there is some lack of homogeneity, as for example, in the case
of the Seddin tabulations the day begins at Greenwich midnight, whereas
for Cheltenham it b^ns at 5^ G.M.T.

The portion of the activity derived from the hour-int^ral is much
smaller than the part depending on the differences between the mean
hourly values and the daily mean, only about one-eighth as great on the
average for all days and only one-twenty-fifth for the less disturbed days.
The fact that the normal diurnal-variation is such a predominant factor
in the total activity as derived by Bidlingmaier's method raises the ques-
tion whether results of greater value would not be obtained if the diurnal
variation was eliminated, at least in part, in computing the activity.

A comparison of the daily mean values of total activity for the two
observatories for 1915 shows a general agreement, but with considerable
difference in detail, largely because of the third point referred to above.
The total activity is on the average about 15 percent greater for Chelten-
ham than for Seddin, as was to be expected on account of its higher
magnetic latitude, but for many days and even for some monthly means
the Seddin values are greater.

If only the hour-integral activity is used in the comparison, the agree-
ment is much closer, and when the results are smoothed out by taking
five-day means the plotted curves for the two places are almost identical
in phase. The agreement is almost as good with the international char-
acter numbers, both in phase and relative amplitude, and speaks well for
that simple method of determining the degree of disturbance.

A comparison of the monthly means with the relative sun-spot numbers
for the same periods shows little evidence of systematic agreement, thus,
as pointed out by Schmidt,^ confirming former experience in comparing
these numbers with terrestrial phenomena. In this case the lack of agree-
ment is no doubt partly due to the exaggerated effect on the mean activity
for a month of a single day of great disturbance.

As to the diurnal variation of the total activity, the predominant feature
is a maximum occurring about noon local mean time. At Cheltenham
this feature is modified in the months May to August to form a two-
peaked stunmit with maxima about 10^ and 14'^ and a considerable depres-
sion between. There is little variation in activity during the night hours.
The range of activity is greater in summer than in winter, though, as

^ Terrestrial Magnetism, Sept, 1920.

TERRESTRIAL MAGNETISM AND ELECTRICITY 59

already pointed out, the effect of a single day of great disturbance is
overpowering. The above characteristics can be traced at once to the
features of the diurnal variation of the magnetic elements, the two max-
ima at Cheltenham corresponding to the minimum horizontal intensity
before noon and the maximum west declination after noon.

If we consider only the hour-integral activity, we find very little evi-
dence of system in its diurnal variation in the (Ufferent months, but there
seems to be a tendency toward higher values toward the end of the 24
hours. This is more pronounced for Seddin and corresponds to the dis-
tribution of disturbed hours arrived at directly. The effect of a few
disturbed hours is so great, however, that it is hardly safe to draw definite
conclusions from the results of a single year.

Schmidt discusses some other phases of the activity at Seddin in 1915,
in the paper referred to above, but time does not permit going into the
matter in greater detail here. He also makes some comparisons with the
results of simpler methods of determining the activity.

As a result of the activity computations for 1915, I am of the opinion,
which is shared by Schmidt and Chree, that equally valuable results can
be obtained by other methods that involve much less time and labor than
Bidlingmaier's. His method gives undue weight to days of large disturb-
ance in any combination of hourly values, and the introduction of the term
depending principally on the diurnal variation of the earth's magnetism
prevents a satisfactory comparison of results at different stations.

Division of Terrestrial Magnetism,

U. S. Coast and Geodetic Survey.

ON MEASURES OF THE EARTH'S MAGNETIC AND ELEC-
TRIC ACTIVITY AND CORRELATIONS
WITH SOLAR ACTIVITY

By Louis A. Bauer

When attempting to find correlations between manifestations of the
sun's activity and those of the earth's magnetic and electric activity, three
points require immediate consideration :

(1) What shall be taken as an adequate measure of the sun's activity
with respect to such radiations and emanations as are likely to have
an effect upon the magnetic and electric fields of the earth ?

(2) What shall be taken as an adequate measure of the earth's mag-
netic activity, or of the earth's electric activity?

(3) What quantities shall be taken as defining the so-called normal or
undisturbed condition of the earth's magnetic field, or of the earth's elec-
tric field?

With respect to the first question, we have at present at our disposal
the sun-spot numbers, sun-spot areas, flocculi areas, prominences, f aculae,
and solar-constant values.

For measures of the earth's magnetic activity, as well as of its electric

60 TERRESTRIAL MAGNETISM AND ELECTRICITY

activity, we may use fluctuations in the magnetic and electric quantities,
which are more or less periodic in their character, as, for example, the
diurnal range or annual range of the magnetic and electric elements. But
it is also found that during a magnetic storm and for some time after-
wards, the earth's permanent magnetic state, as also possibly its electric
state, has been affected. Thus, we have at our disposal both fluctuations
about a mean value for a certain interval, and also change in that mean
value for a given time. The selection of normal or undisturbed values of
any measure taken may be based upon the international list of so-called
magnetically-calm or electrically-calm days. Though it must not be over-
looked that often the values of the magnetic and electric elements on such
days are affected by a peculiar kind of disturbance. In brief it has been
found that the magnetic or electric elements on a comparatively undis-
turbed day are not necessarily normal values. Rather may the values be
"normal" which lie intermediate between those for the "quiet" days and
those for the days of moderate disturbance.

Every analysis thus far undertaken of any particular magnetic fluctua-
tion indicates that the observed effects are to be ascribed to at least two
systems of forces : E, an external system consisting most probably of elec-
tric currents in the upper regions of the atmosphere; and /, an internal
system consisting of electric and magnetic systems below the earth's sur-
face. The two systems E and / are not necessarily related as though /
were the result of an inductive effect caused by the system £. The system
/ would appear rather as a composite system, composed primarily of a
direct effect and secondly of an indirect effect .which may be related to
the fluctuating E system. Indications have also been found of the pres-
ence of a third system, C, consisting of vertical electric currents which
apparently pass through the earth's surface, either from the atmosphere
or from some internal source. What we observe during a magnetic storm
is the combined effect of the three systems, E, I, and C, and this important
fact must be borne in mind in endeavoring to find correlations between
solar activity and terrestrial activity. It may even happen, as apparently
was the case on May 8, 1902, during the eruption of Mont Pelee, that we
have a world-wide magnetic fluctuation of internal rather than external
origin. Hence, were it feasible, a mathematical analysis should be imder-
taken first of a magnetic disturbance in order that the effects coming from
external sources may be separated from those to be related to internal
ones.

The question has also been raised, since at times a magnetic disturb-
ance on the earth apparently precedes some striking manifestation of solar

activity, whether there may not be also the possibility of a universe dis-
turbance-system affecting both solar activity and planetary magnetic
activity.

As the combined result of my investigations to date, it is found that, in
general, the most successful measure of solar activity, of special interest
here, is a qtiantity indicative of the variability of sun-spottedness during a

TERRESTRIAL MAGNETISM AND ELECTRICITY 61

given period. For example, instead of taking the sun-spot numbers (N)
direct for comparison with magnetic or electric fluctuations, the range
(J?) in the sun-spot numbers per month, or the average departure (D)
of the daily sun-spot numbers from the monthly mean, irrespective of
sign, is taken. The R and D quantities are found to run closely parallel
to one another; preference was finally given to the D-measure of solar
activity as it utilizes all the sun-spot numbers {N) during a month,
whereas, the /^-measure depends only on two numbers — ^the maximum and
minimum sun-spot numbers of the month. The annual mean values of R
and D are furthermore found to run closely parallel with the ^-numbers ;
the monthly values of R and D, however, generally follow a decidedly
different course from the iV-numbers and exhibit a closer relationship with
the measures of the earth's magnetic, or its electric, activity than do the
latter (the N's). Some of these relationships between solar activity,
terrestrial magnetism, and terrestrial electricity (earth-currents, atmos-
pheric electricity, and polar lights) are shown in figures 1 and 2 and are
summarized below.

The adopted measure of the earth's magnetic activity is a quantity,
w = cHv, where H is the horizontal intensity of the earth's magnetic field
at the observing station and v, the observed magnetic variation, or range
of the magnetic fluctuation; € is a numerical factor. It may be shown
theoretically that this value of zv, as a first approximation, is representa-
tive of the energy-change which the earth's magnetic field experiences
during a magnetic variation.

Under certain assumptions it may also be shown that the R and the D
measures of solar activity may be regarded, as a first approximation, as
representii^ an energy-change experienced by the sun during a manifesta-
tion of activity.

Obtaining similarly, as just described for the sun-spot numbers, R and
D measures of solar activity from the solar-constant values (£), which
have been observed under the auspices of the Smithsonian Institution at
Calama, Chile, during 1919 and 1920, it is found that these latter measures
run much more closely parallel with the R and D measures derived from
sun-spottedness than do the numbers E and N.

Connections between sun-spot activity, disturbances of the earth's mag-
netism, earth-currents, and polar lights have been worked out by various
investigators. The present investigation shows that there is a fifth natural
phenomenon — atmospheric electricity — ^by which an interesting and sug-
gestive relationship with solar activity is exhibited. Owing to the many
disturbances to which the atmospheric-electric elements are subject, as for
example during cloudy and rainy weather, it has been more difficult to
establish the existence of definite variations of the chief atmospheric-
electric elements during the well-known sun-spot cycle of somewhat over
11 years than in the case of magnetic effects, earth-currents, and polar
lights. The new results found are based upon atmospheric-electric data
obtained chiefly at four European observatories between 1898 and 1919,

62

TERRESTRIAL MAGNETISM AND ELECTRICITY

,»T^ TIM «••* »•»• 1

J_

Ivk

z^^^

40 — .

^'^ \

y ^ (1) WtUm BitMiii

-^ X"^S^ W^^IPOT

»"""-^^

40 . .\.

//SI

18 A ^^

_Hz »-••—)

9 "^^I

>^

« ^

i=^ ^:^

-Mr-

^' ^

-»^-*-^ _

_ _ / „ ^L^iAtl m .-^r*. ■

« ^*.

.. f ^ -^ -._.. ^...

♦10 * - "

.. PrvrlsliMl)

-43*

0

— . .. -' ^ |W| 1 ^■>^ fltel l^aMkl

^ " fOMH iBIMlliJ

*to /

*f JZ

0 r-*S-

y V- onuaai |9»i ■■•i

-8

"^^^ v^

Jf

0

**^ ^r"

«4

-^ r 0^»«iwiijriP»l Wm9i

♦iJ*

: i:

' \ fnl nrlAtl^ !■ aiT ihrl^

0

/ ^^ J^M^MMmA #«kA« '^^mm^ ft Aft

-8

r -. 0B»«* CD^i »t#u i)

7 \ (f ) lHrlh*« aiff. ADUn^FdMirJ

7lr- Cfmi%tiii«,»,An Bi9B)

^^

U) \

r ■ X *^ — |<B«lMBHMBt 10 0|

• ^"'-^^-

. _ 4- .^ (^1) iMTlfc OmiMi AvtlTity OB)

^^

/ fttftl l^ro. •■ til Ttoval

-^^'^"**"— •«..

. ^ ^ (It) »rtb.ouT«B« A0tiTi«ar OB)

40

v*^ — iJifti ■■•• a>li«jjMiy»i

19M mo

_in4 i9i« 1

Fig. 1. Variations in solar activity, terrestrial magnetism, atmospheric
electricity, and earth-currents during 1905-19^.

TERRESTRIAL MAGNETISM AND ELBCTRICITV

63

the combined data in the case of the potential-gradient thus covering about
two sun-spot cycles. Recent observations on board the Carnegie also indi-
cate a decrease in the electric potential-gradient since 1917, when sun-spot
a<:tivity was at a maximum. A half century ago Quetelet at Brussels and

1900 IflOA

19oa 1912 1«« IteftJ

^M%

J

.

ill

12)

M

lbs

nt

Jbi

M

««!*

mi

\

80

**^

n

J

s

>

^

/

Sv

0
%

4lO

0
-10

4lO

0
-10

.«*

0
-10

%

♦10

0

-10

t4

la

0

A

^

s

/

^

"^

..^

X

^^^

~^

V

^^^

^

f

—

[£jLi

d

^

■L

IB.

1&

U

2]b

nti

Al

Ac

idl

ml

J

1^

ifl

111

D>

^

^

-*>

z

^

7

/

^

^—

>/

^s.

z

^

—

r

[21

di

^X'

Shii

p

z

Blj

AX<

flJB

di

tti

((

■i(

t :

nar

I)

\ri.

t£

»

^B-^V ^A 1

^

:^

—

A

/

^

^

/

Al

JQ

cb

ti

-^

Xfi

|V|

T

Ill

(ft!

Ia:

AL

&£

idi

m

ja

A

lI.Kb]

Sc^

Jt

a

u

^

/

"^

V

-—

^

>

•

ffi)

JQ

El

11

^.

Ln

Um

T

In

m

SSk

uJl

lag

L-fl

r I

oft

tti

Al

.^

^

^

V-

>.

ll

ll_lt

AilU

S

IV

i_

\

n

■^

r

It]

MM

»

«•

Ug

CM

^

^

1M

» w

aa

Bfl

^

■t

/

V.

r^

HH

F

1

/"

s^

^

>

r

A

N

z

s

J

t

^

%

<;

y

~\

Pig. 2. Variation of the electric-potential gradient and of its diurnal
range during sun-spot cycle. (See also Fig. 1.)

Wisclxzenus at St. Louis believed that they obtained some definite results
showing a variation in the potential-gradient dependent upon sun-spotted-
ness. Owing, however, to the uncertainty of results obtained by the in-
strumental methods then in use and because of the necessity of thoroughly
eliminating the numerous disturbances dependent upon meteorological

64 TERRESTRIAL MAGNETISM AND ELECTRICITY

condition*, these previous results have not been accepted, and so modern
treatises on atmospheric electricity omit mention of any possible relation-
ship between atmospheric electricity and solar activity.

More complete publication of the results of the investigations here out-
lined is made in the 1921 volume of the Journal of Terrestrial Magnetism
and Atmospheric Electricity, pages 33-68 and 113-115.

CHIEF RESULTS

1. The earth's magnetic energy and average intensity of magnetization,
as well as the strength of the normal electric currents circulating in the
earth's crust, suffer a diminution during increased solar activity. The
electric currents induced in the earth during periods of increased solar
activity are in general reversed in direction to the normal currents, the
strength of these superposed currents increasing with increased solar
activity.

2. The diurnal range of the strength of earth-currents, as in the case
of the diurnal range of the earth's magnetic elements, increases with
increased solar activity ; at time of maximum activity the range, as shown
by the observations at the Observatorio del Ebro, Tortosa, Spain, during
the period 1910-1919, was about 50 percent higher than at the time of
minimum solar activity.

3. The magnetic effect running a concomitant course with the solar-
activity cycle is retarded, on the average, about one year so that there is
a residual, or an acyclic, effect at the end of the cycle. The actual amount
of retardation, in general, increases with intensity of the sun's activity
or energy. This lag in the magnetic effect may be accounted for by the
fact that the electric currents generated inside the earth during magnetic
storms and magnetic variations continue for some time after the apparent
cessation or diminution of solar activity, or after the period of the varia-
tion experienced. The same lag is shown by polar-light frequencies at
times of maximum solar activity.

4. On fine-weather, or electrically-calm, days the atmospheric potential-
gradient, or the deduced negative charge on the surface of the earth,
increases with increased solar activity, the range in the variation between
minimum and maximum solar activity being about 20 percent. The elec-
tric conductivity of the atmosphere, on the other hand, shows but little,
if any, systematic variation during the solar cycle. Accordingly, since
the vertical conduction-current of atmospheric electricity is derived from
the product of the potential-gradient and the electric conductivity, it is
found that this vertical current also increases in strength with increased
solar activity. It would thus appear that atmospheric electricity, like
terrestrial magnetism, is controlled by cosmic factors. The results derived
here may have an important bearing upon theories of atmospheric

electricity.

5. The diurnal range of the electric potential-gradient as deduced from
the observations on the electrically-calm days, made at the Observatorio

TERRESTRIAL MAGNETISM AND ELECTRICITY 65

del Ebrp, Tortosa, Spain, 1910-1919, is found to increase with solar
activity; the minimum occurred in 1911 and the maximum in 1917,
whereas the sun-spot minimum occurred in 1912 and the maximum in
1917. The range between minimum and maximum diurnal range is about
25 percent. (It appears probable that the same fact just stated for the
potential-gradient will also be found true for the vertical conduction-
current.) Department of Terrestrial Magnetism,

Carnegie Institution of Washington.

THE PENETRATING RADIATION AND ITS BEARING UPON

THE EARTH'S ELECTRIC FIELD ^

By W. F. G. Swann

The paper was devoted largely to a description of certain investigations
on the penetrating radiation in progress at the University of Minnesota,
under the author's direction. It opened with a brief review of the status
of our knowledge with regard to the penetrating radiation.

In a hermetically sealed zinc vessel freed from radioactive air, ions are
produced at a rate of about 8 or 9 per c.c. per second over the land. Ac-
cording to the computations of A. S. Eve, the normal gamma-ray radiation
from the atmosphere is capable of accounting for about 0.06 ion per c.c.
per second, while that from the soil will account for 1.6 ion per c.c. per
second, making in all 1.7 ions. On account of the secondary ionization
resulting from electrons emitted from the walls of the vessel by the pri-
mary radiation, this vahie becomes increased to about 2.5 ions in the case
of vessels of the size ordinarily used. If this value be subtracted from
the 8.5 ions per c.c. per second found over land, there remains about
6 ions per c.c. per second attributable to causes which are not directly
obvious, and this corresponds roughly to the ionization (4 to 6 ions per
c.c. per second) observed over the ocean.

The results of several investigators seem to indicate that, in vessels sur-
rounded by thicknesses of water sufficient to absorb practically all gamma-
ray radiation of the ordinary type, the ionization does not become reduced
greatly below that found over the great oceans, nor does it diminish appre-
ciably with increasing thickness of water, so that, if this ionization is due
to an external radiation, this radiation must be of an extremely penetrat-
ing type as compared with ordinary gamma-radiation. By using a vessel
of ice, in order to be as free as possible from radioactive contamination,
McLennan obtained a value as low as 2.6 ions per c.c. per second over
Lake Ontario, and inclines to the view that even this small residual is
attributable to lack of complete purity of the ice. On the other hand,
results obtained by Kolhorster in balloon ascents up to 9.5 kilometers show
an increase to about 80 ions per c.c. per second at this altitude, in a her-

* Abstract of the paper presented at the annual meeting of the American Geo-
physical Union, Washington, D. C, April 18, 1921.

66 TERRESTRIAL MAGNETISM AND ELECTRICITY

metically sealed, light, tight vessel ; moreover, the rate at which the appar-
ent ionization increases with altitude in the neighborhood of 9 kilometers
is such as to suggest that at altitudes but slightly greater, the ionization
might attain enormous values. The validity of Kolhorster's results has
been questioned by C. H. Kunsman in view of possible complications
resulting from the effect of the low temperature on the insulating material
used; if, however, they should be substantiated, they afford one of the
most convincing evidences that one could wish as to the true cosmic nature
of part at least of the so-called penetrating radiation. By the employment
of precautions such as to prevent all uncertainty as regards leakage, and
by the use of small pilot balloons, the writer is at present endeavoring to
extend the observations of Kolhorster to greater altitudes.

VARIATION OF RESIDUAL IONIZATION WITH PRESSURE

The study of the variation of the residual ionization with pressure has
a very important bearing upon the origin of that ionization. If the resid-
ual ionization were due to alpha-rays emitted from the walls of the vessel,
it would show practically no increase with pressure in the case of a vessel
whose linear dimensions were of an order of magnitude greater than the
range of the alpha-particles. Analagous remarks hold for the case of
soft beta-rays emitted from the walls. In the case of a radiation of cos-
mical origin there are three possibilities, viz, (1) a direct ionization by
the primary radiation, (2) ionization by slowly moving electrons emitted
from the gas by the primary radiation, (3) ionization by rapidly moving
electrons emitted from the gas by the primary radiation, the penetrating
power of the electrons being such as to enable them to go right across the
vessel at atmospheric pressure. A fourth possibility resulting from emis-
sion of electrons from the walls of the vessel by the primary rays, is indis-
tinguishable from a corresponding emission resulting from radioactive
contamination. Ionization of the first and second type would increase pro-
portionally with the pressure until the pressure attained was so high that
the primary radiation itself became appreciably absorbed in passing
through the vessel. Effects resulting from the third t)rpe of ionization
require more detailed consideration.

At each point in the gas there will be a definite value of what we shall
call the ionizing intensity /, i.e., the number of ions which would be pro-
duced per c.c. per second in an element of gas at atmospheric pressure
placed at the point in question. For the purpose of this definition, we
may suppose the element of gas to be contained in a non-absorbing cap-
sule, since it is of course not implied that the whole of the gas is at atmos-
pheric pressure. The number of ions produced per c.c. per second in the
gas at some point O where the pressure is p will be Ip,

Let us consider an element of volume As 6r, situated at a point P at a
distance r from O in some definite direction, dr being element of radius
vector, and As being element of cross section of an elementary cone of

TERRESTRIAL MAGNETISM AND ELECTRICITY 67

solid angle d» drawn from O to P. The contribution of the element of
volume to the value of / at O is

^I = ^f(pr)dsdr

where, assuming the secondary radiation originating within an element of
volume to be proportional to the pressure therein, ^A^ dr dr is the second-
ary ionization intensity at unit distance from P in the direction PO, on the
basis of no absorption, and / (pr) is a factor inserted to take account of
the absorption in passing from P to O, and also of the variation of the
ionizing efficiency along the path of the ionizing agent. By writing / (pr)
instead of / (r), we imply that the diminution of the ionizing intensity per
unit distance (apart from spreading) is proportional to the number of
molecules per c.c. In terms of the solid angle dw above referred to we
have:

A/ = iV/(/^r)pdrd»

If we should now increase the pressure from p to p^, and diminish r
to r, so that pr =■ p^r^^ and dr to dr^ so that pAr = /^idr^, the new element
of volume contained between the radii r^ and (r^ + dr^) in the cone d«>
will make the same contribution to the ionizing intensity at O as did the
old element of volume comprised between r and (r + dr). The sum
total of all the elements of volume in the vessel, corresponding to the
lower pressure will, however, correspond to a sum total of elements of
volume which, at the higher pressure, do not fill the vessel. Hence, the
actual ionizing intensity at O will be greater at the higher pressure than
at the lower pressure. Since q, the number of ions produced per c.c. per
second at O, is obtained by multiplying / by the pressure p in atmos-
pheres, we see that the increase in q, per atmosphere increase, should
itself increase with the pressure. We may extend this statement so as to
include in q the ionization due to the direct action of the external radia-
tion, since this increases proportionally with the pressure. Thus, the
actual ionization in the vessel, due to primary and secondary emission
from the gas, will be less at one atmosphere than at any higher pressure.
If then, the ionization-pressure curve should show a very small increase
of ionization per atmosphere increase at high pressures, we know from the
above that such increase per atmosphere is nevertheless greater ^ than the
portion of the ionization due to primary and secondary action in the gas
within the vessel at one atmosphere. We may infer that any greater ioni-
zation found at atmospheric pressure is to be attributed to radiation from
the walls of the vessel ; this radiation, owing to its absorption at the higher
pressures, results in a diminishing rate of increase of ionization with pres-

^ It would not be quite safe to extend this argument to imply that the ionization
here referred to was necessarily greater than the true natural ionization in the open
air, since a portion of the ionization in a volume of the external air occupying the
space of the vessel would result from secondary radiations originating outside of
that volume.

68 TERRESTRIAL MAGNETISM AND ELECTRICITY

sure. The foregoing discussion has been made for the case where the
primary radiation is so penetrating as to be but little absorbed in passing
through the gas in the vessel, even at the higher pressures, this being the
case which is of interest in discussing the action of a radiation whose
degree of penetration is comparable with that attributed to the cosmical
penetrating-radiation.

Experiments on the variation of the residual ionization with pressure
have been made by several investigators, the most recent being those made
under the writer's direction, by Dr. K. M. Downey and by Mr. H. Fruth.
The main feature of the experiments of the latter investigators lay in the
use of a comparatively large vessel (a sphere one foot in diameter), and
the employment of certain special devices to insure freedom from errors
due to leakage and lack of constancy of the batteries. Dr. Downey's obser-
vations extended in the first instance up to 20 atmospheres, giving over
this range a practically perfect linear variation with the pressure and an
increase of ionization of about 1.2 ions per c.c. per atmosphere increase.
On extending the observations to higher pressures, it was found that the
linear relation ceased to hold in the neighborhood of about 27 atmos-
pheres, and the curves finally became parallel to the pressure axis at pres-
sures above 46 atmospheres. If one were to accept this parallelism with-
out reservatiofi, he would be forced to conclude that the portion of the
ionization within the vessel which was attributable to the direct or indirect
action of the radiation on the gas was immeasurably small.

Mr. Fruth's observations have been made for air, oxygen, and nitrogen
up to pressures of 75 atmospheres, and for carbon-dioxide up to its lique-
fying pressure, with sensibly the same results for all the gases used.
While his curves do not attain as complete a parallelism with the pressure
axis as do those of Dr. Downey, they correspond to an increase per atmos-
phere of less than 0.75 ion per c.c. per second at the higher pressures.* It
is worthy of note that the presence of radium-emanation in the gas would
tend to increase the slope of the ionization-pressure curves. The normal
emanation-content of the atmosphere is such as to produce about 2.3 ions
per c.c. per second. Each additional atmosphere of air would, of course,
carry with it its emanation-content, so that the increase per atmosphere
for normal air resulting from the emanation-content alone would amount
to 2.3 ions per c.c. per second per atmosphere increase. It is therefore
necessary to carefully age the air before use. Recalling that the emanation
activity dies to half value in 3.85 days, it will be readily seen that in the
case of air 3 weeks old the effect of the emanation would become reduced
to a negligible amount. In some of Dr. Downey's observations the air
was aged for a month before use.

In experiments of this kind it is of the utmost importance to insure
complete saturation, and this matter was consequently tested very care-

* Since this was written Mr. Fruth has found complete parallelism for air, oxygen,
and carbon dioxide for pressures above 52 atmospheres, when the gases are per-
fectly dry and dust-free.

TERRESTRIAL MAGNETISM AND ELECTRICITY 69

fully, the voltages used being considerably higher than those at which
experiment showed saturation to have been attained at the higher pres-
sures. One dement of uncertainty not usually considered in relation to
the attainment of saturation must be here referred to. It pertains to the
effect of dust nuclei. Such nuclei could theoretically cause a departure
from a saturation which could not be reduced beyond a certain minimum
however great the field might be, since increase of the field intensity could
not reduce beyond a certain limit the probability of an ion's encounter
with a dust nucleus during its passage across the vessel. Departure from
saturation due to such a cause would not show up by the failure to attain
apparent saturation with increasing field, and it would increase, moreover,
with increase of the amount of gas (and consequently of dust nuclei) in
the vessel. The comparatively good agreement between the results of Dr.
Downey and those of Mr. Fruth, and the agreement of the various results
of the latter investigator among themselves, suggest, however, that dust
did not play an important role in the experiments, particularly when one
remembers that the various experiments corresponded to different samples
of gas, samples which had undergone, moreover, entirely different treat-
ments. However, it is planned to make a very careful investigation of the
effect of dust in this connection ; for, if the experiments of Dr. Downey
and of Mr. Fruth represent a primary phenomenon, not explicable by sub-
sidiary considerations of this kind, they carry with them the very remark-
able conclusion that, of the ionization observed in a vessel at atmospheric
pressure and ordinarily attributed to a penetrating radiation, less (and
probably considerably less) than one ion per c.c. per second is to be ac-
counted for as having its origin in a direct action of the primary radiation
on the gas or in the action of a secondary radiation emitted from the gas
by the primary radiation.

EXPERIMENTS ON THE DIRECTION OF THE PENETRATING

RADIATION

An important light would be thrown upon the origin of the penetrating
radiation if it could be shown to partake of a directive character. Experi-
ments on this matter were originally made by Cook and by Wood, who
interposed screens between the apparatus and its surroundings at various
orientations. The experiments were inconclusive, but, as far as they went,
seemed to indicate that the radiation came equally from all directions.
A method of this kind is very seriously affected by lack of constancy of
the residual ionization itself during the various experiments between which
comparisons are subsequently made, and, for this reason, some experi-
ments were undertaken by Miss J. Herrick under the writer's direction
with the object of eliminating the main causes of uncertainty in the earlier
experiments of Cook and Wood. The method used by Miss Herrick de-
pends upon the fact that if gamma-rays pass through a thin sheet of metal,
the ionizing electrons emitted from the side at which the rays enter differ
as regards their number and speed from those which are ejected from the

70 TERRESTRIAL MAGNETISM AND ELECTRICITY

side at which the gamma-radiation leaves. If the penetrating radiation is
of a gamma-ray type, it should show similar characteristics. The ratio of
the subsequent ionization resulting from the incidence of gamma-rays on
a surface to that resulting from the emergence of gamma-rays from the
surface depends upon the material of the surface. The apparatus used
consisted of two similar cylinders mounted with their axes in the same
horizontal line. One semi-circular half of each cylinder was made of lead,
and the other half was made of aluminum. The cylinders were provided
with central rods which were connected to each other and to the insulated
quadrant of an electrometer. By placing potentials of plus fifty and minus
fifty volts respectively on the cylinders, the ionization currents in the gas
could be caused to feed into the electrometer in such a way as to almost
completely compensate. Several precautions to avoid leakage, and errors
due to fluctuation in the potentials of the batteries were taken, the details
of which it will be unnecessary to describe.

To fix our ideas, suppose that an excess of gamma-radiation comes
from above, and that in the case of one of the cylinders the lead half i^
uppermost. Then, as far as ionization due to the emission of electrons
from the wall of the vessel is concerned, the ionization in the vessel in
question (or rather the portion of it due to the excess of radiation com-
ing from above) will be due to the emergence radiation from the lead, and
the incidence radiation from the aluminum. An alteration of the effect
should consequently be produced by rotating the cylinder through 180
degrees, while the other cylinder is left untouched, its function being
simply to act as a compensator for the purpose of minimizing effects re-
sulting from an actual variation in the conditions during the experiment.
Without here entering into details, it will be seen that it would be possible
to obtain a comparison between the radiation coming in different direc-
tions and the total radiation entering the vessel in so far as the ionization
was due to the electrons emitted from the walls of the vessel and in so
far as one assumed the penetrating radiation to partake of the nature of a
hard gamma-radiation as regards the difference between the incidence and
emergence effects in the case of the metals used.

The first experiments performed by Miss Herrick in the laboratory of
the physics building showed marked changes on rotating one of the cylin-
ders, and by plotting a polar diagram representing the ionization due to
radiation coming from the various directions it became possible to locate
small quantities of radium in different parts of the building. The polar
diagram, moreover, showed a hump indicating an excess of radiation com-
ing from above. The apparatus was next moved to the astronomical
observatory, where there was no radioactive material, and experiments
again gave an indication of an excess radiation from above. In order to
be free from all possibility of radioactive contamination, the writer next
set the apparatus up in the attic of his house, and carried on observations
over a period of 6 weeks, the observations being taken between the same
hours each day. The attic of a dwelling house is not the most ideal situa-

TERRESTRIAL MAGNETISM AND ELECTRICITY 71

don for a quadrant electrometer, but on plotting the results for the various
experiments there was again decided evidence of an excess radiation in
the downward direction. The magnitude of the excess was such that,
when the vessel was turned in the most favorable direction, the ionization
was about 9 percent greater than the average, a result in comparatively
good agreement with the observations of Miss Herrick made on the uni-
versity campus.

These experiments are only cited as of a preliminary nature, for there
are certain sources of complication which must be removed before a cos-
mical interpretation may be made of the results. Thus, the potential-
gradient in the atmosphere will deposit active material from the atmos-
phere on the roof of the building in which experiments are made. A
simple calculation will show that the amount of such deposition may well
be enough to seriously affect experiments of the kind described above.
Similar remarks apply to the effect of radioactive material in the soil itself,
and that deposited on the surface of the soil by the atmospheric potential-
gradient. It is planned to continue the observations under conditions
which, it is hoped, will eliminate these causes of uncertainty.

THE EARTH'S PENETRATING RADIATION AND THE ORIGIN OF THE

EARTH'S CHARGE

In 1915, the writer proposed a theory of the origin of the earth's charge
based on the assumption that high speed electrons were shot into the earth
from the atmosphere as a result of a very slight radioactivity of the
atmosphere itself, or as a result of the breaking up of the emanation nor-
mally in the atmosphere. If one assumes that an electron of sufficiently
high speed can have a range as great as 5 kilometers in the atmosphere,
it is only necessary to postulate the emission of one such high speed cor-
puscle per c.c. in the downward direction, each 100 seconds in order to
account for the maintenance of the earth's charge. Or, viewed from
another standpoint, since we know that about 5 pairs of ions are formed
per C.C. per second in the atmosphere, it is only necessary to suppose that
in the case of one out of every 500 pairs of ions formed a high speed cor-
puscle of the above kind is emitted. The theory of passage of electrons
through matter is not at all inconsistent with the postulation of g^eat
ranges such as those required by the theory ; however, in 1917 Swann ^
put forward another theory in which the expulsion of the electrons from
the atoms of air is brought about by the penetrating radiation from above,
the hard nature of this radiation resulting in its emitting electrons from
the air molecules almost exclusively in the downward direction. Under
the influence of the electronic bombardment, the earth would charge up
imtil the conduction current back to the atmosphere just sufficed to balance
the effect. It appeared that if one assumed as many as 3 corpuscles to be
emitted per c.c. per second from the atmosphere, it would only be neces-
sary to postulate a range of penetration of about 9 meters in order to

' Phys. Rev., 9, 555-557, 1917.

n TERRESTRIAL MAGNETISM AND ELECTRICITY

account for the maintenance of the earth's charge. As pointed out by
the writer at the time, the chief difficulty facing a theory of this kind is the
explanation of why the swiftly moving corpuscles do not produce, in the
atmosphere, a much greater ionization than is observed. Difficulties of
this kind assume a much less formidable aspect, however, when viewed in
the light of modem views as to the properties of swiftly moving electrons.
In 1918, V. Schweidler independently put forward the second of the above
theories, and described an experiment carried out with the object of test-
ing it. The aim of this experiment was the endeavor to observe a charging
effect in a thick piece of metal as a result of corpuscles entering it, the
piece of metal being surrounded by a shield from which it was insulated.
Failure to observe any charging effect caused v. Schweidler to conclude
that the replenishment of the earth's charge could not be brought about
by a corpuscular radiation of the type discussed. As a matter of fact, the
writer had performed an experiment somewhat similar to v. Schweidler's
experiment in 1915, in connection with his earlier theory of corpuscular
chai^ng. In this experiment an earthed vessel surrounded an insulated
hollow cylinder connected to an electrometer. The rate of rise of poten-
tial was noted, and a solid copper bar was then inserted in the hollow
cylinder, the rate of rise being then again noted. By this device of per-
forming two experiments in which the surfaces exposed were the same,
surface effects were eliminated. As in v. Schweidler's experiment, no
certain charging effect was observed ; and, while this weighed against the
former of the theories above referred to, it was not felt that it formed so
weighty an argument against the latter, for on that theory it might result
that the penetrating radiation would shoot as many electrons out of the
bottom of the cylinder as it shot in at the top, except for the absorption of
the penetrating radiation itself within the cylinder. In other words, it is
the coefficient of absorption of the penetrating radiation in the cylinder
rather than the coefficient of absorption of the corpuscles which is of
in:q)ortance. In a recent paper,^ R. Seeliger discusses the origin of the
earth's charge. He considers v. Schweidler's experiment as conclusive
evidence against any theory which postulates a corpuscular replenishment
at all parts of the earth, but raises the question as to whether a corpus-
cular replenishment may not take place in certain limited areas, in polar
regions for example. It is to be observed that any asstunption of this
kind invites, in their most serious form, difficulties associated with the
ionization which might be expected to result from a passage of the cor-
puscles through the atmosphere. For a concentration of the corpuscular
current in a limited r^on would result in greatly increased ionization in
that region, and on the assumption that a corpuscle produces 50 ions per
centimeter of its path and that the coefficient of recombination of ions is
1.6X10^, it can readily be shown that, unless the area of precipitation
were more than one thousandth of the area of the earth, the conduc-
tivity produced in the air in the region of precipitation would be so great
that, for a potential gradient of 150 volts per meter, one would calculate

^AnnaUn drr Physik, 63, 464-481.

TERRESTRIAL MAGNETISM AND ELECTRICITY 73

for this region alone a total conduction current greater than the corpuscu-
lar current. In other words, there is a lower limit to the value which one
may assume for the region'of precipitation. Thus the avoidance of diffi-
culties concerned with the failure to directly measure a corpuscular cur-
rent, by relegating that current to regions where experiments have not
been made, does not avoid what is perhaps one of the most serious diffi-
culties confronting any corpuscular theory, that of reconciling the com-
paratively small ionization of the atmosphere with the passage through
it of about 1500 high speed corpuscles per square centimeter per second.

THE CONDUCTIVITY OF THE UPPER ATMOSPHERE

The paper concluded with a reference to the importance of a knowledge
of the conductivity of the upper atmosphere in relation to the origin of
the earth's charge and allied phenomena, and the author described an ex-
periment in progress at the University of Minnesota designed with the
object of measuring the distance of the supposed conducting layer by
measuring the time taken by wireless waves to reach that layer and return.

RECENT RESULTS DERIVED FROM THE DIURNAL-VARIA-
TION OBSERVATIONS OF THE ATMOSPHERIC-
ELECTRIC POTENTIAL-GRADIENT ON
BOARD THE CARNEGIE "^

By S. J. Mauchly

The Department of Terrestrial Magnetism, in accordance with its direc-
tor's plans, has for many years been making not only magnetic but also
atmospheric-electric observations aboard its survey vessel, the Carnegie,
It is thus contributing the chief data for mapping both the earth's mag-
netic field and its electric field. Furthermore, since 1915 numerous obser-
vations have been made aboard the Carnegie to determine the nature and
magnitude of the changes in the electric condition of the atmosphere which
take place during a 24-hour cycle.

For the potential^gradient the general procedure in the diurnal-variation
observations is to make a set of 20 observations during each of 24 consecu-
tive hours. The observations for such a set require about 20 minutes
and their mean value is referred to the mean time for the set. From
deductions .based on the observations made prior to April, 1916, it ap-
peared that the diurnal variation of the potential-gradient over the oceans
probably did not differ much from that which has been found at many
land stations ; that is, they indicated two rather pronounced maxima and
two minima during a 24-hour period.' However, very few data were

* Preliminary report presented before the American Geophysical Union, with
amplifications.

""Researches of the Department of Terrestrial Magnetism," Vol. Ill, pp. 416-420^
Washington (1917).

74

TERRESTRIAL MAGKETISM AND ELECTRICITY

available from oceans other than the Pacific, and as pointed out in the
report just cited, a large percentage was derived from series of observa-
tions which were terminated by the advent of unfavorable weather. It
should also be noted in passing that Swann ^ a year later in discussing

Fig. 3. Diurnal variation of electric potential-gradient on the oceans,

plotted according to Local Mean Time.

the results of the observations for the year ending February 20, 1917,
states that **the effect of the 12-hour Fourier wave is less important in
the present curves than in those already published."

The largely increased amount of material which has accumulated since
1915 makes it now possible to reject nearly all data corresponding to less
than a 24-hour series and still have 45 practically complete 24-hour series
available. The data for each series, therefore, correspond to an actually
occurring sequence of phenomena, and the mean results are free from the
errors which would result from combining the results of partial series of
observations.

Of the 45 diurnal-variation series referred to, 30 were made in the
Pacific, 5 in the Atlantic, and 10 in the Indian Ocean ; the combined data
represent about half the earth's surface. The means corresponding to

*W. F. G. Swann. "Supplementary report on atmospheric-electric observations
made aboard the Carnegie from May 17, 1916, to March 2, 1917," in "Annual Report
of the Director of the Department of Terrestrial Magnetism" for the year 1917.
Year Book of the Carnegie Institution of Washington, 1917, p. 282.

TERRESTRIAL MAGNETISM AND ELECTRICITY

75

the separate oceans, as derived from 39 series, are represented in figure 3.
They show : ( 1 ) That the mean diurnal-variation curves for the Pacific,
Atlantic, and Indian oceans are similar in form; (2) that the principal
component of the variation consists of a 24-hour wave, and (3) that the
times of occurrence of the chief phases of this wave, when referred to

CQivii)

Pig. 4. Diurnal variation of electric potential-gradient on the oceans,
plotted according to Greenwich Mean Time.

local time, diflfer for the several oceans by amounts which correspond
approximately to the differences between the respective mean longitudes,
for the several oceans, of all the points at which observations were made.
Since the curves of figure 3 suggest the simultaneous occurrence of
maximum (or of minimum) phase over all three oceans, it was decided

76 TERRESTRIAL MAGNETISM AND ELECTRICITY

to refer the results of each series of observations to Greenwich Mean
Time (civil), and recompute the means for the separate oceans on this
basis. The results are shown in figure 4, together with a curve which
includes the data from 6 recent series received from the vessel after the
curves in figure 3 had been prepared. The differences between the several
curves of figure 4 are of course not to be thought of as representative of
separate characteristics, since the smoothness of the respective curves is
seen to be closely related to the number of component series.

The curves of figure 4 show a decided similarity to land results for
high latitudes and also to many of the winter curves obtained in ordinary
latitudes. Indeed, if differences in local mean time are taken into account,
it appears that for many land stations at which the single diurnal wave
predominates, there is approximate simultaneity as to the time of occur-
rence of maximum (likewise, of minimum), and this at a time which is
in general agreement with what is indicated by the curves of figure 4.
For the summer, however, as is well known, most land stations show, in
addition to the 24-hour wave, a decided secondary wave which seems to
occur in general at about the same local mean time at different stations.

The minimum value of the potential-gradient, according to figure 4,
occurs at about 4^ A.M., G.M.T.,.and in view of the fact that for observa-
tories in western and central Europe the difference between local and
Greenwich time is not great, this may account for the fact that various
authorities have assumed the occurrence of the principal minimum at
about 4^ A.M., local time, to be a rather general characteristic for most
stations. It is also significant to note that Mache and v. Schweidler ^ long
ago p<Mnted out that the phase angle of the 24-hour wave varied greatly
from station to station while the phase angle of the 12-hour wave was
approximately the same for nearly all stations. Although the phase angles
of the 24-hour Fourier waves for the European stations show among
themselves very much greater differences than can be accounted for by
the rather small differences in longitude, it must be borne in mind that
the results of harmonic analyses are dependent upon local meteorolc^cal
and cultural, and sometimes topographical and instrumental, factors as
well as upon any general characteristics which the potential-gradient may
possess.

In the present investigation no account has been taken of possible
changes in the characteristics of the diurnal variation with latitude and
with time of year, except to ascertain that the preponderance of the
24-hour wave and the approximate progress on a universal-time basis
seem to hold throughout the year and for wide ranges of latitude. The
present results are, therefore, to be considered as provisional and repre-
senting only a general yearly average. In fact, investigations under way
show that considerable modification in detail is to be expected as more
observational material becomes available. The data from 45 practically

^ H. Mache tind E. v. Schweidler, "Die Atmosphirische Elektrizitat," p. 27. Braun-
schweig, 1909.

TERRESTRIAL MAGNETISM AND ELECTRICITY 77

complete series of diurnal-variation observations aboard the Carnegie,
representing a general distribution over most of the accessible ocean-areas
indicate* therefore, as a preliminary result, that the chief component of
the diurnal variation of the potential-gradient over the major portion of
the earth {especially the oceans) is a wave of 24-hour period which occurs
approximately simultaneously in all localities.

A fact of considerable interest is that the diurnal- variation curves for the
potential^adient derived from the Carnegie observations are very similar
to curves which represent the relative frequencies of the Aurora Borealis,
as observed at several European stations, and also to curves representing
the diurnal distribution of certain classes of magnetic disturbances, when
all are referred to the same time-basis. It may also be pointed out that
owing to the non-coincidence of the earth's magnetic axis with its axis of
rotation, the time of daily potential-gradient maximum, as indicated by
the ocean curves, corresponds approximately to the time when the earth's
north magnetic pole, for example, is nearest to the sun, while the daily
minimum occurs, in a general way, when this pole is farthest from the
sun. The actual times of maximum and minimum, however, appear to
depend upon the positions of both magnetic poles and the fact that their
longitude difference is not 180°. These correlations appear to support the
assumptions of various investigators that the earth's electric charge and
resultant field may be very intimately related to an electric radiation from
the sun. The best evidence as to the extent of this support will probably
result from a study of the details of the diurnal-variation curves' corre-
sponding to various positions of the earth in its orbit. Reductions with
this end in view are under way and it is hoped that sufficient data will
soon be available to yield some information on this point.

The making of diurnal-variation observations in atmospheric electricity
by eye readings is always a burdensome procedure; the carr3dng on of
such work aboard a vessel is not only arduous but also difficult. In this
connection the utmost credit is due the several commanders of the Car-
negie, during her various cruises, and to all the observers who participated
in the observational work.

I am indebted to the director, Dr. L. A. Bauer, for his constant interest
in and encouragement of the work in hand, and for a suggestion of the
possibility of finding in the asymmetry of the earth's magnetic field an
explanation of the observed diurnal variation on a universal-time basis.
I am also greatly under obligations to the members of the Department of
Terrestrial Magnetism who assisted in the reduction of the observational
data, especially to Dr. G. R. Wait, both for valuable assistance and helpful
suggestions.

The full publication of the observational data and discussion of results
will be deferred until after the completion of the present cruise of the
Carnegie. Department of Terrestrial Magnetism,

Carnegie Institution of Washington.

SUGGESTIONS RELATIVE TO THE APPLICATION OF MATH-
EMATICAL METHODS TO CERTAIN BASIC PROB-

LEMS OF DYNAMIC OCEANOGRAPHY

By G. F. McEwbn

Investigations of the ocean have generally been carried on by geogra-
phers and geologists, oftentimes incidentally to those of other divisions
of these extensive fields of science. Accordingly, qualitative methods so
characteristic of geography and geology have been widely used in oceano-
graphic investigations. Such qualitative methods and the empirical treat-
ment of quantitative field observations have been very suggestive, have
stimulated interest, and led to certain broad generadizations that are
essentially correct. However, there has been a tendency toward rather
loose reasoning and lack of consideration of established quantitative prin-
ciples of physics, which has resulted in certain erroneous conclusions.

Must we admit that the complexity of such geophysical phenomena
renders careful reasoning and quantitative treatment impossible of attain-
ment ? Probably many would at first answer yes, but let us consider the
matter further before expressing an opinion. Within the past fifty years
a few scientists have undertaken, by means of a definite formulation of
specific problems, to apply mathematics to ocean data, and thus to con-
tribute to a system of demonstrable principles applicable, in general, to
all similar cases ; and attention has been increasingly directed to this type
of research. Important advances have thus been made, and serious errors
in certain former conclusions have been discovered, although, especially
in some of the earlier attempts at mathematical applications, significant
errors arose from incorrect assumptions and failure to appreciate impor-
tant attributes of such "field," or natural problems. At first, men accus-
tomed to the problems of laboratory physics attempted to deduce physical
laws of the sea from results of laboratory studies, and certain precon-
ceived assumptions regarding oceanic conditions. They also worked
under the disadvantage of having very inadequate data. As more accu-
rate and exhaustive field data accumulated, attention was directed more
to interpreting and coordinating field observations rather than to depend-
ing on the speculative and unsound method of imposing on the sea purely
theoretical laws deduced from laboratory researches.

Mohn's pioneer investigation of 1887,^ based on the deduction of the
changes in form of the surface that would give rise to currents actually
produced by winds, variation in barometric pressure, and specific gravity,
was a great advance beyond earlier attempts at a precise treatment of
ocean data, and doubtless contributed greatly to the development of the
more satisfactory methods of today. Among the later results thus worked

*Mohn, H. 1887. The Norwegian North Atlantic expedition, 1876-1878: The
North Sea, its depths, temperature and circulation. (Christiania, Grondahl), 212
pp., 48 pis.

78

PHYSICAL OCEANOGRAPHY 79

out are Ekman's * hydrodynamical theory of oceanic circulation, which
pertains especially to wind-driven currents, and was undertaken at Nan-
sen's ^ suggestion ; and B jerknes's * convection theory which pertains
especially to the determination of ocean currents due to differences in
specify gravity. Later his pupil, Sandstrom,* devised a much more rapid
and accurate method of computing such currents.

The Norwegian investigator, Jacobsen,** in certain more recent quanti-
tative investigations pertaining to the Atlantic near Denmark, obtained
encouraging results by giving special attention to the alternating con-
vective motion of small masses of the water, or to the "mixing phenom-
ena," as he called it. His researches afford strong evidence in support of
the idea suggested by earlier qualitative studies, that in lakes and oceans,
very small or elementary masses of the water are moving at random in a
manner somewhat analogous to the motion of molecules in a gas, except
that the direction of motion in large bodies of water is mainly vertical,
although the resultant vertical flow may ze zero. This phenomenon of the
interchange of small masses of water has been variously referred to as an
alternating convective circulation, mixing phenomenon, eddy or vortex
motion, and turbulence. Jacobsen's and other recent investigations in this
field indicate that this phenomenon is the cause of the processes of diffu-
sion, heat conductivity, and f rictional resistance peculiar to oceanic condi-
tions. Comparable values of the **Mischungsintensitat," a coefficient indi-
cating the intensity of the rate of interchange of small water masses, have
been deduced independently from the distribution of temperature and
salinity, and also from the dynamical treatment of current observations.
Thus studies of temperature and salinity distributions may yield appro-
priate values of the f rictional resistance, an essential factor in the dynami-
cal solution of ocean-current problems as well as certain tidal problems.
This f rictional resistance about which there is so little definite information
appears to vary widely with the locality, wind velocity and other factors.
It is not a "physical constant" of the substance, water.

The precise nature of this mixing motion can not be directly determined,
but various reasonable assumptions regarding it can be made, and com-
bined with known fundamental facts into a quantitative theory or general-
ization, from which deductions can be made, and tested by comparison
with observations. Encouraging results already reached appear to justify

'Ekman, V. W. 1905-06. On the influence of the earth's rotation on ocean cur-
rents. Arkiv for Matematik, Astronomi och Fysik, 2, 1-53, 1 pi. and 10 figs.

'Nansen, Fridtjof. 1902. The Norwegian North Polar Expedition, 1893-1896.

Scientific Results, Vol. Ill, Longmans Green & Co. London, part IX, pages 1-427.

33 pis.
' Bjerknes, V. F. K. 1901. Circulation relative zu der Erde. Ofversikt af Kongl.

Vet.-Akad. ForhandL. 58, 739-757.
*Krumniel, O. 1911. Handbuch der Ozcanogriphic (Stuttgart, Engelhorn), 2,

xvi, 766 pp., 182 figs, in text. .

■Jacobsen, J. r. 1913. Beitrag zur Hydrographie der Damschen Gewasser.
Medd. Komm. Havandersogelser (Hydrografi), 1, no. 94, pp., 14 pis., 17 figs,
in text.

80 PHYSICAL OCEAMOGRAPHY

further efforts in this direction and point to the possibility of a satisfac-
tory coordination of the various phenomena of conduction, diffusion, and
fluid friction by means of a single mathematical theory of the mixing
motion. Such an investigation, if successful, would enable one to deduce
the temperature distribution in a body of water gaining heat from solar
radiation of given intensity, and losing heat by evaporation. Investiga-
tions of this simplest case would thus correlate under definite physical
principles all of these various thermal phenomena. Again, by so amend-
ing such results as to include the effect of a given flow or current on the
distribution of temperature determined for the above simplest case, esti-
mates of a current could be made from the difference between the undis-
turbed and the actual temperature distribution. This has been partly
worked out and applied to the determination of the velocity of upwelling
in the San Diego region. Thus the temperature disturbance can be
quantitatively treated as an effect of a current, without regard to its
dynamical causes. Qualitative ccmclusions relative to ocean currents have
long ^o been reached from essentially the same general idea, and this
fact points to the possibility of such a quantitative theory. Such general
quantitative laws of the relation between currents and temperature de-
partures from the undisturbed state might be combined with Bjerknes's
dynamical theory, and thus afford a means of deducing answers to more
involved questions, such as the following : Given the distribution of solar
radiation over the surface of a body of water having given boundaries,
and a known initial temperature distribution, to determine the resulting
temperature distribution and circulation for any later time. The results
of similar determining conditions have not infrequently been either
assumed or surmised in order to form a basis for more far-reaching
oceanographic conclusions. But it is by the precise formulation and suffi-
ciently accurate solution of suitable specific problems sufficiently in accord
with actual conditions that general laws of oceanic phenomena can be
discovered and tested. And the greater the variety of such ideal problems
that we are in position to attack, the greater will be our progress in the
precise and detailed study of the physics of the ocean. Also, it is to be
expected that a satisfactory physical theory, especially of the mixing
phenomenon, would greatly aid in the solution of certain problems of
sedimentation and ocean chemistry.

One problem of ocean physics, whose simplest special case would be to
deduce the vertical temperature distribution in a body of water exposed
to solar radiation of approximately uniform intensity over the surface
and losing heat by evaporation and conduction from its surface, has
received very little attention except of a qualitative or speculative nature.
Yet this problem appears to be fundamental in precise oceanographic
investigations. Accordingly, the author has for some time attempted to
work out a solution, and after trying and rejecting various assumptions,
has reached encouraging preliminary results by using certain concepts
from statistical mechanics, combined with elementary laws of heat and
radiation. It is hoped that these studies will have progressed far enough
for publication within a year or two. It is also the intention, after this

PHYSICAL OCEANOGRAPHY 81

work on temperatures is in a more finished state, to investigate the prob-
lem of diffusion in the sea by similar methods applied first to certain of
the numerous salinity determinations made by the Scripps Institution.
Problems of the t3rpe mentioned in this paper form an extensive and
promising field of fundamental importance in oceanography, and demand
the attention of all investigators interested in promoting quantitative
studies of the sea, but probably only a few will desire to engage actively
in their solution.

It has formerly been necessary to make a great deal of use of such
scattered data as the investigator could find as a basis for theoretical
work. Much has been, and doubtless will be, accomplished in that way.
And all original detailed data, as well as summaries and deductions there-
from, should be accessible in some way to investigators, even if publication
is not always practicable. But such a procedure has obvious disadvan-
tages, such as insufficient or unknown precision, incomplete data, or lack
of significant factors that may impair or greatly restrict the conclusions.
Therefore it is also necessary to conduct special programs of observations,
designed with reference to particular problems, in order to improve and
supplement the above more extensive and preliminary type of work. Thus
selection of the locality, season, etc., and the observation of all relevant
phenomena affords as nearly as possible a realization of the advantages
of the physicist who controls the conditions affecting his laboratory ex-
periments. For example, serial temperatures observed in the central part
of a high-pressure area away from currents or land, and where prevailing
great depths provide results corresponding to the simplest conditions,
would be of great aid in the study of ''normal" temperature gradients.
Such observations should also be accompanied by observations on the
intensity of solar radiation, turbidity, salinity, and evaporation, and should
be continued through different seasons and times of day in order to pro-
vide the most important kinds of data. Very little of this intensive type
of work, carried on with sufficient continuity and completeness, has been
done, and it has been restricted to certain portions of small inland seas or
inshore regions. Although results thus obtained are valuable in them-
selves and as a means of interpreting such fragmentary and widespread
data as may be available, they can not take the place of similar intensive
work at selected stations throughout the ocean. Actual conditions in
typical areas of the great ocean must be carefully observed and studied,
if any reasonable approach to exhaustive oceanographic investigations is
to be realized.

In this paper I have dwelt especially on the deductive treatment of
ocean problems, because of the great need of improvement in this aspect
of the subject. Although admitting that purely empirical or statistical
methods are indispensable in assembling and coordinating various kinds
of field data, it seemed desirable to urge the need of progress from such
empirical results toward the goal of a complete deductive treatment
carried out in accordance with known generalizations of physics.

82 PHYSICAL OCEANOGRAPHY

STATE OF PROGRESS IN CONTINUOUS RECORDING
OCEANOGRAPHICAL INSTRUMENTS

By Albixt L. Thubas

The modern tendency in physical research is to replace indicating instru-
ments by recording instruments wherever possible. This has bmi espe-
cially true in the science of meteorology where the recent advances have
been brought about almost entirely by the remarkable improvements and
developments in recording instruments. In the related science of physical
oceanography there are practically no recording instruments now in gen-
eral use, except possibly the tide-gage. If meteorology has been so greatly
benefited by such instruments, surely in oceanography, where the changes
in the physical properties are so much more regular and therefore more
easily interpreted, great advances should be looked for through the addi-
tion or substitution of recording instruments.

Heretofore the methods of collecting physical data have been such that
no complete knowledge of the physical characteristics of the particular
body of water under investigation have been obtainable as the work is
progressing. The procedure has been to lay out stations, as intelligently
as possible along courses throughout the region of the ocean to be studied
which will give the most important information. At these various sta-
tions with the use of water bottles and reversing thermometers samples
and temperatures of the ocean water are obtained at various depths down
as far as the investigations are carried. The thermometers are read as
soon as the water bottles are drawn up and samples of the water are
stored in bottles which are later chemically measured for salt content in a
laboratory on shore. The several disadvantages of this method are ap-
parent: (1) No working knowledge of the ocean water is immediately
obtainable and consequently no rearrangement or addition of stations can
be made from an examination of the data taken. This is very important
especially where our knowledge of the ocean is limited and one wishes to
explore the magnitude and extent of the surface and submarine currents.

(2) It is impossible to obtain a corroboration of any data where there
may be doubt as to the accuracy or reliability of single observations.

(3) The data taken are usually inadequate and especially so at those sta-
tions where vertical lines of observations pass through various strata of
water of different salinity, temperature and density. Curves and cross
sections plotted from data taken in these regions are usually a matter of
approximation and give very little information as to the mechanism of the
mixing of waters of widely different properties. These observations are
particularly inadequate in such regions as the southern end of the Grand
Banks of Newfoundland where the cold waters of the Labrador Current
merge into the warm saline waters of the Gulf Stream.

With the object of improving the technique of the science of physical
oceanography, an effort has been made in recent years to develop practical

PHYSICAL OCEANOGRAPHY 83

recording instruments which are sufficiently rugged and simple to be used
on shipboard. The most important physical properties of the sea of
which a continuous record should be made, are temperature, salinity, den-
sity, velocity and direction. The first three properties are so related that
any one can easily be deduced from a measurement of the other two. The
properties most easily measured are temperature and salinity. Salinity
is defined as the number of grams of salts in a liter of sea water. From
a c(Hisideration of the properties of sea water that vary with the salinity,
the electrical conductivity seems to be the most susceptible to continuous
measurement if the difficulty due to the variation of conductivity with
temperature can be overcome. Such a method consists in measuring the
ratio of the resistance of sea water in two equal or nearly equal electro-
lytic cells, one cell containing sea water of a known salinity and the other
having the sea water to be measured flowing through it. The ratio is
obtained by a Wheatstone bridge, using alternating current to eliminate
polarization effects in the cells. A record of the resistance ratios of the
two cells is made by an automatic electrical recorder. By immersing the
two cells in the same temperature bath almost complete compensation of
temperature changes is effected.

A continuous record of the temperature of the ocean is most easily
obtained with a platinum resistance thermometer and an automatic regu-
lating Wheatstone bridge quite similar to the continuous salinity recorder.
This instrument has been used successfully on shipboard for several years
in the region of the Grand Banks of Newfoundland and some very inter-
esting records have been obtained which show the distribution of tem-
perature and thereby indicate the location j)f ocean currents and also give
a knowledge of their boundary conditions which could hardly be obtained
by repeated single measurements of temperature.

The continuous recording instruments for temperature and salinity de-
scribed above give only surface measurements but they could easily be
constructed for making measurements below the surface. This would
require the use of an insulated cable of 4 or 5 conductors which would
be sufficiently strong and flexible. During 1918-19 in connection with
submarine listening experiments there were constructed reinforced cables
similar to these, which could be repeatedly wound on to and off of a drum
and would withstand a weight of 400 to 500 pounds. Data from these
instruments gave accurate curves of temperature, salinity and density
from the surface down to a depth of probably 500 meters, which is the
most interesting part of the ocean dynamically.

R. A. Daly of Harvard University has developed and built a thermo-
graph which can be anchored in very deep water and will give a record of
temperature for a period of several days. The instrument has an inter-
mittent mechanism which gives periodic photographs of a mercury col-
umn. This instrument was specially designed to withstand very high pres-
sures and it should be especially valuable in studying the small variations
of temperature at great depths in the ocean.

84 PHYSICAL OCEANOGRAPHY

The measurement of the movement of waters in the ocean has been
quite difficult to perform experimentally. This difficulty has been due
chiefly to the non-continuity of measurements and the unknown move-
ments of the vessel from which the measurements are made. Dr. Hans
Pettersson of Goteborg, Sweden, has solved this problem by his photo-
graphic current meter which will give a continuous record of both direc-
tion and velocity for a period of two weeks. This instrument with the
use of special anchors and buoys, can be firmly secured at any depth down
to 300 meters. The difficult problem which had to be solved in this instru-
ment was the transfer of the motions of a propeller through a water tight
case containing the recording apparatus without the addition of friction.
This was accomplished by a magnetic coupling.

Dr. Pettersson has also developed densimeters to be used from shore
stations which give a record of the movements of the waters of various
salinities into and out of the Swedish Fjords. These instruments consist
of vessels or cans whose density is equal to the average density of the
Fjord waters. As the submarine waves of high salinity come in from
the ocean these vessels will rise and a record of their height is automati-
cally recorded. Some interesting theories of submarine waves have re-
sulted from this work and the correlation between the variations of salinity
and the abundance of fish in these Fjords is being studied.

This briefly describes the recent developments in recording instruments
and in conclusion I wish to suggest the possible application of these in-
struments to future research in physical oceanography.

A comparison of the yearly observations in the region of the Grand
Banks of Newfoundland shows that the volume and strength of the
Labrador Current have a decided influence on the course of the Gulf
Stream in that vicinity. In some years the Gulf Stream was found almost
up to the southern end of the Grand Banks and in other years as far
south as the 40th degree of north latitude, a variation of over 100 miles.
An accurate knowledge of the volume, velocity and location of these cur-
rents from time to time and correlation with meteorological conditions
might yield results of great interest.

This information could be obtained by the use of recording instru-
ments in the straits of Florida and across the Gulf Stream before it
branches out east of the Grand Banks of Newfoundland.

With continuous salinity and temperature recorders placed on trans-
Atlantic vessels a complete record of the variation of temperature salinity
and density could be secured across the Atlantic from month to month.
These instruments would make measurements at a constant depth below
the surface and might throw considerable light on the hydrodynamics of
this part of the Atlantic.

It seems to me that the science of physical oceanography has passed the
period of exploration and has now reached that stage in its development
which calls for a program of research on a large scale with most carefully
thought out plans of systematic investigation extended over long periods

PHYSICAL OCEANOGRAPHY 85

of time. Results from such an undertaking I believe can be most suc-
cessfully accomplished by the use of recording instruments.

Western Electric Company,

New York City.

PRESENT STATUS OF RESEARCHES ON MARINE

SEDIMENTS IN THE UNITED STATES f

By Thomas Wayland Vaughan * ">

INTRODUCTION

The ocean of today stands at the end of a succession of oceans that
* have existed since land and water were first divided from each other on
the earth's surface. This fact, admitted by everyone, needs to be empha-
sized in order to make clear the transcendent importance of the study of
marine sediments. It is possible to measure the depth and the tempera-
ture of the waters of the present ocean, to sample its waters from the sur-
face to the bottom of its greatest abysses and examine them chemically,
and to measure directly or to infer from measurable factors its currents.
It is also possible to study the sediments deposited on the floor of the ocean
and around its margins. These and other features of the present ocean
can be known by direct processes but over a large part of the earth's sur-
face where there was once sea there is now only land, and the depth, tem-
perature, chemical composition, and currents of bodies of waters no longer
existent cannot be measured. That seas once extended over regions now
land is known through the record made by the sediments and these sedi-
ments supply the fundamental data for recognizing the physical features
of the vanished oceans.

Considerable information has already been acquired on modern marine
deposits and preliminary maps of parts of the ocean floor have been made.
Among the sources of this information are the studies of Bailey and
Pourtales, the classic work of Murray and Renard, Murray, and Murray
and Lee, Murray and Philippi, and Philippi, the many papers by Thoulet,
several papers by Boggild, including his recently published "Meeresgrund-
proben dcr Siboga-Expedition," papery by Walther, and the studies of the
shoal-water deposits of Florida and the Bahamas and Murray Island,
Australia, with which I have been associated.' Of course there are many
other authors but I have given the names of those who have done most in
areally mapping deposits on the bottom of the sea. The great leaders are
Murray and his associates, among whom Philippi is to be reckoned, and
Thoulet. It is believed that the characteristics of some deposits and the
relations of these deposits to the conditions under which they formed have

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

*K. Andr^e is the author of a useful bibliography on literature on marine sedi-
ments published between 1841 and 1911. See his article, Uber Sedimentbildung am
MeeresDOden, Literaturzeichniss : Geolog. Rundschau, 3, 1912, 524-338.

86 PHYSICAL OCEAXOGRAPHY

been ascertained with enough accuracy to admit their use in interpreting
geological history ; but how inadequately some relations are understood is
exemplified by the presence of red clay at comparatively shallow depths,
4000 meters, in the enclosed deep basins of the East Indian Archipelago.
Boggild says it is necessary to conclude that the capacity of the water to
dissolve calcium carbonate is greater in the enclosed basins of the East
Indian Archipelago th^n in the open ocean.^ Although considerable is
known about marine sediments, the information is far below what is
needed to understand many important features of sediments in the modem
oceans and to supply a basis for interpreting ancient sediments.

RESEARCHES ON MARINE SEDIMENTS IN AMERICA

There is under the Division of Geology and Geography of the National
Research Council of the United States a Committee on Sedimentation
composed of fourteen members, of which I have the honor to be chair-
man. This committee is divided into seven subcommittees, as follows:
universities and colleges east of the Allegheny Front; universities and
colleges between the Allegheny Front and the Rocky Mountains ; univer-
sities and colleges west of the Rocky Mountains ; state geological surveys ;
chemical and physical researches on sediments ; field description of sedi-
ments ; preparation of a treatise on sedimentation. The report of the com-
mittee for the year ending on April 28 has been transmitted to the chair-
man of the Division of Geology and Geography and is available in
mimeographed form to interested persons.

No attempt will be made to give an account of the work of the com-
mittee, as that would consume too much time, but it will be said that its
scope includes both modern and ancient sediments and both continental
and marine deposits. One of the purposes of the committee is to ascer-
tain and to follow all current investigations on sediments within the United
States and the machinery for accomplishing this purpose is good.

The U. S. Bureau of Fisheries is trying to arrange for a study of the
sediments of the Bay of Maine but the plans have not yet been completed ;
and an attempt is being made to have the sediments of Chesapeake Bay
studied cooperatively by the Bureau of Fisheries, the U. S. Geological
Survey, and Johns Hopkins University, but the actual work on the bot-
tom samples has not begun. Prof. G. D. Louderback has been trying to
study the bottom deposits collected principally by the Bureau of Fisheries
in San Francisco Bay. Some preliminary information on the samples has
been published but the investigation has progressed slowly. Two re-
searches with which I have been concerned have had grave difficulties.
One of them, the study of the sediments off the mouth of Mississippi
River, as representing an area in which great quantities of terrigenous
material are being deposited, has come to a standstill with the resignation
of Mr. E. W. Shaw from the U. S. Geological Survey. The other study

* Boggild, O. B., Mceresgrundproben der Siboga Expedition : Siboga Expcditic,
Mon. 45, p. 11, 1916.

PHYSICAL OCEANOGRAPHY 87

is on the shoal-water deposits of southern Florida and the Bahamas, as
representing areas in which very little or no terrigenous material is being
deposited, except at the north end of the Florida reef. Fortunately several
papers on the Floridian and Bahamian samples themselves and on cor-
related phenomena, such as papers by Dole and Qiambers and Wells on
the chemistry of the waters, bacteriological studies by Drew and Keller-
man, and temperature records by me, have been published, but a large
quantity of data remains unpublished. I am hopeful that within a rela-
tively few months all data already acquired, which include Ekman current
meter measurements at about 15 stations, may be prepared for printing.

The researches of F. W. Clarke and W. C. Wheeler on the inorganic
constituents of the skeletons ol marine organisms is of prime importance
but such work needs to be correlated with studies on the bottom deposits
themselves. The research on which Clarke is now engaged, the composi-
tion of river water discharged into the sea, is also of great value. Wells's
researches, such as his published "New determinations of carbon dioxide
in water of the Gulf of Mexico" and the studies he is now making on the
waters of Chesapeake Bay, are also of much value in understanding prob-
lems of sedimentation, but the sediments themselves need to be studied.
Richard Field of Brown University is studying some features of modem
shoal- water limestones, and E. M. Kindle of the Canadian Geological Sur-
vey is conducting important researches on modern limestones ; but Kindle
may not be credited to the United States.

Of the researches above enumerated, five deal with bottom samples and
areal surveys of the sea bottom. The areas are the Bay of Maine, Chesa-
peake Bay, southern Florida and the Bahamas, the mouth of Mississippi
River, and San Francisco Bay. No one of these researches is progressing
in a satisfactory way. Furthermore, all these researches deal primarily
with shoal-water deposits — ^there is no work on deep-sea deposits. The
only modern deep-sea samples recently described from America are two
I described in 1917 from the Tongue of the Ocean, Bahamas. This, it
seems to me, is a very poor showing for the United States.

FACTORS THAT RETARD RESEARCHES ON SEDIMENTS

During the period that the United States were participants in the
World War, investigations on sediments suffered as did many other kinds
of scientific work and our country has not yet finished its readjustment
after the conflict. Besides the interruption caused by the war, several
competent investigators have been diverted by other duties and a new
crop of investigators has not yet ripened.

The interruption of investigations and the diversion of investigators are
not the only difficulties in the way of studies of sediments. The subject
is one that does not belong exclusively in any one of the sciences as the
sciences are currently classified, although those engaged in several kinds
of scientific endeavor recognize the value of knowledge of certain aspects
of it in the proper performance of some of their work. The engineer,

88 PHYSICAL OCEANOGRAPHY

for instance, wishes to understand shore-drift in certain places and the
rate of the deposition of sediment in harbors; the student of fisheries
wishes to know the relations between bottom materiafand organisms that
may be used as food ; the navigator may keep his course through fog and
snow by detailed knowledge of the bottom ; and the geologist may utilize
knowledge of sediments in interpreting some geological formation of eco-
nomic significance. Of the different kinds of scientific men the geologist
is the most broadly concerned, because only by an adequate knowledge of
the modern can he understand the ancient deposits and it is part of his
work to study the mechanical features and the constituents of sediments,
both modem and ancient, though he usually feels that his attention should
be directed to past rather than to present history. This is a practical day
and students inquire how they can make work on sediments pay. It has
been possible to provide funds for some work on sediments but the re-
muneration is far below that offered by oil companies.

MEANS FOR PROSECUTING RESEARCHES ON SEDIMENTS

In remarks already made I have tried to bring to your attention the
present status of researches on marine sediments in this country and I
have indicated some factors that I believe retard such investigations. How
can the backward condition of researches in this important subject be
remedied? I will venture a few suggestions.

My first suggestion is that those interested endeavor to impress upon
students the scientific importance of investigations on sediments. This
may best be done by the establishment of courses in sedimentation in our
universities and the offering of fellowships to graduate students for inves-
tigations in the subject. At present courses in sedimentation are given at
the universities of Wisconsin, Iowa, and California, and at the University
of Iowa a research fellowship is maintained. Courses should be given at
more universities and there should be more research fellowships. The
Geophysical Union might combine with the divisions of Geology and
Geography and of Biology and Agriculture and endeavor to establish two
or three more fellowships in sedimentation.

In addition to university work of the kind indicated an institution or
institutions in which complicated special studies may be undertaken are
needed. An institution comparable to the Geophysical Laboratory of the
Carnegie Institution would fulfil the g^eat need but an endowment that
will yield an income between $50,000 and $75,000 per year is not easily
obtained. However, it is worth striving for. As such an institution does
not exist it may be preferable to try to utilize existing institutions by ap-
pealing to them and trying to strengthen them. The study of sediments
is a fundamental of geology and the U. S. Geological Survey has recog-
nized this and has tried to develop researches on sediments as a part of its
wprk. Furthermore, many geologists, because of their training, are pre-
pared to undertake such investigations. It is, therefore, suggested that
those interested in such work make their desires known to the director of

PHYSICAL OCEANOGRAPHY 89

the U. S. Geological Survey, that it be pointed out to him how the Geo-
logical Survey by doing such work can help science and serve other gov-
ernmental bureaus, and that he be requested to do as much as the circum-
stances of the Survey will permit. The Geological Survey has already
done enough to place students of sediments under deep obligations to it.
If it could study and prepare reports on bottom specimens one of the
present difficulties in the way of advance in knowledge of marine bottom
deposits would be removed.

Until now it has been possible to obtain larger collections of properly
taken bottom samples than it has been possible to study. The U. S.
Bureau of Fisheries is fully equipped to collect samples precisely as they
should be taken and the heads of that Bureau are anxious to do all they
can to aid researches on sediments. Perhaps if provisions could be made
to study the samples, the U. S. Coast and Geodetic Survey might make
systematic collections. Other than governmental agencies, especially the
Department of Marine Biology of the Carnegie Institution, have shown
willingness to help in procuring bottom samples for study. The material
available for investigation is large in quantity and much of it has been
properly collected and is accompanied by all needed data. If these col-
lections could be properly studied and reports on them published, what
fine contributions would be made to our knowledge of marine sediments !

CONCLUSION

In conclusion I wish to emphasize the value to science of a proper
understanding of the marine sediments in the ocean of today. A proper
understanding of these sediments includes knowledge of the depth, tem-
perature, and salinity of the waters above them, the distance from shore
to where they were deposited, their relations to currents, and if near land,
the relief of the land, its climate, and the rocks composing it. Through
such knowledge of modern sediments the criteria for interpreting the
sediments of ancient seas are discovered. Having established the needed
criteria, the boundaries of the old seas may be traced ; the physiography,
constitution, and climate of the neighboring lands may be recognized, and
the depth, temperature, chemical composition, and currents of the waters
of the ancient oceans and the organisms that inhabited them may become
known. Modem sediments, though important in understanding what is
today, are doubly important because knowledge of them supplies the only
key to what would otherwise be an unknown past.

90 PHYSICAL OCEANOGRAPHY

THE INTERVALS THAT SHOULD OBTAIN BETWEEN DEEP-
SEA SOUNDINGS TO DISCLOSE THE OROGRAPHY

OF THE OCEAN BASINS

By G. W. LnTLSHALBS

The intervals between sounding stations must be gauged by the dimen-
sions of the orographical features whose presence it is intended to disclose.
Leairing out of consideration details of topography and confining the
attention to features of the greatest prominence, inquiry must be made
as to the form and dimensions of the slenderest isolated submarine peak
that could be raised from the floor of the ocean to a mountainous height
and remain standing under the stresses of its own weight and of the
superincumbent body of water. For if the spacing of soundings be such
as to give indication of the presence of the slenderest form that could
stand, then evidence of the presence of any orographical forms that may
exist is likely to be afforded. Theoretically, the shape of an isolated
submarine peak would be that of a solid of revolution in which the
resistance to crushing of any horizontal section is equal to the combined
weight of the portion of the formation above that section and of the
superincumbent body of water.

Let y denote the radius of any horizontal section and z its distance from
the top of the formation. Let K denote the coefficient of resistance to
crushing of the material composing the formation; w, the weight of a
unit of its volume ; and w', the weight of a unit volume of sea water.

Accordingly, irw f y*dz = the weight of the formation above any

section whose distance from the top is z,2rw'fy,zAy—Tyr'jy*dz^

the weight of the water upon the formation above any section whose
distance from the top is z, assuming the top of the formation to reach
to the surface, xKy*— the strength of any section to resist crushing, and

TW f y «dz + 2rw' f y .z .dy — rw' Jy'dz = xKy* ( 1 )

By differentiation, equation ( 1 ) becomes

Twy*dz+2Tw'y.z.dy — irwV.dz = 2TK.y.dy (2)

which expresses the condition that the increase of strength of any section
in excess of that of the section next above is equal to the sum of the
increases of the weight of the formation and the weight of the water
upon any section in excess of their combined weight imposed upon the
section next above.

PHYSICAL OCEANOGRAPHY 91

Letting S denote the area of any horizontal section whose radius is y,
and dS, the differential of S, equation (2) may be written in the following
forms:

w.S.dz +w'.z.dS — w'S.dz = K.dS
(w - wO S.dz = (K - w'z) dS
dS_, « dz w— w' dz

c- = (W-WO

S ^'^ ^K-Vz w' K

V-"

By integration, equation (3) becomes

log

s.-— l„,(|-.)+c

in which C is the constant of integration.

(K \ w' w'
-> — Z J = ; C ; log S
w' / w— w' w— w'

>c

-, w— w
K £

or "7 — z =
w

E

{:^^<^^^)

In the absence of knowledge of the value that should be assigned to K,
the coefficient of resistance to crushing, this equation has been used in
the generalized form,

B B B .

A— z =

/ w^ , ^.\ / 1.03 , c\" £1-46 log loS

to find the equation of their average form from the observed bathymetri-
cal data on Seine Bank in latitude 33'' SCX N. and longitude W 20 W.,
Cocos or Keeling Island in latitude W 06' S. and longitude 96** 53' E.,
Enderbury Island in latitude 3** 10^ S. and longitude \7V 10' W., Funa-
futi Atoll in latitude 8^ 25' S. and longitude 179** 07' E., Taviuni Bank
in latitude 12' 05' S. and longitude 174** 35' W., and the shoal near Mid-
way Island in the North Pacific Ocean in latitude 28** 00' N. and longitude
177** 4a W.

For this purpose the values of z and y, expressed in nautical miles,
were inserted in the above equation, and a conditional equation was
formed for each pair of coordinates relating to each of the submarine
formations. From these conditional equations normal equations were

92

PHYSICAL OCEANOGRAPHY

Fig. 1. Profiles of isolated submarine peaks.

PHYSICAL OCEANOGRAPHY 93

found by the method of least squares, which gave the values of the con-

1.87
stants A and B. The resulting equation is 1.87 — z = r-n >|>l6lQg ^

and the corresponding curve, which by revolution around the vertical
axis would generate the average form, is shown in figure 1, together with
others which have been plotted for purposed of comparison from measured
data. This investigation shows that isolated formations occupying com-
paratively limited areas at the bottom can and do occur in the ocean
depths, and we are able to assign at once the maximum interval that
should obtain between deep-sea soundings taken in operations directed
toward the development of the orography of the bottom of the sea. An
interval of 8 miles coupled with a differential interval of 2 miles would
serve for general development, and would prove with certainty the exist-
ence or absence of any formation rising close to the surface. Of all the
possible ways in which an 8-mile interval could lie with reference to a
submerged peak, that which would be most advantageous for a prompt
discovery of its existence is the condition in which one end of the interval
is at the bottom of the slope and the other near the apex, and that which
would be least advantageous is the condition in which the interval is bi-
sected by the position of the apex. In the latter case, there would be
nearly equal soundings at both ends, but the soundings at the ends of the
adjacent two-mile intervals would in all probability give indication of the
slopes.*

^ Following the presentation of this paper the following suggestion was made by
Harry Fielding Reid:

Dr. Lattlehales' remarks about the soundings in the oceans bring up a matter that
I have had in mind for some time; that is, the value of a detailed sounding of a
single deep. We know very little indeed about the shape or conformation of the
great ocean deeps ; a detailed set of soundings of a particular deep, to bring out not
merely the general slope of the bottom, but also details of configuration, would be
of great value. If, as seems probable, the great deeps are due to faulting, the sound-
ings should be close enou^ together to show the existence of fault-scarps. A deep
which offers especial facilities for such determinations is the Virgin Islands or
Bronson Deep. It is a long east and west trough, lying a little north of Porto Rico,
with a recorded sotmding of 4,662 fathoms (the greatest depth measured in the
Atlantic) ; although but few soundings have been made in its deeper parts. Its
situation is very convenient; San Juan could be used for a base for the western
part and St. Thomas for the eastern part

There are, of course, other parts of the Caribbean region where soundings would
be valuable, but I think a detailed sounding of a single deep would yield more
valuable results than scattered soundings over a larger area.

94 PHYSICAL OCEANOGRAPHY

NEW METHODS OF OBSERVING WINDS AT FLYING

LEVELS OVER THE OCEAN

By Alexander McAoie

Aerography may be defined in a general way as a study of the structure
of the atmosphere. There are various ways of obtaining information
regarding the flow of air at different levels and the conditions of density,
pressure, and temperature of the mixture of air and vapor. Exploration
of the upper air has been accomplished by means of close study of the
clouds; the establishment of mountain observatories; the ascent of
manned balloons; kites and kite balloons; sounding balloons and pilot
balloons. To these we propose to add another where measurement is
made from the deck of a vessel by employing certain predetermined lapse
rates, or rates of fall in temperature with elevation.

The principle in brief is that provided sufficient water vapor is present
and condensed as cloud, the height of the level of condensation is a func-
tion of the lapse rate. The height can be obtained then from observations
of the actual temperature, the temperature of evaporation and the tem-
perature of saturation at sea-level, making proper corrections for surface
speed and direction.

It is a little more than twenty years since Teisserenc de Bort at Trappes,
and Lawrence Rotch at Blue Hill, close friends and co-laborers, began
the systematic sounding of the atmosphere by means of sounding balloons.
With the war came a widespread use of pilot balloons. Today, sondages
are made (or are supposed to be made) at all United States naval air
stations, and at many land stations. During the war information regarding
the speed and direction of the winds at flying levels thus obtained was
of great value — not alone to the airmen, but also to artillerists and gas
men.

There is no special difficulty in using sounding balloons or pilot bal-
loons on land; but at sea the sounding balloon is out of the question,
owing to difficulty of recovering the record. Pilot balloons, however, can
be used ; and during the trans- Atlantic flight of the N. C. boats, I obtained
fully a hundred observations from sea-level up to 4 or 5 kilometers, while
stationed on the U. S. S. Baltimore (mine layer).

We can not, however, expect navigating officers of our merchant marine
to send up balloons, follow them with theodolites, record the elevations
and angles, plot the trajectories and deduce from these the speed in
meters per second and the direction of motion for the different levels.
Not but that it would pay to do so ; for it will pay any navigating officer
to be posted concerning the structure of the air. It may sometimes mean
the safety of the ship. And an intelligent aerographic officer with a
moderate outfit of aerographic apparatus on a ship like the Mauretania
could tell from the upper air movements studied in connection with the
surface circulation, the location of the ship with reference to the true

PHYSICAL OCEANOGRAPHY 95

centers of gyratory and translatory flow, and could forecast the future
path of the storm. A daily weather map or one at more frequent inter-
vals, based on reports received by radio, could be made and used to great
advantage with this added knowledge of the upper air conditions.

In cloudy weather the pilot balloon may soon be lost and it is therefore
advisable to substitute for the balloon method a method which makes use
of clouds, especially lower clouds within 35 degrees of the zenith. Of
course in dense fogs, neither method cajrbe used.

The new method makes use of a specially stabilized nephoscope with
automatic sighting rods, and an arc with tangent values ; also a new type
of hygroscope. The combination may be called a marine altoscope.

The nephoscope consists of a black mirror suitably mounted (for de-
tails of construction see Blue Hill Report, 1910) to permit of motion in
azimuth, proper leveling devices, and graduated circle, reading clockwise
and in either degrees and tenths or in grads. To this mirror is attached
a stabilizing device, suggested by Professor R. W. Wilson of Harvard
University. The mirror thus keeps a horizontal position regardless of
the ship's motion.

A metal arc or quadrant springs from the plane of the mirror and is
graduated in degrees, and also in natural tangents, the reason for which
will appear later.

At the free end of the arc a vertical rod is mounted and carries a panta-
graph or diamond-shaped rectangle supporting two rods for sighting the
cloud. Use is made of the reflection of the line joining cloud and eye,
and the second sighting rod forms a straight line prolongation of the line
from the center of the mirror to the cloud. The value of this is in fixing
the eye, whatever the ship's motion may be. When once set, the eye can
be withdrawn or rested for a few seconds and then brought back to the
original position without delay or uncertainty. The radials can be pro-
vided with sleeves permitting extension.

In observing, first level the instrument. Bring the zero which is also
360* or 400 grads of the horizontal or azimuth circle to the true south
point. The circle is graduated clockwise and the true west will therefore
be 100 if scale is in grads (90' if in ordinary units).

Since the reflection of the cloud crosses the mirror in the same direction
as the cloud is moving, the reading on the azimuthal circle where the
cloud image passes off the black mirror will be the direction or angle
from which the wind is blowing.

The quadrant is now swung into position, making the same angle. With
the control screw provided for the purpose of raising or lowering the
sighting rods, bring the nearer sighting rod into perfect alignment with
the reflection of the other sighting rod. This latter rod joins the cloud
point and the center of the mirror. We have now the angular elevation
of the cloud from a true horizon. When this angle is 50 grads or 45
degrees, it is plain that the distance the reflection of the cloud moves in
the mirror is equal to the height of the intercept corresponding to the

96 PHYSICAL OCEANOGRAPHY

height of the cloud; that is, the sine and cosine of the angle are equal,
and the natural tangent is unity.

In such a case, we have only to divide the height of the cloud (to be
determined later) by the number of seconds to get the rate in meters
per second.

If, however, the cloud line does not make an angle of 45, we use directly
the value of the tangents. The following condensed table gives these
values :

Tangent Grads Degree Tangent Grads Degree

A = 24

22

.5 = 29

26

.6 = 34

31

.7 = 39

35

.8 = 43

39

.9 = 47

42

1.0 = 50

45

1.1 = 53

48

1.2 = 55

50

1.3 = 58

52

1.4 = 60

54

1.5 = 62

56

1.6 =

64

58

1.7 =

66

59

1.8 =

68

61

1.9 =

69

62

2.0 =

70

63

2.5 =

75

68

3.0 =

80

72

4.0 =

85

76

5.0 =

87

79

6.3 =

90

81

11.4 =

95

85

00

100

90

One has only to divide the height of the cloud by the arc reading (i. e.,
tangent value) to get the horizontal distance. This last divided by the
number of seconds gives the speed of the cloud in meters per second.
We thus have direction and speed of the air at the cloud level, provided
the height of the cloud is known.

To get the height we use a special t)rpe of psychrometer (McAdic
cryoscope). The improvements over the usual psychrometers are:

(1) The amount of air passing over the wet-bulb is under control;
i. e., a definite value is given to the wind factor in evaporating the film
of water.

(2) The method of wetting the bulb is novel. The old method of usiii^
a wick or muslin cloth, bringing a constant supply of water by capillary
action, is replaced by a fine metallic mesh shaped to slide over the bulb,
easily wetted and containing a known small weight of water, to be evapo-
rated in a given time.

(3) The conversion of vapor pressure into units of force permits the
use of a simple equation connecting the actual temperature, evaporation
temperature, and saturation temperature.

Of the above factors, the wind velocity is of great importance and must
be known definitely if the humidity records are to be regarded as reliable.
It may be pointed out that even in official meteorological services at home
and abroad the records of relative humidity are open to criticism on the
ground of uncertain ventilation. In the best forms of sling and whirling
devices no record is kept of the time and number of revolutions.

PHYSICAL OCEANOGRAPHY 97

In the present instrument a definite wind velocity is automatically main-
tained and the beginning and ending of the movement of the air over the
evaporating surface, or what is approximately the same, the movement
of the wet-bulb through the air, is definite. The wetted bulb can be swung
either vertically or horizontally at any desired speed from 4 to 10 meters
per second. The thermometers are carried by a frame which slides on the
rod and their distance from the top of the rod or axis of rotation deter-
mines the velocity of the equivalent wind. Thus at a distance of 100
centimeters (39.3 in.) the bulb when whirled will travel in one complete
revolution 6.283 meters (approximately 20 ft.). It is then only necessary
to know the number of rotations and the time to get the speed of the wind.
An automatic counter is so connected with the handle that at the comple-
tion of every hundred revolutions an alarm bell rings. With a little
practice one makes 100 swings per minute.

If desired a watch may be used and the number of seconds counted.
The rate mentioned, one hundred per minute, is equivalent to a wind of
10.5 meters per second (23.5 miles per hour).^

Now, the rate of evaporation varies as the square root of the wind
velocity. Thus the rate at 10.5 m/s is to the rate at 4 m/s as 16 to 10.
The hygrometric tables in common use were based on experiments in
which the speed of rotation was approximately 4.5 meters per second,
although no definite statements are made and there appears to have been
no special attention paid to the speed of rotation or the rate of fanning
of the wet bulb. Naturally discordant results are obtained by different
observers. The speed mentioned (4.5 m/s) is somewhat too low for a
good circulation of air, and is indeed below the average wind value at
most places. The value of 10 meters seems to be a more representative
figure.

In the present instrument the pressure of the water vapor at any tem-
perature ordinarily met with above the freezing point is expressed in
units of force, and so far as known this is the first instrument employing
these units for water vapor. A kilobar is that pressure which if exerted
as force would give an acceleration of one centimeter per second per sec-
ond to a mass of weight one kilogram. Roughly, it is the pressure given
by a wind of 12 meters per second on a plane one meter square and at
right angles to the wind. Thus, temperature, pressure and weight are
expressed in a uniform, consistent and scientific set of units, namely, the

^ In the sling psychrometer used by the Bureau of Mines, if we assume a speed of
100 revolutions per minute the equivalent wind would be about 2.9 m/s (6 miles per
hour). There is no counting device and while a higher rate can be obtained, it is
difficult to count by the eye more than 120 per minute. In the whirled psychrometer
used by the Weather Bureau, the radius of rotation of the bulbs is about the same
as in the Bureau of Mines instrument, but a geared handle permits of varying the
rate from 175 to 260. The velocity equivalents will vaiy from 9 to 16 miles per
hour, the rate of evaporation in the former being only 75% of that in the latter.
McAdie has suggested a simple form of counter for this instrument to standardize
the results and has used such a device at Blue Hill Observatory for two years.

98 PHYSICAL OCEANOGRAPHY

kilobar, kilograd, kilogram. These are strictly in accord wiA the C. G. S.
system of um'ts.

To determine relative and absolute humidities, and the temperature
of saturation, the so-called dew-point, there is used an equation given
by the author in the Physical Review, Vol. XIII, No. 4, page 285.

in which p, is the pressure of the water vapor at the saturation or dew-
point, p^ the pressure of evaporation — that is, the wet-bulb — p the pres-
sure of the atmosphere expressed in kilobars, C a constant, t the tempera-
ture of the dry-bulb expressed in kilograds, and t^ the temperature of
the wet-bulb.

When the wind velocity exceeds 2 m/s, pC may be written as 0.18 ; and
for purposes of quick calculation we regard it as 20 percent without
materially affecting the result.

I stop at this point to read part of a letter just received from Sir Napier
Shaw. He says :

As to the inter-relation of meteorology and oceanography, I think that homidity
probably offers the most promising line of attack, if we could be quite certain oi
getting true humidities on board ship. I suppose that there must be a mathematical
expression for the absolute humidity depending upon the air current and the eddy
motion which it carries. I could imagine a very useful expedition tracing the in-
crease in absolute humidity down the Trade Wind and ultimately to uie West
Indies; but it is very difficult to get humidities on board ship because the dry bulb
is apt to get wet and the wet bulb to get dry; and both of them to be spoiled by
spray. But he will be a great benefactor who will give us a map of the distribution
of absolute humidity over the Atlantic Ocean.

Three things in the quotation are important : the suggestion of the map,
the expression of belief in the humidity problem as a most promising
liaison between meteorology and oceanography, and the remark about the
difficulty of getting accurate humidities aboard ship.

Granted, then, that we can get these humidity values at sea with much
greater precision by these new instruments, we proceed to use these
values in determining the cloud heights.

The temperature of saturation can be obtained without the use of tables,
which are always troublesome to use aboard ship, owing to high winds,
from the cryoscope, or, if desired, from the accompanying chart (figure 2).

An example will show how this is done.

Let the dry reading be 1063 and the wet, after proper precautions,
1053. The relative humidity is at once shown by the dotted line to be
74, and the dew-point, obtained by running back to left-hand edge of chart
parallel to the solid lines, 1046. If the absolute humidity is desired, one
has only to follow the 1046 line horizontally to the right-«nd edge ; and
one reads 1 1 grams per cubic meter of space.

We will call 1046 the cloud point or temperature of condensation
(heretofore called dew-point, but the new name has some advantages).

What we now want is the difference between the surface temperature

TEMPERATURE IN KILOGRADS

Fig. 2. Absolute and relative huinidity

PHYSICAL OCEANOGRAPHY

Fig. 3. Cloud heights from surface hutnidtty

PHYSICAL OCEANOGRAPHY 101

axid the cloud levd ; or what may be caOed the depression of the cloud
temperature (see figure 3). 1063 — 1046 = 17. The cloud height oppo-
site 17 is 750 meters for a day of light winds and 600 meters for a
windy day. A correction for percentage of saturation and type of struc-
ture is desirable.

I The height of the cloud being known, the direction and velocity are

j obtained as described, and the observer can compare these values with

the surface values. Nearly always there will be differences. In fair
weather there is generally a steady shifting of the wind to a higher value
for both speed and direction. At Blue Hill the mean deviation for the
1000-meter level is 7 grads, or 6 degrees to the right. The increase in
speed is variable, often 100 percent in the first 500 meters, and we have
instances of 200 per cent. On the land we get all sorts of structures, in
some of which, such as sea breeze, the depth of the surface flow is shallow
and essentially different from the flow above. The values obtained by
this nephoscope-cryoscope method are approximately gradient velocities
and directions. It is possible to construct a chart when gradient velocity
direction and latitude are known, from which the pressure gradient can
be deduced ; and thus in a rough way the isolated observer could obtain
the curvature of the isobar and pressure tendency. In former years this
would have meant much ; but now, of course, full reports can be obtained
by radio and the surface isobars easily drawn.

It only remains to explain the variation in the value of the lapse rate
on different days, or rather with different structures.

While the adiabate rate is 35.5 kilograds per 1000 meters, an average
rate of cooling of mixed air and vapor is 21 kilograds.

For moist air saturated, a value of 18 may be taken. In windy weather,
a fair value is 25 grads.

PHYSICAL OCEANOGRAPHY

THE STEERING LINE OF HURRICANES
Bv ALSXAMnn UcAdik

As a frontispiece to the "Manual of Meteorology," Part IV, "The ReU*
tion of Wind to the Distribution of Barometric Pressure," Sir Napier
Shaw gives three storm paths of unusual duration and remarkable re-
curvature (see figure 4),

Fig. 4. The tracks of some storms of long duration (after Shaw)

Perhaps the most striking of these is a track of a typhoon or bagnio
charted by McAdie. This storm path was determined by the usual method
of connecting pressure minima. The readings were obtained from ab-
stracts of ships' logs, available through the courtesy of the Hydrographtc
Office. Surface winds and cloud directions were utilized as much as
possible.

PHYSICAL OCEANOGRAPHY 103

It was agawimrd that the miniHHun pressure and the center of circula*
tkm as indicated by surface winds were identical. It is, however* to be
remembered that the wind direction as noted on the deck of a moving
vessel may need correction. Fnrthermore the center of a cyclone is not
necessarily the center of ascending air; and still further there must be in
the convergence of the surface winds a certain distortion due to the travel
of the storm.

The storms referred to above are perhaps best described in the words
of Sir Napier Shaw (page 119).^

There b evident stability in motion of this cfaaracler because beginning wtdi ex-
amples of wliirb lasting for some seconds there is a^iparently an umntermpted
sequence by way of rcvoivina sandstonns or dnst-devils, tornadoes, or whirlwinds,
to tropical revolving storms and large cyclonic areas with radii of 10 degrees or
more.

The onty limit of the scries is a revolving air-cap covering the hemisphere or a
large part of it And just as a belt of west wind or a belt of east wind may lie
over dese [British] Isluids for weeks, so the other type of quasi-permanent atmos-
pheric motion, which has always been diought of as a column of air in continuous
revolution, may preserve its identity for days or weeks. Through the kindness of
Professor McAdie of Blue Hill Observatory, Harvard University, we are enabled
to give two notable examples.

The first is that of a tropical revolving storm which started on a westerly track
toward die Philippine Islands (where visitations of that kind are known as
"Bagnios"), turned round toward the north and northeast, crossed the Pacific Ocean
and, after some vagaries on the North American continent, continued its journey
eastward and crossed the Atlantic in the usual track of cyclonic depressions over
that ocean. The whole journey lasted from 20th November, 1895, to 22d January,
1896.

The second is a cyclonic depression of October, 1913, in the outer region of which
the tornado was formed whidi caused so much destruction in South Wales on the
27th of that month.* The track of the main depression shows an anomalous path
from Canada to the north of the British Isles. [See figures 4, 5, and 6 from the
"Geographical Review."]

To these notable examples has been added the long track of cyclonic depression
whidi was figured in the Meteorologiod Office chart of the North Atlantic and
Mediterranean for August, 1904.' The cyclone was first noted on 3rd August, 1899,
m that part of the North Atlantic Ocean where West Indian hurricanes often take
their rise. It moved westward to the West Indies, skirted the coast of Florida and
turned eastward over the Gulf Stream. After some hesitation about latitude 40* W.
it made for the mouth of the English Channel and, missing that, crossed to the
Mediterranean, where it lost itself on 9th September, after a life of thirty-eight days.

In each of the above described storms it is evident that causes other
than those developed by the rotating mass of air, operated to retard these
storms in their eastward progress.

Let us now trace the path of a West Indian hurricane where the evi-
dence is seemingly more direct.

On the morning of October 15, 1910, this storm was centered between
Havana and Key West moving very slowly northward. The maximum
wind velocity at the former place was 39.4 m/s (88 miles per hour) ; on
the a. m. of the 14th; and at Key West 26.8 m/s (60 miles per hour).

* See also ** Wandering Storms," McArdie, Geographical Review, 10, no. I. July,

1920.

* Geophysical Memoirs, no. 11. M. O. Publication, no. 22a.

* M. O. Publication, no. 149.

PHYSICAL OCEANOGRAPHY

Fic. 5. Track of storm of September 27-October 28, 1913.

PHYSICAL OCEANOGRAPHY 105

The stoim's progress northward was checked by a continental area of
high pressure moving southward. Thus on the 17th we find the hurricane
actually retrograding and centering again over Havana. As the conti-
nental anticyclone moved east, the hurricane developed a northerly com-
ponent of motion and on the 18th moved across Florida. It then fol-
lowed the usual hurricane track passing south of Cape Hatteras on the
20th. The hourly speed increased from 30 kms. to 50 kms. per hour and
the direction of motion 40 degrees east of north.

The speed continued to increase averaging 60 kms. per hour and the
direction shifted more to the east, approximately 65 degrees east of north,
and so at noon of October 21 the center was in the latitude 37 degrees
north and 67 degrees west.

Professor Bjerknes has remarked that "anticyclones are bom as
cyclones die" but the behavior of this and similar storms gives the impres-
sion that the path and speed of West Indian hurricanes, off the coast of
Florida, are dependent upon the intensity and direction of advancing
highs. These in turn may be but the surface expression of an advancing
polar front.

Two types of south moving sub-Arctic surges which seem to control
the path of hurricanes from the Caribbean Sea to the North Atlantic can
be identified. The first of these is a Nichikun high. This is a more appro-
priate designation of what has heretofore been known as a Labrador high.
According to Dr. Klotz * there is nowhere else in Canada "so distinct a
Pamir or Roof of the World as the neighborhood of Lake Nichikun (in
English, Otter Lake)." The lake itself is in latitude 53* N., longitude
71** W., and on the northwest slope of the Height of Land. The drainage
is into Hudson Bay. On the south and east the drainage is into the River
St. Lawrence. It is this southern slope which concerns us because south
moving masses of air pass over the ridge, elevation 730 meters ; and being
both cold and dry and therefore heavy, fall to sea level in a comparatively
short distance, 200 to 600 kilometers.

The other type of sub- Arctic surge is the "Labrador," essentially
oceanic.

Both of these tongues may be portions of what Bjerknes has called the
polar front. They undoubtedly play an important part in determining
the speed and path of storm centers in the North Atlantic States and effec-
tively control the path of tropical storms or hurricanes as they move from
the south and change into North Atlantic cyclones.

On the Pilot Chart of the North Atlantic Ocean for October, the path
of the hurricane under discussion ends abruptly in the position and on
the date given above (Oct. 21). One might in consequence infer that the
storm dissipated at sea.

Careful study of pressure conditions shows a depression on the 22d in
latitude 35** N. and longitude 60** W. A day later it appears as one of
two centers in a large depression extending from New Brunswick to

* In a letter to the writer.

PHYSICAL OCEANOGRAPHY

PHYSICAL OCEANOGRAPHY 107

Bennada. The other center can be traced back to a storm over Lake
Superior on October 21. The previous history of this depression, while
somewhat obscure, is deserving of study. It appeared as an unexpected
abnormal devel(qmient and invalidated all forecasts made for the Lake
R^on, Upper Mississippi and Ohio Valleys. Where cold weather, frosts
and an absence of precipitation were reasonably anticipated from an ad-
vancing high pressure (1030 kb.), there suddenly developed warmer
weather with rain. On the face of the map we are unable to connect this
low with a more northern slow moving depression of the Alberta type.
The weather map of October 20, 1910, will repay study in connection
with the steering line of cyclones.

To return to the hurricane and its further history, we have seen that
when centered over Florida, there were in juxtaposition two air masses
of different origin, one from the tropics with a vapor content of not less
than 20 grams per unit volume (one cubic meter of space) and an average
northwest speed of one kilometer per hour, while the other air mass was
of sub-polar origin, approximately 20 kilograds (5.5 degrees C.) colder,
and with an average vapor density of 12 grams per cubic meter. The
densities of the two air masses at a pressure of 1 megabar would be
approximately 1170 and 1220 grams. Air motion is initiated by differ-
ence of pressure rather than difference of density ; but it is plain that the
south moving air mass would continue to gain momentum and underrun
the less dense northbound air. The horizontal pressure gradient was
1 kb./20 km. and hence surface velocities of 30 meters per second or
higher would and did occur. The gradient velocities were 23 m/s or
higher ; and the radii of survature of isobars approximately 100 kilometers.

Figure 7 shows the path of the hurricane from October 13 to 25, and
also the path of the lake "low" from October 21 to 24. Other charts
show the surface pressure distribution on various dates.

It is much to be r^^etted that there are no records of winds aloft.
When such data shall be available then perhaps definite relations between
path, velocity and duration of hurricanes with upper' winds will be forth-
coming.

Recently it has been claimed by meteorologists of the Bergen (Norway)
Institute that the storms of the Northern Hemisphere can be traced back
to a "surface of junction of polar and equatorial air." This surface can
be detected at the ground as "a line of discontinuity" in surface condi-
tions. In other words, it is the boundary between air masses of different
densities, pressures, and vapor content.

(jiven then a mass of warm moist air moving north of east, under the
combined effects of general drift, pressure gradient and rotational deflec-
tion, and a second mass of cold dry air moving south, the surface of dis-
continuity should be detectable as a moving front.

Professor Bjerknes has come to the conclusion from the study of the
structure of moving cyclones that a broad belt of rain accompanies the
moving (and ascending) warm moist air, and a second smaller rain belt

108 PHYSICAL OCEANOGRAPHY

follows, where cold dry air underruns the warm air, that is, along the
wind shift or squall line.

A more important point, however, is the discovery through the use of
detail maps, that the discontinuity or contrast can be traced from any
cyclone to another. As expressed by Bjerknes, ^dones follow each other
along a common line of discontinuity like "pearls on a string/'

Furthermore this line of discontinuity surrounds the polar regions as a
closed circuit. It shows how far the cold air flowing along the ground
has penetrated. Shaw describes it as a kind of polar front Hne.^

The following substance of the discussion at the Meteorological OflSce
on "new methods of forecasting" may make plain the leading features of
Bjerknes's views.'

In the case of a cyclone making progress towards tfie east, a sector to the south
is occupied by a warm current; this warm area on the earth's surface is bounded
to the north by the "steering line," to the west by the "squall-line." Bjerknes'
generalization is "that these squall-lines and steering lines of all the cyclones of
Sie northern hemisphere are parts of a single line — *the polar front'" We are to
think of two great streams of air, both flowing from the west, the more northerly
stream being colder and carrying less moisture. The boundary between these two
streams is imstable and its oscillations manifest themselves as cyclones. The warm
stream overrides Uie cold one, which retaliates, so to speak, by turning round and
kicking its partner in the back.

Charts 4, 5 and 6 are reproduced through the courtesy of the Geo-
graphical Review, published by the American Geographical Society, New
York City.

^ Nature, January 24, 1920, p. 524.

* Meteorological MagoMme, November, 1920, p. 213.

Bulletin of the National Research Council

Volume 1

Number 1. The national importance of scientific and industrial research. By
George Ellery Hale and others. October, 1919. Pages 43. Price 50 cents.

Number 2. Research laboratories in industrial establishments of the United
States of America. Compiled by Alfred D. FUnn. March, 1920. Pages 85.
Price $1.00.

Nwnber 3. Periodical bibliographies and abstracts for the scientific and tech-
nological journals of the world. Compiled by R. Cobb. June, 1920. Pages
24. Price 40 cents.

Number 4. North American forest research. Compiled by the Committee on
American Forest Research, Society of American Foresters. August, 1920.
Pages 146. Price $2.00.

Number 5. The quantum theory. By Edwin P. Adams. October, 1920. Pages
81. Price $1.00.

Number 6. Data relating to X-ray spectra. By William Duane. November,
1920. Pages 26. Price 50 cents.

Number 7. Intensity of emission of X-rays and their reflection from crystals.
By Bergen Davis. Problems of X-ray emission. By David L. Webster.
December, 1920. Pages 47. Price 60 cents.

Number 8. Intellectual and educational status of the medical profession as
represented in the United States Army. By Margaret V. Cobb and Robert
M. Yerkes. February, 1921. Pages 76. Price $1.00.

Volume 2

Number 9. Funds available in 1920 in the United States of America lor the en-
couragement of scientific research. Compiled by Callie Hull. March, 1921.
Pages 81. Price $1.00.

Number 10. Report on photo-electricity including ionizing and radiating poten-
tials and related eitects. By Arthur Llewelyn Hughes. April, 1921. Pages
87. Price $1.00.

Number 11. The scale of the universe. Part I by Harlow Shapley. Part II by
Heber D. Curtis. May, 1921. Pages 47. Price 60 cents.

Number 12. Cooperative experiments upon the protein requirements for the
growth of cattle. First report of the Subcommittee on Protein Metabolism
in Animal Feeding. By Henry Prentiss Armsby, Chairman. June, 1921.
Pages 70. Price $1.00.

Number 13. The research activities of departments of the state government of
California in relation to the movement for reorganization. By James R.
Douglas. June, 1921. Pages 46. Price 60 cents.

Number 14. A general survey of the present status of the atomic structure
problem. Report of the Committee on Atomic Structure of the National
Research Council. By David L. Webster and Leigh Page. July, 1921.
Pages 61. Price 75 cents.

Number 15. A list of seismological stations of the world. Compiled by Harry
O. Wood. July. 1921. Pages 142. Price $2.00.

Volume 3

Number 16. Research laboratories in industrial establishments of the United
States, including consulting research laboratories. Originally compiled by
Alfred D. Flinn; revised and enlarged by Ruth Cobb. December, 1921.
Pages 135. Price $2.00.

Number 17. Scientific papers presented before the American Geophysical Union
at its second annual meeting. January, 1922. Pages 108. Price $1.50.

Reprint and Circular Series of the National

Research Council

Number 1. Report of the Patent Committee of the National Research CounciL
Presented for the Committee b^ L. H. Baekeland, Acting. Chairman.
February, 1919. Pages 24. Price 30 cents.

Number 2. Report of the Psychology 'Committee of the National Research
Council. Presented for the Committee by Robert M. Yerkes, Chairman.
March, 1919. Pages 51. Price 60 cents.

Nmnber 3. Refractory materials as a field for research. By Edward W. Wash-
burn. January, 1919. Pages 24. Price 30 cents.

Nmnber 4. Industrial research. By F. B. Jewett. 1918. Pages 16. Price 25

cents.

Nmnber 5. Some problems of sidereal astronomy. By Henry N. RosselL
October, 1919. Pages 26. Price 30 cents.

Number 6. The development of research in the United States. By James
Rowland Angell. November, 1919. Pages 13. Price 25 cents.

Nmnber 7. The larger opportunities for research on the relations of solar and
terrestrial radiation. By C. G. Abbot. February, 1920. Pages 14. Price
20 cents.

Number 8. Science and the industries. By John J. Carty. February, 1920.
Pages 16. Price 25 cents.

Number 9. A reading list on scientific and industrial research and the service of
the chemist to industry. By Clarence Jay West. April, 1920. Pages 45.
Price 50 cents.

Nmnber 10. Report on organization of the International Astronomical Union.
Presented for the American Section, International Astronomical Union, bv
W. W. Campbell, Chairman, and Joel Stebbins, Secretary. June, 1920.
Pages 48. Price 50 cents.

Number 11. A survey of research problems in geophysics. Prepared by Chair-
men of Sections of the American Geophysical Union. October, 1920.
Pages 57. Price 60 cents.

Nmnber 12. Doctorates conferred in the sciences by American universities in

1920. Compiled by Callie Hull. November, 1920. Pages 9. Price 20 cents.

Number 13. Research problems in colloid chemistry. By Wilder D. Bancroft
January- April, 1921. Pages 54. Price 50 cents.

Number 14. The relation of pure science to industrial research. By John J.
Carty. October, 1916. Pages 16. Price 20 cents.

Number 15. Researches on modern brisant nitro explosives. By C F. van
Duin and B. C. Roeters van Lennep. Translated by Charles E. Munroe.
February, 1920. Pages 35. Price 50 cents.

Number 16. The reserves of the Chemical Warfare Service. By Charles H.
Herty. February, 1921. Pages 17. Price 25 cents.

Number 17. Geology and geography in the United States. By Edward B.
Mathews and Homer P. Little. April, 192K Pages 22. Price 20 cenU.

Number 18. Industrial benefits of research. By Charles L. Reese and A. J.
Wadhams. February, 1921. Pages 14. Price 25 cents.

Number 19. The university and research. By Vernon Kellogg. June, 1921.
Pages 10. Price 15 cents.

Number 20. Libraries in the District of Columbia. Compiled by W. I. Swan-
ton. June, 1921. Pages 19. Price 25 cents.

Number 21. Scientific abstracting. By Gordon S. Fulcher. September, 1921.
Pages 15. Price 20 cents.

Number 22. The National Research Council. Its services for mining and
metallurgy. By Alfred D. Flinn. October, 1921. Pages 7. Price 20 cents.

Number 23. American research chemicals. By Clarence J. West. September,

1921. Pages 28. Price 50 cents.

Vol. 3. Part 3. AUGUST, 1922 Number 18.

Bulletin

OF THE

National Research

Council

THEORIES OF MAGNETISM

Report of the Committee on Theories of Magnetism of the

National Research Council

BT

A. P. Wills^ S. J. Barnett, L. R. Ingersoll, J. Kunz^ S. L. Quimby,

E. M. Terry, S. R. Williams

PUBUBHED BY ThB NATIONAL ReSEABCH COUNCIL

OF

Thb National Academy of Sciences

Wabhinoton, D. C.

1922

Announcement Concerning Publications

of the

National Research Council

The Proceedhigs of the National Academy of Sciences

has been designated as the official organ of the National Research
Council for the publication of accounts of research, committee
and other reports, and minutes.

Subscription rate for the "Proceedings" is $5 per year.
Business address: Home Secretary, National Academy of
Sciences, Smithsonian Institution, Washington, D. C.

The Bulletin of the National Research Council

presents contributions from the National Research Council,
other than proceedings, for which hitherto no appropriate
agencies of publication have existed.

The "Bulletin" is published at irregular intervals. The
subscription price, postpaid, is $5 per volume of approximately
500 pages. Numbers of the "Bulletin" are sold separately at
prices based upon the cost of manufacture (for list of bulletins
see third cover page).

The Reprint and Circular Series of the National

Research Council

renders available for purchase, at prices dependent upon the
cost of manufacture, papers published or printed by or for the
National Research Council (for list of reprints and circulars see
fourth cover page).

Orders for the "Bulletin" or the "Reprints and Circulars"
of the National Research Council, accompanied by remittance,
should be addressed: Publication Office, National Research
Council, 1701 Massachusetts Avenue, Washington, D. C.

PREFACE

The present report attempts to sketch in bold outline the evolution
and development of magnetic theories from the time of Poisson and
Ampere to the present, including some reference to experimental results,
particularly in the domain of magnetostriction where theory and ex-
periment are in the greatest need of reconciliation.

It will be noticed that the table of contents for the complete report
does not contain any reference to the HaU Elffect or allied phenomena.
The reason for this omission is that these topics have been assigned by
the National Research Council to another committee.

Space limitations have debarred from inclusion in the report some
material which appropriately might have found place there. Certain
portions of the subjects treated may have been emphasized more than
their importance deserves, while others have been unduly slighted.
The report, being a composite compilation by contributors so widely
separated geographically that close collaboration was not always pos-
sible, may lack somewhat in coherency.

But in spite of such deficiencies it is hoped that the report may furnish
a perspective of the subject which in its chief outlines is reasonably free
from distortion and that the reader may obtain from its perusal a fair idea
of the present status of magnetic theory.

The committee is indebted to Professor L. R. Ingersoll of the Univer-
sity of Wisconsin for his contribution on Magneto-optics.

i

BULLETIN

OF THE

NATIONAL RESEARCH COUNCIL

Vol. 3. Part 3 AUGUST. 1922 Number 16

THEORIES OF MAGNETISM

Report of the Committee on Theories of Magnetism of the

National Research Council^

CONTENTS

Magnetic theories prior to the discovery of the electron. By S. L. Quimby 3

Theories of para- and of diamagnetism. By A. P. Wilis 16

Theories of ferromagnetism — ^intrinfiic fields. By E. M. Terry 113

Theories of magnetic crystals and the magneton. By J. Kuns 165

Magnetostriction and its bearing on magnetic theories. By S. R. Williams 214

Theories of magnetostriction. By S. L. Quimby 225

The angular momentum of the elementary magnet. By S. J. Bamett 235

Magneto-optics. By L. R. IngersoU 251

MAGNETIC THEORIES PRIOR TO THE DISCOVERY OF

THE ELECTRON

Bt S. L. Quimbt

Instructor in Physics, Columbia University

THE BEGINNING OF THE SCIENCE OF MAGNETISM. GILBERT'S

"DE MAGNETE"

The science of magnetising as well as of electricity, began with the re-
searches of William Gilbert (b. 1540, d. 1603). The ancients were ac-
quainted with the fact that amber, when rubbed, attracts light bodies,
that the lodestone has the power of attracting iron, and that this power
can be conmiunicated to the iron by bringing it near to or stroking it
with a lodestone. The latter had been used as a mariners' compass at
least since the time of the Crusades. No attempt had been made, how-
ever, to order or extend the small amount of available knowledge con-
cerning these phenomena. Rather was it lost in a mass of false doctrine
bmlt about it by the medical profession, who were chiefly interested in

^ This oonmiittee of the Division of Physical Sciences of the National Research
Council consists of the following members: A. P. Wills, Columbia University, Chair-
man; S. J. Bamett, Carnegie Institution; J. Kuns, University of Illinois: S. L.
Quimby, Columbia University; E. M. Terry, University of Wisconsin; S. R. Williams,
Oberlin College.

3

4 EARLY MAGNETIC THEORIES: QUIMBY

utilizing the mysterious property of the lodestone for the curing of
disease. Gilbert, himself a physician, dissipated these erroneous
notions by clearly separating the medicinal from the magnetic properties
of the lodestone, and then proceeded with an exhaustive investigation of
the latter.

He differentiated sharply between electrical and magnetic attraction
by pointing out the difference in behavior of electrified amber and mag-
netized iron. He emphasized the dual nature of the magnetic element
and examined the effect of the shape of a magnet upon its strength.

After pointing out that the earth is itself a huge magnet, Gilbert
investigated the variation and dip of the magnetic needle over its surface
and coordinated a vast mass of data which he secured from mariners.

Apart, however, from the intrinsic worth of Gilbert's researches, his
work may be regarded as the forerunner of the modem scientific method.
His De Magnete (1600)^ contains the first formulation of natural law
based entirely upon the results of experiment. In it Gilbert applied
the method which was later set forth with logical precision by Francis
Bacon.

No material advance upon the knowledge of magnetic phenomena
recorded in GUbert's book was made for nearly two centuries. During
this period developments along different Unes were taking place which
eventually made possible the great progress in magnetic theory which
marks the nineteenth century. One of these was the improvement of
methods of experimentation. With the fundamental importance of the
experimental method once definitely established advancement along
this hne reached a point where Coulomb in 1785 was able to prove
satisfactorily the inverse square law of magnetic attraction and repulsion.

Another important factor in scientific progress about this time was
the rapid growth of mathematical analysis which followed the discovery
of the infinitesimal calculus by Newton and Leibniz. Under the in-
spiration of Laplace, Lagrange, and Legendre, mathematicians, par-
ticularly Poisson and Fourier, about the beginning of the nineteenth
century began to apply mathematical analysis to physical problems.
In 1812 Poisson published a memoir on electrostatics and in 1820
another on the theory of magnetism which remains to the present day a
correct mathematical formulation of the phenomenon of magnetic
induction.^

POISSON'S THEORY OF MAGNETISM

The starting point of Poisson's mathematical theory is Coulomb's
law that two magnetic poles attract or repel each other with a force

> EDgliah traoBlation by P. Fleury Mottelay; New York, John Wiley and Sons, 1893.
*PoiB8an, "Sur la Theorie du Maenetisme," M^moires de I'lnstitut, V (1820), p.
247 and 488.

EARLY MAGNETIC THEORIES: QUIMBY 5

inversely proportional to the square of their distance apart. As a
mechanism for the utilization of this principle he adopted the "two
fluid'' theory of magnetism which had been previously advanced by
Coulomb and others. In accordance with this theory Poisson assumed
that all magnetic substances consist of a large number of small particles
or magnetic elements containing equal quantities of positive and negative
magnetic fluid. These elements are themselves perfect conductors for
the fluids, but the spaces between them are impenetrable to the fluids,
which cannot be allowed to pass from one element to another. In the
unmagnetized state of the body the two fluids are united to form a single
neutral fluid. The process of magnetization consists in the separation
of the two fluids within the magnetic element, one being displaced in
one direction under the action of the magnetizing force and the other
in the opposite direction.

In applying Coulomb's law to calculate the interactions between
these magnetic elements, Poisson assimied that the force of repulsion
exerted by a quantity, qi, of magnetic fluid upon a quantity, qs, of the
same kind situated a distance r from it, is proportional to

and is independent of the substance of which the magnetized body is
composed.

Using this conception of the phenomenon of magnetization Poisson
solved the problem of calculating the magnitude and direction of the
resultant force exerted by a magnetized body of any shape upon a imit
magnetic pole situated at any point outside the body. He exhibited
this force as the negative gradient of a function V, which may be ex-
pressed as follows:

= j - (I- n) dS - j - (div I) dr.

s

where n is a unit normal to an element dS of the surface S bounding a
magnetized body of volume t. I is a vector such that if 5r be any
physically small element of volume within the body, then Br will be
the magnetic moment of that element of volume. It is therefore the
"intensity of magnetization" of the substance at a point within 5r.
The form of the fimction V shows that the magnetic effect of any mag-
netized body is the same as that which would be produced by a layer of
magnetic fluid of density I-n over its surface, together with a distri-

6 EARLY MAGNETIC THEORIES: QUIMBY

bution of density — divl throughout its volume. These are called
''Poisson's equivalent surface and volume distributions of magnetism/'

It is evident that for points inside the magnetised body r^^ will become
infinite for an element of the second integral in Poisson's expression.
This difficulty may be removed if we consider the point situated inside
a cavity in the medium, small in dimensions, yet very large compared
with the dimensions of the elementary nuignets themselves.^ A part of
the surface density I- n will be on the wall of the cavity and this part
will give rise to a finite force at the point inside it, whose value will
depend on the form of the cavity and on the magnetic polarisation at
the place. If we omit this purely local part of the magnetic force in
the cavity, the remaining part, which is that due to the polarised mass
as a whole, will be derived from the general volume density div I and
surface density I* n just as at an outside point. This latter part arising
from the system as a whole, omitting the local term depending on the
molecular structure at the point considered, is thus quite definite, and
is named the magnetic force H. In this way we arrive at a definition
of the magnetic force within a magnetised medium which is consistent
with the way it is defined for points external to the substance.

Though the h3rpotheses regarding the nature of the magnetic element
which Poisson adopted have not proved to be correct, the formuls of
magnetostatics which he developed remain valid and useful since they
rest upon the experimental fact of induced nuignetization and not upon
the nature of the mechanism by which this is brought about.

The mathematical labor of developing a complete theory of magnetic
induction foimded solely upon experimental data was later undertaken
by Lord Kelvin. In addition to freeing Poisson's theory from the hy-
pothesis of two magnetic fluids, Kelvin greatly enriched it and simplified
the conceptions involved by introducing the terminology which is used
today.* One such extension in particular should be mentioned here.

Poisson had pointed out that in general the intensity of magnetisation
in a homogeneous body is a linear vector function of the field intensity,
so that in general the specification of I in terms of H would require
nine constants depending upon the nature of the substance. If the
medium be isotropic as well, these nine constants reduce to one, so that
for this case:

I»kH.

The subsequent researches of Faraday, Pllicker and Tyndall having
revealed the fact that crystals possess different magnetic properties in
different directions, Kelvin extended the theory to a treatment of the

1 ef. Larmor. "Aether and Matter," p. 257.

t KelTin, "Reprint of Papers on Electrostatiot and Magnetism/' XXIV.

EARLY MAGNETIC THEORIES: QUIMBY 7

problem of magnetic induction in non-isotropic media. He showed^
that for such media the nine constants introduced by Poission reduce
to three, so that:

in which the linear vector function # is self-conjugate.

As it has been handed down to us by Poisson, Green, and Kelvin, the
mathematical theory of magnetic induction may be regarded as com-
plete. It is inadequate to meet the demands of the modem viewpoint
because it is essentially a statistical theory. It deals with the phenomena
exhibited by matter in bulk, without attempting to account for the
ultimate causes of these phenomena. Just as Thermod3mamics pre-
ceded Kinetic Theory, so the theory of magnetostatics has preceded a
study of the dynamics of the ultimate magnetic particle.

AMPERE'S THEORY OF MAGNETISM

In July 1820 Oersted announced the discovery that a magnetic needle
placed near a conductor carrying an electric current tends to assume a
position at right angles to the conductor. This discovery inspired
Ampere to imdertake a series of researches on the relation between
current electricity and magnetism which extended over a period of three
years and resulted in the publication in 1825 of a memoir on the mathe-
matical theory of electrodynamical phenomena which has been charac-
terized by MaxweU as, ''one of the most brilliant achievements of
science."^

Ampere based his theory of magnetism upon the identity which he
established between the magnetic properties of Poisson's "two fluid"
magnetic element and a solenoid of molecular dimensions in which an
electric current is continually flowing. According to Ampere the
molecules of a magnetic substance are perfect conductors about which
or within which are flowing perpetually minute currents of electricity.

The process of magnetization consists in changing the orientation of
these molecular currents either by changing the plane of the current
relative to the molecule, or by turning the molecule as a whole, so that
their axes, initially pointing in all directions at random, wiU tend to
align themselves parallel to the magnetizing field. . Ampere showed
that this sort of magnetic element would suffice to explain not only the
phenomena of magnetostatics in accordance with the formulse deduced
by Poisson, but also the laws expressing the mutual actions of magnets
and conductors carrying currents, which had been discovered by Biot,
Arago, and himself.

1 ibid. XXX.

«Ampere, ''M^moim de rinstitut." VI (1823), p. 175.

8 BARLY MAGNETIC THEORIBS: QUIMBY

At the time at which Ampere wrote, electromagnetic current induction
had not yet been disooveredy and therefore he was able to make no
h3rpothe6i0 as to the origin and strength of the molecular currents.
Ampere's great contribution to the science of magnetism consisted in
showing that all the then known interactions between magnets, and
between these and electrical conductors, could be reduced to a single
cause.

THE MAGNETIC RESEARCHES OF FARADAY

The fact that an electric current is invariably accompanied by a
magnetic field led Faraday to search for a converse effect.^

In a paper read before the Royal Society in 1831 he described a series
of experiments in which the phenomenon of electromagnetic current
induction was discovered. The establishment of this reciprocal relation-
ship between magnetism and current electricity afforded added support
to the molecular current h3rpothesis of Ampere as against the two
fluid theory of Poisson. Another discovery by Faraday, however,
sufficed to clinch the argument in favor of Ampere's theory and to
demonstrate that of Poisson to be untenable.

In 1845 while investigating the rotation of the plane of polarization of
a beam of light traversing a piece of glass placed in a strong magnetic
field, Faraday observed that the glass itself possessed magnetic proper-
ties opposite to those of iron and other "magnetic" metals.* While a
piece of iron would tend to set itself with its greatest length parallel to
the field, the glass, if left free to turn, placed itself across the field.
Faraday gave the name ''diamagnetism" to this new phenomenon,
and proceeded to make a thorough examination of the magnetic proper-
ties of a vast number of substances, solids, liquids and gases. He
definitely established the fact that all substances possess either the
diamagnetic or the magnetic property. He even concluded that,
"If a man could be suspended, with sufficient delicacy, and placed in
the magnetic field, he would point equatorially, for all the substances of
which he is formed, including the blood, possess this property."

In accordance with the two fluid theory of magnetism, the elementary
magnets of all substances would, when placed in a magnetic field, be
polarised in the same direction. Faraday showed, however, that the
direction of polarization of diamagnetic bodies in a magnetic field is
opposite to that of noagnetic bodies in the same field. The two fluid
h3rpothesis, therefore, fails in this respect to account for the facts.

Adopting Ampere's theory, a substance whose molecules were them-
selves elementary magnets due to the existence of permanent electric
■ ■ ■ ' ' ' — — — >r .

> Faraday, Eiperimantal Reaearchae, I, p. 2. The diamagnetic property of Biamuth
had prevYoualy been observed by Brugmana.
• Faraday* op. dt.. III. p. 27.

EARLY MAGNETIC THEORIES: QUIMBY

9

currents JSowing about them, would be magnetic. On the other hand,
if no such currents existed initially, then the action of an applied mag-
netic field might induce such molecular currents, and these, by Faraday's
law of current induction, would polarize the molecule magnetically in
opposition to the external field : that is, the substance would be diamag-
netic.

In 1852 Wilhehn Weber, adopting Ampere's h3rpothesis and the results
of Faraday's researches, developed mathematically a theory which it
will be profitably to outline here in some detail, for it laid the foundation
for certain of the modem theories of magnetism.

WEBER'S THEORY OF MAGNETISM

Weber starts by assuming that the molecules of a magnetic substance
are small permanent magnets whose axes in-
itially point in all directions at random.^ Let
NM (Fig. 1) be such a magnet, which is
capable of turning about its center C, under
the action of an external field H. If the
molecule were perfectly free to rotate then
the body would be magnetized to saturation
by any applied field, however small. This
Weber knew was not the case, and he there-
fore assumed a constraint upon the rotation
of the elementary magnets in the form of a

molecular nuignetic field, D, whose direction for each molecule coincides
with the initial equilibriimi position of its axis, and whose magnitude is
constant throughout the body.

The magnet will be in equilibrium imder the action of the two fields
when

_ Hsing (1)

^*""D + Hcos«'

If M denote the magnetic moment of the molecule, its component parallel
to H is, before the application of the field,

fACOBd

which, upon the establishment of the field, becomes

M cos (^ — ^).

Hence the increase in the magnetic moment parallel to H, say Mb>
due to the presence of the external field is given by:

Fig. 1

Mh

■'{

cos (^ — ^) — cos ^

}•

(2)

> W. Weber, " Uber den ZuBammenhang der Lehre vom Diamngnetiamua mit der Lehre
▼on dem Magnetismus und der Elektrioit&t," PogO' Ann. 87 (1864), p. 146.

10 EARLY MAGNETIC THEORIES: QUIMBY

Eliminating 4> between equation (1) and (2) we have for a aingle molecule :

( H + D cos ^ )

This expression must now be summed for all the molecules imder
consideration. Let there be n molecules per unit volume. Assiuning
initially a random distribution of the axes in space, the fraction of the
molecules whose axes make an angle less than 6 with H will evidently
be the ratio of the area of the zone cut from a sphere by a cone of semi-
angle 0, to the area of the sphere, that is }4{l — cos 6). The number
of molecules whose axes make angles with H lying between 6 and
$ + dSis, therefore,

The net increase in the magnetic moment per unit volume due to the
rotation of all the elementary magnets is, then, given by:

\J i

Mh sin ^ d^.

2 H

If H < D this integral has the value I = q M n fi'

2m n

Tf IT la D " " " " " I ^ - •

If H > D " " " " " I = M n

\ 3BP/

If H = 00 " " " " " I = M n.

An examination of these formulae shows that the intensity of
magnetization should increase proportionally to the impressed field
until it has reached ^ of its maximum value, after which it should
approach the latter as3rmptotically. Weber obtained experimental
results for iron in close agreement with this conclusion. His theory,
however, is unable to account for residual magnetism, and more accurate
experiments have shown that the initial variation of intensity of
magnetization with field strength is not linear.

Before proceeding to a discussion of the various modifications which
have been suggested to resolve these discrepancies, we will review briefly
Weber's theory of diamagnetism.

According to Weber's theory, there exist in the molecules of a
diamagnetic substance closed channels in which electricity can flow
without resistance. If a magnetic field is established through one of
these channels an electric current will be set in motion in it. The
magnetic field of this induced current will be opposed to the external

EARLY MAGNETIC THEORIES: QUIMBY 11

field. In the mathematical development of his theory Weber made use
of electrodynamical formulae derived from assumptions regarding the
nature of current electricity which have since been abandoned. It will
therefore be more profitable to examine the theory in the form in which
it was afterwards interpreted by Maxwell.^

If L is the coefficient of self induction of a molecular circuit, and M
is the coefficient of mutual induction between this circuit and some
other circuit, and if, furthermore, i is the current in the molecular
circuit, and i' that in the other circuit, then:

~ (li + MiO = ~ Ri.
dt

But by h3rpothesis R = O, and we get by integration :

li + Mi' = lio,

where io is thus the initial value of the molecular current.

If the current i' produces a magnetic field of strength H which makes
an angle 6 with the normal to the plane of the molecular current, then :

Mi' = HA cos e,

where A is the area of the molecular circuit. Hence :

li + HA cos ^ = lio.

Diamagnetic substances dififer from magnetic in that in the former
there are no permanent molecular currents. Hence for diamagnetic
substances io »0, and we have for the value of the induced current:

HA ^

1 = z- COS 0.

Li

The magnetic moment, /i^ of this current is expressed by:

HA*

M = iA = — cos 0;

L«

and the component of this parallel to H by

HA«

M cos ^ — COS* 0.

Lt

If there are n such molecular currents per unit volume with their axes
distributed at random, the number of axes lying between 0 and 0 +

d0 will be, as before, - sin d d0.

Hence the resultant magnetization per unit volume will be given by:

HA*

-/: - °-

2L
1 n H A*.
3 L

cos* $smed$

> Maxwell, Treatiae II, §838.

12 EARLY MAGNETIC THEORIES: QUIMBY

and the diamagnetic susoeptibility per unit volume becomes:

^ 3 L

It is evident that Weber's theory of diamagnetism offers a satisfactory
fundamental explanation of the phenomenon provided that his assump-
tion of the existence of perfectly conducting channels about the
molecules be granted. This assumption, however, did not appeal
strongly to his contemporaries, as is evident from a remark by Tyndall
in the Bakerian Lecture for 1855 that, ''This theory, notwithstanding
its great beauty, is so extremely artificial, that I imagine the general
conviction of its truth cannot be very strong."

The discovery of the electron fiunished an adequate mechanism for
the verification of Weber's h3rpothesis, and some of the more recent
attempts to explain diamagnetism are nothing more than efforts to fit
this mechanism into the fundamental theory which Weber established.

MAXWELL'S MODIFICATION OF WEBER'S THEORY

It has been noted that Weber's theory fails to account for residual
magnetism. MaxweU introduced a new assumption designed to re-
move this deficiency by providing for a permanent alteration in the
position of equilibrimn of a molecular magnet.^ He s«ippo8ed that if
the deflection of the magnetic axis of a molecule under the action of
a magnetizing field is less than some fixed value /So, then it will return
to its original position on the removal of the deflecting force. If,
however, the deflection, /3, is greater than /So, then, when the external
field is removed the magnetic axis of the molecule^ will not return to
its initial position but will remain permanently deflected through an
angle /S-/3o. Incorporating this hypothesis into Weber's theory leaving
the remainder of it unchanged, MaxweU obtained theoretical magneti-
zation curves which exhibit the phenomenon of retentivity. But
while the main hysteresis loop of a ferromagnetic substance may be
roughly accounted for in this way, the modified theory fails to explain
the smaller loops which may be superimposed on this by only partially
removing the magnetizing field and then reapplying it. Furthermore,
a physical justification for the assumption of the critical angle /So as
well as for the controlling field D of Weber's theory seems to be lacking.

Maxwell made a further extension of Weber's theory by investigating
the diamagnetic effect which is, on the hypothesis of molecular currents,
sure to be present in all magnetic substances.

In the molecules of such substances the primitive current, io, will
be diminished by the action of the applied field so that we have, in
accordance with the analysis of the previous section,

1 Maxwell, op. dt., S4i4«

EARLY MAGNETIC THEORIES: QUIMBY 13

. . HA ^

1 =» lo — COS 6.

JL

The magnetic moment of the molecule is given by:

•A • A HA« ^

/I ^^ lA = loA — cos $,

L

and its component parallel to H by:

HA«

M cos ^ = ioA cos 0 =— cos ^d

JL

=" ioA cos ^

/. HA« \

HA

If — - is small compared with unity, /i = ioA, and we return to Weber's

HA

theory of magnetism. If — is large compared with unity, then

HA« ^^"^

M = r— cos* 6, and Weber's theory of diamagnetism foUows. It is

L«

evident that the greater the value of io, the primitive value of the
molecular current, the smaller will be the diamagnetic effect. More-
over, a large value of L will bring about the same result. In any
event, it follows that the intensity of magnetization should diminish
if the impressed field be made sufficiently great. Such an effect has
not been observed, but it is evident that it wiU be very small and the
experimental difficulties which must be overcome in order to detect it
correspondingly great.

EWING'S THEORY OF RESIDUAL MAGNETISM AND HYSTERESIS

The accurate and extensive researches of H. A. Rowland^ and others
definitely established the inadequacy of existing theories to explain
h3rsteretic phenomena in iron and other ferromagnetic substances. In
attacking the problem Ewing discarded the arbitrary postulates re-
garding the controlling field and angle of. permanent set, and endeavored
to account for the magnetic behavior of these substances by investigating
the effect of the constraint which the molecules exert upon one another
by reason of the fact that they are magnets.'

Consider, for simplicity, a group consisting of two equivalent mole-
cular magnets, free to rotate about fixed centers. (Fig. 2) In the
absence of any disturbing force the two molecules will arrange themselves
with their magnetic axes coincident with the line joining their centers.
If an external field, H, be appUed which makes an angle 0 with this
line the two magnets will each be deflected through an angle ^, seeking

i PhU. Mag. 46 (1873). p. 140. 48 (1874), p. 321.

• Ewing, " Magnetic Induction in Iron and other Metala," p. 287.

14 EARLY MAGNETIC THEORIES: QUIMBY

a new position of equilibrium for which, evidently,

2 m H r sin (^ - «) = m« CN/PQ»,

where m is the pole strength of the magnets, and 2r is their length.

This position of the molecules corresponds to the initial stage of ihe
magnetization in which there is a small increase in induced magnetism
with increasing external field.

When:

the equilibrium becomes neutral and any further increase in H will
result in instability. The magnets will then swing violently toward
a new position of equilibrium with their axes nearly parallel to H.
This sudden shift corresponds to the second stage in the nuignetisation
in which a large increase in magnetic moment accompanies a small
increase in the magnetizing field.

Any further increase in H will not appreciably alter the positions of
the molecules and we have the condition of approximate saturation.

It remains only to note that if H is now decreased the magnets will
not retrace the same path in returning to their original positions. The
deflection accompanying a small decrease in H will be small until a

^ second state of instability is reached,

when they will swing back into posi-
tions approximating the initial ones.

A single pair of magnets of this sort
would give a discontinuous hysteresis
loop. If, however, we inuigine a large
nmnber of such elements with their
axes initially distributed at random it
is evident that some of these will reach
— -H the position of instability earlier than

p 2 others, and the '^ magnetization curve"

of the aggregate will be a smooth one.
"Hysteresis loops" have been obtained experiment^y with a group
of only twenty-four magnets, which are in perfect qualitative agree-
ment with those observed for iron.

The theoretical retentivity of a substance may be obtained by
assiuning it to be composed of a large number of groups, with the mole-
cules of each group arranged in some sort of symmetry. This is in
agreement with the fact that iron and other magnetic metals are
known to be composed of minute crystal matrices of the cubic system
irregularly oriented throughout their mass.

^

— H

• *

• /
I*

EARLY MAGNETIC THEORIES: QUIMBY 16

It is characteristic of such a cubical formation that the permanent
deflection of the molecules must necessarily be either 0^, 90^, or 180^.
Referring to Fig. 2, it is clear that if ^ be the angle of permanent de-
flection, we have three cases to consider:

(1) Molecules for which 6 is less than 45^. These will su£Fer no
permanent deflection. This is because the original lines are more
favorably directed than lines at right angles to them. For these mole-
cules (t> ^^ 6.

(2) Molecules for which 6 is greater than 45^, and less than 135^.
These wiU be permanently turned through one right angle. In this
case * = d - 90^

(3) Molecules for which B is greater than 135^. For these molecules
0 « ^ - 180^

If the axes of the molecules are initially directed at random, we have,
as before, for the number of molecules whose axes lie between 6 and
^ + d^,

-smBde,

and if the nuignetic moment of each molecule is fi, the contribution
of these molecules to the net intensity of magnetization will be

iin
2

flenoe the whole residual magnetism wiU be given by:

1 = ^/ sin^cosdd + ^/ sin« ^ d

4

2 J iw

+ ^ / sin ^ cos (^-180**) AS = 0.8927 /in.
2 J iw

4

More recent researches seem to indicate that the behavior of the
magnetic elements in crystals is not as simple as Swing's theory would
lead us to believe. The theory is, however, a step in the right direction,
for it attacks the problem which is fundamental in the explanation
of ferromagnetism, namely, the evaluation of the mutual actions of
the elementary magnetic units.

In the preceding review we have not considered the various theories
of magnetostriction which belong to the period under consideration.
A discussion of these theories will be found in a later section of this reports

i p. 225.

16 PARA- AND DIAMAONBTISM: WILLS

PROGRESS IN THE DEVELOPMENT OF THEORIES OP
PARA- AND OF DIAMAGNETISM FROM 1900 TO 1920

Bt a. P. Wills
Professor of Mathematical Physics, Columbia University

CONTENTS

Introduction 16

I The electric and the magnetic field due to a moving electron 19

II The magneton 23

III The distribution function in theories of paramagnetism 37

IV Early attempts at electron theories of magnetism 48

V The theory of Langevin 55

VI Modifications of the theory of Langevin independent of quanta hypotheses 68

VII Theories of paramagnetism based on quantum hypotheses 85

VIII Diamagnetism in metals due to the motion of free electrons 103

INTRODUCTION

The development of theories of magnetism during the period which
the present survey attempts to cover is characterized by successive
efforts on the part of theorists to explain magnetic phenomena in terms
of the properties of electrons in motion.

Early in the period under review it was found that the assumption
of motions of electrons in independent closed orbits in a material body
was incompetent to produce a satisfactory explanation of magnetisation
in the body. Some type of sub-molar structure of electrons was found
to be needed. For convenience we shall designate such a structure a
'^ magneton." The electron theories of magnetism to be reviewed are
naturaUy differentiated through the more or less arbitrary structural
properties assumed for the magneton.

Any molecular theory of magnetism is, of course, essentially statistical
in character and therefore continually faced with the weU known dif-
ficulties of statistical mechanics. These difficulties assume rather
formidable proportions in a theory which claims a generality sufficient
to account for magnetic susceptibiUties observed at low temperatures.
For it then appears that the theory has to part company with the law
of equipartition of energy of classical mechanics and introduce in its
place a law of distribution of energy among the magnetons of a body,
depending upon some more or less plausible quantmn hypothesis.

A primary object of all magnetic investigations on material bodies
is, of coiurse, to find out as much as possible concerning the nature of

PARA' AND DIAMAONETISM: WILLS 17

the magneton. So far as we know it cannot be segregated and ex-
amined; and our empirical knowledge of the magnetic properties^of
a material body is of necessity derived from an experimental study
of its magnetic quality in bulk. It is the bulk susceptibility which
is experimentally determined. This is a statistical quantity, repre-
senting the contributions of the statistical units, the magnetons, to
the magnetisation of the body in bulk. In the consideration of
any molecular theory of magnetism it is therefore necessary to bear
in mind that the theory may weU stand the test of experiment, and yet
the model of the magneton which it assumes be far from a true one,
since different types of magnetons might have the same statistical
properties.

As far as fundamental physical ideas are concerned the reader of
the following report wiU probably conclude that the interpretation
and the extension of old conceptions, those of Ampere and of Weber,
rather than the introduction of new ones, save those relating to quantmn
theories, characterize in general the developments in molecular theories
of para- and of diamagnetism during the years from 1900 to 1920.

The development of electron theories of magnetism which began
early in the period covered by the present report was stimulated in
large measure by the theoretical writings of F^ofessor H. A. Lorenta
and of Sir Joseph Larmor. Their results constitute a lai^e part of
what is now termed classical electron theory with which the reader is
supposed to have some acquaintance.

Kinetic theories of magnetism are of necessity somewhat mathematical
in character and the pages of the literature dealing with them are often
encumbered with many rather formidable appearing formuke, which,
while oftentimes necessary, operate as a deterrent to the average reader,
who is more interested in the physical content of a theory than in the
mathematical dress in which it is clothed.

With the object of divesting, so far as possible, the various theories
discussed below of the mathematical features which are shared by many
of them in conunon the first three sections have been written.
These sections are intended more for reference during the reading of
the rest of the report than for continued perusal. The reader who so
desires may therefore begin with Section IV, dealing with early attempts
at electron theories of magnetism.

For the purposes of the present review it has been found convenient
to use a vector notation. That of Gibbs has been adopted.

Vector quantities are printed in the heavy Bookman type — A, B,
a, b . . and the corresponding scalar values in ordinary t3rpe — A, B,.
a, b . . .

18 PARA' AND DIAMAGNETISM: WILLS

The reader who is unfamiliar with Vector Analyofl and who desires
to follow those parts of the argument in the text in which vector methods
are used will find ''Vector Analysis" by J. G. Coffin a very convenient
book for reference.

As regards units, for electric and magnetic quantities the Gaussian
4system is used throughout. For other quantities c. g. s. absolute units
are always used. To denote the velocity of light the letter c is used.

">

t

PARA' AND DIAMAGNETISM: WILLS

19

THE ELECTRIC AND THE MAGNETIC FIELD DUE TO A MOVING

ELECTRON

The explanations of magnetisation on the theories of magnetism
which we shall notice later are referred back to the electric and magnetic*
properties of free electrons in motion, or to the corresponding properties^
of some sort of rotating magneton.

In the present section we shall therefore consider the electric and
the magnetic field of a moving electron; and in Section II we shall
consider the electric and magnetic properties of rotating magnetons,
and also the mechanical moments to which they are subject when

placed in an external electric or magnetic field.

The electron, considered as a point charge,

will at first be considered to be moving in any

arbitrarily assigned manner. The electric and

the magnetic force due to the moving electron

may be calculated for any field point directly

from its retarded scalar- and vector potentials..

Referring to Fig. 1, O represents the origin

of a cartesian S3rstem of axes fixed in space; Q

^ the position of the electron at the instant under

consideration; s the position vector of Q with

reference to O; P the field point; r the position

vector of P with reference to O; and q a vector

drawn from Q to P.

The cartesian coordinates of Q and P are represented respectively

by f , ri, f and x, y, z. From the figm^ :

Fig. 1

(1)

= V (x - {)« + (y ^ r,y + (z - f).»

If e be the charge on the electron and v its velocity, then by classical
electron theory the scalar- and the vector potentials at the field point
P are respectively expressed by:

(2) *-

K'-m_3

A =

ev

h(-Tj')J

.-9

c

where the quantities in square brackets are to be evaluated not at the
time at which the electric and the magnetic forces are required but
at a time previous by the interval required for radiation to travel
from the point Q to the point P, that is at a time t — q/c.

The potentials having been evaluated the electric and the magnetic
force at the field point will be given respectively by:

•20 PARA- AND DIAMAGNETISM: WILLS

(3) E - - V * — ; H = curl A.

e at

Upon carrying out the operations here indicated the following ex-
pressions for the electric and the magnetic force at the field point P
are foimd:

q*

c*L q* \ cq/ cq* \ cq/ J

The details of the calculation are somewhat involved and may be
found in standard treatises dealing with electron theory, e. g., in The
Theory of Electricity by G. H. Livens, p. 506.

For the cases which will come under our consideration the velocity
of the electron may be considered small in comparison with that of
light, and the field point may be chosen so that its distance from the
electron is small in comparison with the wave length of the radiation
emitted by the electron. The general expressions for the scalar and
the vector potential given by (2) then reduce to the ample ap-
proximate expressions:

e ev

(6) * = - , A = - •

q cq

It may be noticed that here the potentials are not retarded.

The corresponding expressions for the electric and the magnetic
force due to a moving electron may be obtained directiy from (6) by
taking the negative gradient of ^ and the curl of A. It is thus found
that:

<7) = = ^s « '

(8) H = — V X q.

cq'

These approximate equations might have been obtained, of course,
from the general expressions (4) and (5) by introducing the restrictions
above made.

If the origin O be so chosen that s is small in comparison with r,
the quantity 1/q in the expressions for the potentials may be developed
in a series in which only the first three terms need be retained:

\-i{^-f-im

PARA- AND DIAMAGNETI8M: WILLS 21

Inserting this expression for 1/q in (6) we find:

(9)

-^{(-^H.|(f)'}.

Taking the negative gradient of ^ and the curl of A we now obtain
the following expressions for the electric force and the magnetic force
at the field point:

(10) E . 1 {(l + ^■) (. - .),}.

(.1) H-^.x{(l+i£^)(r-.)}.

The mean value of H for an electron describing a circular orbit with
constant speed will later be required. If H denote the mean value of
H for this case, it is easily foimd from (11), upon observing that v
= s, that:

(12)

H = - — :(3 s X s- -r — s xs J.
2cr*\ r* /

Thus, an electron describing a circular orbit with constant speed is,
as far as its mean magnetic field is concerned, equivalent to a small
magnet whose moment, t^, is given by:

(13) V = - S X 8.

This expression can be put in a somewhat simpler form as follows.
Let <d be the angular velocity of the electron about the center of its
orbit, then s = <dXs = a)nxs, ifnbe a unit normal to the plane of
the orbit in the direction of u. We now have :

sxs = a)sx(nx8) =a)S-sn= |Sn,

where r is the orbital period, and S the orbital area. Then, from (13) :

(14) »=^n.

Cr

The moment of the orbit wiU be subject to change if a magnetic field
be created through it. Let H be the strength of the magnetic field at
any instant and E the corresponding electric force. Supposing the
area of the orbit, S, to be very small and its plane invariable.

22 PARA- AND DIAMAQNETI8M: WILLS

by making use of Stokes' theorem and Maxwell's field equation, curl
E « — d H/c d t, we obtain:

^ S S

Upon integration the integral on the left gives 2x s E, and hence :

2tsE » - — (n.HS);

cdt

the expression on the right representing the time rate of decrease of the
magnetic flux through the orbit. If A (n-H S) denote the increment
of this flux in the orbital time r, then:

(15) 2 T s E = ^^ '

cr

Again, since the moment of the force e E must equal the time rate
of increase of the moment of momentum of the electron in its orbit,
we have:

„ d, ,. 2mcdfi

seE= — (mB««)= — ;

at eat

consequently, if A /i denote the increment in fi in the orbital time, r:

,-^v ^ 2 m c A/i

(16) B e E = - .

e T

From (15) and (16) it follows that:

e^ e'

(17) A M = ; A (n.H S) = A (H S cos 6),

where 6 is the angle between the directions of n and H.

Mechanical Action upon a Moving fUectron in an External Electro-
magnetic Field.

If E and H now denote respectively the strength of the external
electric and magnetic field, from fimdamental electron theory we have
for the mechanical force, F, upon any electron:

(18) F = eE + - V X H;

c •

and for the mechanical moment, N, of this force about the origin O,

(19) N = es X (E X - V X H).

PARA- AND DIAMAGNETISM: WILLS 23

II
THE MAGNETON

Since, as mentioned above, the assumption of motions of free elec-
trons in independent orbits is incompetent to lead to a satisfactory
explanation of magnetisation, the concept of the magneton made an
early appearance in modem theories of magnetism.

The magneton is conceived to be a minute aggregate of positive and
negative electrons, possessing certain arbitrarily assigned constitutional
or structural properties. We first consider these properties.

Fundamental Assumptions Concerning the Structural Properties of the

Magneton.

The algebraic sum of the charges of the electrons in a magneton are
assumed to be zero. If the charge on a typical electron be e the struc-
tural condition implied by this assumption is expressed by writing:

(1) Se = 0.

The distribution of the electrons in the magneton is supposed to
be such that the electric moment of the magneton is zero. We now
suppose that the typical electron of the magneton is the electron of
Section I, and that the origin O coincides with the centroid of the
magneton. Then (see Fig. 1, Sect. I) the condition that the electric
moment of the magneton shall be zero is expressed by:

(2) Ses « IDeJ + jSei? + kZef = 0,

where i, j, k are imit vectors in the directions of the axes (x, y, z) respec-
tively.

It wiU appear presently that the electric and magnetic properties of
the magneton depend in an important way upon the following quantities
of the second degree in (, 17, T*

(3) Pi = 2)6?, Pi^Dei?^, P8 = Sef«,

(4) Di - Zeiyf, D, « Sff, D, = SJiy,

(5) Qi = P2 + P», Q« = P» + Pi, Qs = Pi + Pi.

From the analogy of these quantities with corresponding quantities
in mechanics it is appropriate to call the Q's and D's respectively
Moments of Inertia of Charge and Products of Inertia of Charge.

The Electric Potential and the Magnetic Potential for a Rotating

Magneton.

In the applications of the present theory with which we shall be
concerned in our review of theories of magnetism the velocity of any
electron will be small in comparison with that of light, and the distance

24 PARA-' AND DIAMAQNETI8M: WILLS

of the field point from the magneton will be small in comparison with the
wave length of the radiation emitted by it and yet large in comparison
with the dimensions of the magneton.

The appropriate equations for the potentials will therefore be fur-
nished by the equations (9), Sect. I, (for the potentials of a sin^ electron)
through summation over aU the electrons in the magneton, the origin O
being supposed at the center of the magneton. We thus obtain for
the electric and the magnetic potential of the magneton respectively:

^ «e / r.s . 3/r.sVl

c r I i« 2\ r« / J

where v is the velocity of an electron, s and r the position vectors of
the typical electron and the field point, respectively.

Taking account of the structural conditions given by (1) and (2)
these expressions reduce to:

^ ^ ^ 1«« /r-8 . 3/r.sVl.

In general, the approximation will be suflScient if only the first term
in the expression for A be retained; then:

(8) A = - 2 e V — .

The right hand member of this equation may be transformed as
follows — noting that v » s we have identically:

^{vr.8-8r.v + |(88.r)|-^rx(7X.)+ii(88.r);
and, therefore:

where

(10) V- — 2esxv.

The Mean Value of the Vector Potential for a Rotating Rigid

Magneton.

For a rigid rotating magneton the mean value during one revolution
of the second term on the right of (9) will vanish, and if we denote the
mean value of A by X and of t^ by ^, then:

(11) A - i^

▼ r.s ^
2

PARA- AND DIAMAONETISM: WILLS 25

Ftom the form of this expression for the mean vector potential it
appears that the mean field of a rotating magneton is the same as the
field of a small magnet with a moment t^; and it may be easily seen that
the direction of the vector ^ will coincide with that of the axis of rotation
of the magneton.

If a, /3, 7 be the direction cosines of the axis of rotation, and therefore
of tf , the scalar components of the mean vector potential of the magneton
will be given by:

Ai = - OSz - 7y),
(12) A, - ^, (7X - OS),

Ai - - (ay - /3x).
r*

The Mean Value of the Magnetic Force Due to a Rotating Rigid

Magneton.

Taking the curl of A ¥^ find for the mean value, H, of the magnetic
force:

(13) H-^,5 rr-^»;

and the scalar components of this force are easQy seen to be given by :

Hi = 3-rax + ftr + 7z)x-^'
r' r*

(14) Hi»3-,(ax + ftr + Tz)y-^'

r r^

H, = 3-.(ax + ftr + 7z)z-3•
r* r*

It appears from these equations that the mean magnetic field is
symmetrical to the axis of rotation of the magneton; that the lines
of force Ue in planes through the axis of rotation; and that the mean
field is equivalent to the field of a magnetic doublet whose axis is parallel
to the axis of rotation and whose moment is equal to «f.

This equivalence, of course, hol(is only for the mean value of the
magnetic field of the magneton, and not for the instantaneous value.
For the latter the second term on the right of (9) comes into considera-
tion; and accordingly the instantaneous value of the field will vary
with the time, giving rise to radiation, with which, however, we are not
B specially concerned.

26 PARA- AND DIAMAGNET18M: WILLS

The Magnetic Moment of a Rotating Rigid Magneton

As has been seen above the mean value of the quantity t^ represents
the mean time value of the moment of a rigid rotating magneton; it
will therefore be convenient to refer to the quantity t^ itself as the
moment of the magneton.

When the magneton is rigid, v = u x s, and we have from (10) :

I^=~2)e8x(«x8)

(16) «--2e(B»« - tt-ss)

2c

iM

(? + 1|^ + f*) « - («1 f + «|1| + «8f) 8.

where (, fi, f are, as usual, the scalar components of s, and ah, «i, wi
are the scalar components of u, the angular velocity of the magneton.
From the last of these equations it follows, with the aid of (3), (4)
and (5), that:

^-i{

(Qi«i - D,w, - I>,»,)i

(16) +(- D,«i + Qiw, - D,«,)j

+(- DiG)! - Di«, + Qi«i)kV

From this equation it appears that t^ is a self-conjugate linear vector
function of u. In fact, the relation between v ftnd u is precisely analo-
gous to that of the moment of momentiun of a rigid body to its angular
velocity of rotation, the Q's in the present case corresponding to the
moments of inertia about the axes and the D's to the so-called products
of inertia.

The Torque upon a Magneton due to an External Electromagnetic Field.

We now suppose the magneton to be placed in an electromagnetic
field which may vary in space and in time. The electric force and the
magnetic force of this field will be denoted respectively by E and H.
The torque, N, acting upon the magneton due to the action of this
field has now to be found.

With reference to the origin O, this torque, from (19) Sect. I, will
be given by:

(17) N = Se8x(E-f-vxH),

c

where the summation is over all the electrons in the magneton.

PARA' AND DIAMAONETISM: WILLS 27

Since E and H may be assumed continuous, they may respectively
be developed into the series:

<18) E = E^+(8.VE)o + ,

<19) H = H,+ (s.VH)^ + ,

where the subscripts indicate that the quantities to which they refer
are to be evaluated at the point O.

If N' and N'" denote respectively the turning moments upon the
magneton due to the external electric and magnetic force, then:

<20) N « N* + N."

In the evaluation of N^ attention must be paid to (2), expressing that
the total electric polarization of the magneton vanishes. On this
accoimt the first term on the right of (18) contributes nothing to the
value of N.^ If furthermore we restrict ourselves to terms of the
second order of smallness in the small quantity s, only the second term
in the development of E need be considered and the evaluation of N*
then gives:

(21) ir = Se8xsVE,

where it is to be imderstood that the derivations in the factor s . VE are
to be effected at the point 0 although the zero subscript is not explicitly
carried forward.

In a similar manner the evaluation of N"" to the same order of approxi-
mation gives:

(22) N" = 52esx(vxH),

where H is to be taken as the external magnetic force at O.

If the triple vector product in the sum on the right of (22) be expanded
and accoimt taken of the perpendicularity of s and ▼, it may be seen
that (22) transforms into:

N" = ZesHv;

or, in case the magneton is considered as rigid :

(23) N" = ^«xc,

where u is the angular velocity of the magneton about an axis through O
and:

(24) c^ZesHs.

The scalar components of the vector c with the aid of (3) and (4),
remembering that (, 17, f are the scalar components of s, may be ex-
pressed as follows:

28 PARA' AND DIAMA0NETI8M: WILLS

Ci = PiHi + DA + DtHi,

(25) C = DA + PiH, + D,H„

C, « DtHi + DiH, + P,H,,

showiBg that c is a self-conjugate linear vector function of H.

Making use of (3) and (4) tiie scalar components of IT given by (25)
may be expressed by:

Ni* = D,— - + ?,— -• + Di— ' - Di^ - Di^ - P,~'
dx dy dz dx dy dz

(26) N,' = Di^ + Di^ + Tt-f^ - Pi—' - D,— • - Di-— ''

dx dy dz dx dy dz

N, - Px- + D.- + Di-^ - D.— - P.- - D.-

Ftom (23) with the aid of (25) the corresponding expressions for the
scalar components of the turning moment upon the magneton due to
the external magnetic field are seen to be given by:

Ni~ = -V«,(D,Hx + DiH, + P,H,) ~ «,(D,H, + P,H,+DiH.)l,

(27)

N," = ^|«,(PiHi + DA + DiH.) - ch(D^i + DiH, + P,H,)l,
N." = ^|«i(D A + P«H, + DiH.) - a),(PiHi + DA + DiH,)|-

Equations for a Rotating Rigid Magneton Referring to its Principal

Axes of Charge.

It is always possible to choose three mutually perpendicular axes
through the centroid of a magneton such that for them the products of
inertia of charge vanish:

Di = Di « Di « 0.

These axes are called Principal Axes of Charge.

The equations foimd above for the magnetic moment of a rigid mag-
neton and for the scalar components of the torques upon it due to the
action of an external electric and an external magnetic field assume much
simpler forms when the axes of reference are Principal Axes of Charge.

Thus^ from (16), we have for the magnetic moment of a magneton:

(28) II = - (Qioni + Qm j + Q««*) ;

zc

PARA' AND DIAMAGNBTISM: WILLS 29-

and, from (26) and (27), for the scalar components of the torques:

(29) N,- = P.f^ - P.'^'.

due to an external electric field E; and :

Ni" = - (P«H*»« - P,H*«,),
c

(30) N," = -(P,Hx«, - P,H,«0,

c

N,» = -(PiHiWi - PiH,«*).
c

due to an external magnetic field H.

Equations for Rotating Spherical and Axial Magnetons Referring to

Principal Axes.

For the purpose of the present review it is only necessary to consider
two special types of magneton, known respectively as the Spherical
Magneton and the Axial Magneton.

The Spherical Magneton is defined as rigid and one for which the
principal axes of charge and of inertia coincide and for which the prin-
cipal moments of inertia of charge Qi, Qs, Qs and the principal moments
of inertia, A, B, C, are respectively equal:

(31) Qi = Q, = Q, = Q .-. p, = p, = p,;

A =B =C =J.

The Axial Magneton is defined as rigid and one for which the prin-
cipal axes of charge and of inertia coincide and for which the principal
moments of inertia of charge and the principal moments of inertia are
respectively equal for two of its principal axes, say 1 and 2:

m^ Ox = Q. = Q, ••• Px = P,;

<32) A = B = J.

For the magnetic moment we have, from (28) :
(33) » = - «

for the Spherical Magneton; and:

30 PARA' AND DIAMA0NBT18M: WILLS

<34) t^ - ^(Q«ii + Qwij + Q,»»k)

for the Axial Magneton.
For the torque due to an external electric field E we have, from (29) :

^* 2U a,;

*T. Q/5E, dE,\ „ Q ,«
<^> N.--|(---)..Mr-^curiE;

^''2\dx dy)
«nd, from (30), for the torque due to an external magnetic field:

Ni" = |(«*,H, - «A),
<36) N," = ^(«A - «,H,), .% N- - J« X H,

N,- = |(«iH,-«,Hi);

or a Spherical Magneton.

From (29) and (30) the corresponding expressions for an Axial Mag-
neton will be given by:

<»7, W - P. f - Sf •

«.-f(f-f)-f<-«..
for the torque due to an external electric field E; and:

N.- - ?(PA-. - fa^)

N,- - ^(E^ - H,«,) - ^(« X H),,

ZC iSC

lor the torque due to an external magnetic field H.

\

PARA' AND DIAMAONETISM: WILLS 31

The Rotary Motion of a Rigid Magneton Subject to an External Elec-
tromagnetic Field.

We assume the reference axes to coincide at the instant xmder con*^
cdderation with the principal axes of inertia of the magneton, for which
the moments of inertia are A, B, and C ; and also that the principal axes,
of charge coincide with those of inertia.

By Eukr's dynamical equations of motion :

Acii - (B - C)«j«, = Ni' + Nr,

(39) Bci, - (C - A)«,«i = W + N,",

Cci, - (A - B)«i«i = Ni' + Ni",

where the N* and N"' torque components in the general case are given
by (29) and (30).

Ftom these equations the rotary motion of the magneton may be
theoretically determined when no dissipative forces are assumed.

Special Case of the Spherical Magneton.

In this case we have A = B=:C«J and, upon introducing the
expressions for the torque components given by (35) and (36), in the
equations of motion (39) it appears at once that they are equivalent to
the single vector equation:

(40) Jii « %curl E + Kx H);

2 c

Since, by virtue of one of Maxwell's field equations, curl E "^
— dH/cdt, this equation may be written:

^^'^ -^dt^-^U"""^}

This equation assumes a simpler form if the time derivations are
taken with respect to the moving space of the magneton instead of

fixed space. If 37 denote time derivation with respect to the former^

at

then:

d tt dtt d H dH ^

and equation (41) may therefore be written:

d « Q d'H

(43) J^-^ = - -5L )L^

^^ dt 2c dt

Integration of this equation gives:

32 PARA' AND DIAMAGNETISM: WILLS

where u, denotes the value of <d before the application of the external
field.

From the last equation it appears that the establishment of an external
electromagnetic field brings into existence a rotation of the magneton
about an axis parallel to the lines of force of the external field of amount

will depend upon whether Q is negative or positive.

The magnetic moment of a spherical magneton is given by (33),
from which with the aid of (44) :

Qi
(45) ,^ « 1^^ - 4^^*

Therefore the effect of the establishment of the external field upon the
moment of the magneton is to bring into existence a component
— Q'H/4c'J directed parallel to the lines of force of the external mag-
netic field; since J is a positive quantity the coeflident of H in (45) will
be negative.

In the particular case where the external magnetic field remains con-
stant in time the equation of motion (41) for a spherical magneton
reduces to:

<46) J« - ^« X H.

In accordance with this equation, since u x H is a vector which is
perpendicular to w, the magnitude of !■» will remain invariable; but,
except in the special case where » is parallel to H, the direction of the
axis of rotation will continually change both in fixed space and in the
magneton. The component of <d in the direction of H will not change
but the component perpendicular to H will rotate about the direction
of H with the constant angular velocity

(47) "• - - ^-

The vector u itself will rotate about an axis parallel to H with this
same angular velocity; and the magneton will perform a reg^ular pre-
cession about this axis. From (44) and (47) :

(48) « = «o + «!•

The angular velocity w of the magneton may thus be regarded as the
sum of two components; cjo, representing its angular velocity before the
application of the external field, and ui, representing an induced angular
velocity about the direction of the lines of force, due to the creation of
the external field.

PARA' AND DIAMAONETISM: WILLS 33

The precessional motion of the magneton takes place in a manner
similar to that of a synmietrical top in a gravitational field, but with the
di£Eerenoe that the applied torque in the present case, QuxH/2c, is
proportional to the angular velocity, while in the case of the top it is
independent of the velocity; thus it comes about that the precessional
velocity, in the case of the magneton is independent of its angular
velocity, while in the csuse of the top it is inversely proportional to the
angular velocity.

Since, by virtue of one of Maxwell's field equations, curl E == 0 f or
a magnetic field of constant strength, it follows from (35), (36) and
(38) that the torque on a spherical magneton in a constant external
magnetic field is t^ x H or, on accoimt of (45) :

(49) »o X H.

The magneton is thus subject to a couple equal to that which would
be experienced by a magnetic needle of moment yo placed in the same
magnetic field H. But the motion of the needle would be quite different
from that of the magneton, in that the needle would move in a plane
containing its axis and parallel to the lines of force, while the magneton,
due to its gyroscopic properties, performs a precessional motion about
the direction of the lines of force. If either the needle or magneton is
to assume a position with axis along the lines of force it is necessary
in general that dissipative forces come into play.

Special Case of the Axial Magneton.

For the axial magneton A — B — J and the general equations of
rotary motion (39), with the aid of (32) reduce to:

J^ - (J-Oo^, = ^ ^' - P.'^ + -f P,«.H. - %M,) '

2 dy dz c\ 2 /

(50) Jci, - (C- J)«^ = ?»?*-% V-* + Y%H» - P»"iH»V

dz 2 dx c\2 /

«■ - Kf - f ) + *'-«• - "-«■'•

The third of these equations refers to rotation about the axis of the
magneton and may be put in the form:

(51) Ci, = |*|(curl E), + ^(« X H),V
By virtue of one of Maxwell's field equations:

(curlE), = ---jr;

c at

34 PARA' AND DIAMAGNETISM: WILLS

BO that (61) may be written:

da)s

Qs /dH \ ;

2cCV dt /,

dt

or, if the derivations be taken with respect to the moving space of the
magneton,

d 0)3 Qs d Hs

"dt "" " 2cC^

Integration of this equation gives:

(52) Wj = 0)03 - 2^^»'

where €003 represents the angular velocity of the magneton about its
axis before the application of the external field. The external field
thus produces a change in the angular velocity about its axis of amoimt
— Q3Ha/2cC. It wiU also produce changes in the velocities of rotations
about two perpendicular equatorial axes the equations of which are
the first two of equations (50), assuming no dissipative forces. Owing to
radiation due to the disynmietry of structure of the magneton with
respect to these axes the motions about them would in course of time
be damped out leaving only the motion about its axis.

To the latter there corresponds a magnetic moment which, from
(34) and (52), will have for its scalar value:

(53) M = Mo3 - ^^^^^

where

(54) tu>z = ^"03;

Aioa is the scalar value of the axial component of the moment of the
magneton before the application of the external field.

Energy of a Rotating Axial Magneton in a Constant External

Magnetic Field.

In what folbws the axial magneton will be supposed to consist of
a rigid system of negative electrons symmetrically spaced about their
centroid and rotating about it, the corresponding positive charge being
in the form of a nucleus at their centroid or of a concentric sphere.
In this case we may write in equation (34) for the magnetic moment of
the magneton:

Q = eJ/m and Qs = eC/m.

For the total energy, U, we may write:

U = Ui + U, + U,,

PARA' AND DIAMAQNETISM: WILLS

35

where Ui lepiesents the energy due to the translatory motion of the
magneton, Ui its energy of rotation and Us the mutual energy of the
magneton and the external field, which, according to the point of view,
may be regarded either as kinetic or potential.

If M denote the mass of the magneton and x, y, z the coordinates of
its centroid we have for its translatory energy:

(55)

Ui = y (i« + ^ + 2«)

In the calculation of the rotatory energy of the magneton we suppose
((y i7y r) to be axes coinciding with its principal axes and therefore fixed
in the magneton, A, B, C being its moments of inertia about the axes
of (, 17, i respectively; since the magneton is now supposed axial, we put
A = B = J.
To specify the position of the magneton with reference to the external

field and fixed space we use Eulerian
angles 0, ^, 0.

Referring to Fig. 2, 0 is the angle
between the positive directions of
the external field H and the axis f;
^ is the longitude of the line of
nodes, on defined as a line per-
pendicular to the plane determined
by the directions of the field H and
the f-axis; and 0 is the angle be-
tween the line of nodes and the
f-axis.
If 0)1, 0^, 0)1 be the scalar compo-

r

Fig. 2

nents of the angular velocities of the magneton about the axes (, 17, f ,
respectively, then:

coi — ^sin0sin0+0oos0,
.(56) ctf|sj^sin0cos0 — dsin0,

0)3 = ^ cos 0 + 0.

We shall therefore have for the energy of rotation of the magneton:

(57)

u, = ^(^ + i^ sin« ^) + ^(0 + ^ cos ey.

Considering the mutual energy Us of the magneton and the external
field as kinetic we may write^-

(58)

U.-i»H.

1 Cf. R. GaoB, iifm. d, Phyt. 49, p. 164; 1916.

36 PARA" AND DIAMAONETISM: WILLS

where i^ is the xnagDetic moment of the magneton. Upon noting that
Q ^ eJ/m and Qs » eC/m where m is the majBS of a constitutive
electron of the magneton, it follows from (34) that:

(59) V = ^ (''"I* + J«ij + C«»k).

Zmc

Upon taking the scalar product of H with this expression for i^, sub-
stituting the expressions for 0)1, (at and <ai given by (56) and inserting
the resulting expression in (58) we obtain:

(60) Us = * v-H = :^H{J^ sin* ^ + C(0 + ^ cos d) cos $].

4mc

Finally, upon adding the expressions for Ui, Us and Us given by (55),
(57) and (60), we obtain for the total energy of the axial magneton in a
constant magnetic field:

M

(61) U = -(x' + y' + i*)

+ ^(^ + ^ sirf e) + ^(0 + ^ cos $y

+ '

-^h/j^ sin* ^ + C(0 + ^ cos d) cos e\.
4mc ( j

In the writing of the present section the treatment of the subject of
the magneton as presented in Abraham's "Theorie der Elektrizitat"
has been of much assistance.

PARA- AND DIAMAONETISM: WILLS 37

III
THE DISTRIBUTION FUNCTION IN THEORIES OF PARAMAGNETISM

In kinetic theories of magnetism the problem of the determination
of the distribution of the axes of the constitutive magnetons of a body
placed in an external magnetic field arises. A knowledge of this distri-
bution is necessary before the contribution of the magnetons to the
resultant magnetic moment due to the action of the external magnetic
field upon the magnetons can be calculated. For convenience of
reference later some results of statistical theory will be considered in
the present section.

Let us consider a S3rstem, subject to no external field of force, con-
sisting of a large number of like statistical imits, the t3rpical one of which
is specified as regards its configuration by the generalized coordinates

qi qr, subject also to the condition that the total energy of the S3r8-

tem is constant. Let the n generalized momenta of the system be
denoted by pi p^.

We suppose the generalized coordinates and the momenta to be
subject to statistical variation, through thermal agitation for instance.
Then if N be the niunber of units per unit mass, in accordance with
statistical theory, when the system is in a state of equilibriiun the
probable niunber of units per unit mass, say dN, which have values of
their coordinates and momenta lying respectively within the specified
limits

qi and qi -h dqi q„ and q„ + dq„,

Pi and Pi + dp, p„ and p„ + dp^,

wiU be expressed by the law of distribution:

(1) dN = ae'^dOf,

where

dl2 := dqi. . . .dq^dpi. . . .dpo,

c is the total energy of a unit which is subject to statistical variation
expressed in terms of the q's and p's and a and h are constants.
For the determination of the constant a we have the condition:

(2)

fae'^'dif = N,

where the integration is to be extended over all possible values of the
variables whose differentials appear in the expression for dQ\
The fimction

oe-^"^-

is called the distribution function for the system of units.

In some cases it may be convenient to introduce new variables in
place of some of the generalized momenta. Thus, let us suppose m

38

PARA' AND DIAMAQNETI8M: WILLS

of the generalised momenta, say pi pm, to be expressed in terms of m

new variables, say ri rm, through the equations:

Pi • fi(ri. . . .r^,

fai(ri. . . .r^).

By differentiation:
dpi

dp

9p\

■dri+ +Z~*''

ori dr.

m)

«P-+....+?Psdr..

dp.

dr,

dr.

From a theorem due to Jaoobi:

dpi. . . .dp„ ■• Adri. . . .dr„,

where A, the modulus of substitution, is given by the determinsntal
expression:

(3) A-

dpi
dr, • • • •

dpi

«P-
dr. •••

dP-
• •dr-

K we write:

do — dqi. . .dqndri. . . .drndpoH-i dpn,

then:

(4) do" - 6dSL

The law of distribution (1) is therefore equivalent to:

(5) dN « ae-**AdO,

where the energy c is now supposed expressed in terms of qi . .
Ti- . . .r^^ and Pn£fi p^.

q->

Case of a System of Axial Magnetons in a Constant External Magnetic

Field.

It will be assumed for the present that the density of distribution of
the magnetons is so small that the molecular field at any given magneton
due to the others is neglible. It will also be assumed that the accelerar
tions of the magnetons are so small that their loss of energy by radiation
may be neglected. Furthermore the restrictions, whereby Q=eJ/m

PARA- AND DIAMAGNETISM: WILLS 39

and Qs^eC/m, imposed upoD the axial magneton in the last part of
the preceding section will be supposed to hold.

The total energy of the system may then be considered as constant,
since the constant external magnetic field can do no work upon the
magnetons, the corresponding mechanical force upon the constitutive
electrons of the magnetons being perpendicular to their directions of
motion.

We may now take for the total energy of the t3rpical magneton of
the system the expression (61) Sect. II:

(6) u = ^(i« + ^ + i«) + ^(^ + ^ sin« ^) + ?(^ + ^ cos ey

+"•

|j ^ sin«^ + C(0 + ^ cos e) cos sX'

4mc(

where x, y, z are the coordinates of its centroid and $, ^, 0 ita Eulerian
coordinates.

From this expression, since the total energy is kinetic, by partial
differentiation we obtain for the corresponding generalized momenta,
say u, y, w, p, q, r, the following expressions:

u = Mx, V = My, w = Mz, p — Jd,

. •

eH

(7) q « J^ sin« ^ + C(«+^ cos $) cos $+-- — (J sin* B+C cos» 6),

4mc

eH

r = C(0 + ^ cos d) + ' — C cos e.

4mc

The statistical variables of the system are now x, y, z, u, v, w, 0,
^1 01 Pf Qi c^<l ^' But it will prove convenient to replace the momenta
P> <h f by new variables P, Q, R, using the following equations of sub-
stitution :

p = PcoB0 — Qsin0,

eH

(8) q- (Psin0 + QcoB0)sin^ + Rcos^+7— ( J sin* ^ + Cco^d),

4mc

r = R + -:^HCco8d;
4mc

from which by (3) we find for the modulus of transformation:

(9)

cos ^, — sin 0, o

sin 0 sin 0, sin 0 cos 0, cos ^

o . o . 1

= sin ^

40 PARA' AND DIAMAGNETI8M: WILLS

From (6), with the aid of (7) and (8) :

(10) u-2li(u* + V + W) + ^ + |

eH
+ z — (Pain^8in^ + QcoB08in^ + RcoBd).
4nic

We are now dealing with a system of statistical units, the magnetons,
which is subject to an external field of magnetic force, and the question
arises as to the form of the function t appropriate to this case. Gans,
in the paper cited above, has shown this to be equal to this expression
for U modified through multiplication of the last term by the factor
u; so that:

1 P'-l-O' R*

(11) -^("*+-*+-*)+^+ic

eH

+ „p(P sin ^ sin ^ + Q cos 0 sin ^ + R cos ^).

In this expression the coordinates x, y, z, ^ do not appear explicitly.
Therefore the law of distribution for the remaining statistical variables
will be independent of these coordinates. Furthermore the expression
involves the statistical variables u, v, w only as a sum of squares and
therefore, as a well known result of statistical theory, the law of dis-
tribution for the remaining variables will be independent of u, v and w;
moreover the constant h in the law of distribution will have the value
given by:

2kT

where T denotes the absolute temperature and k the gas constant for
a single molecule, known as Boltsmann's constant.

Now from the point of view of magnetic theory we shall be concerned
only with the law of distribution of the statistical variables B, 0, P, Q, R;
and the appropriate expression for t for this case is obtained from (11)
by simply ignoring the terms involving u, v and w.

If then dN now denote the number of magnetons per unit mass whose
statistical variables. By 0, P, Q, R, have values which lie within the
element of phase dQ given by:

dQ » d0d0dPdQdR,

the equilibrimn law of distribution for these variables will be expressed
by:

(12) dN-oe'^sin^dO,

PARA- AND DiAMAGNBTISM: WILLS 41

where

(13) €« ^^;J^+ -^ + ;^H(Pan08ind+^

2J 2L/ 2mc

We shall have occasion to consider another case in which the number
of statistical variables involved is still further reduced. In this case
the angular velocity <at of any magneton about its axis of symmetry is
considered constant and the same for all magnetons. This requires
that the quantity R shall be constant, since:

(14) R = C(^ + ^ cos d) = C«s.

Consequently R may no longer be considered as a statistical variable
and the statistical variables of the present case are therefore $, 0, P, Q.
The appropriate expression for the energy function for this case is now
required.
From (52) Sect. II, noting that Qs/C » e/m and that Ht » H cos tf :

«» = Was - ^H cos e.

Since R — Ccoa, we have, with the aid of this expression, for the sum of
the terms in (13) involving R:

(15) S,+ R;r-H cos ^ = iC«o,« - mH cos d,

2C 2mc

where

2mc

fi denotes the constant scalar value of that portion of the magnetic
moment of the magneton which is due to its rotation about its axis of
symmetry.

The appropriate expression for the energy function in the present case
may now be obtained directly from (11) through elimination of R by
means of (15), thus:

(17) U = JL(u.+v«+w^) + ^±^ + ^^'

2M^ ^ 2 ^ 2

eH
H (P sin ^ sin ^+Q cos ^ sin ^)— JmH cos $.

4mc

The modulus of transformation is easily seen to be sin 0 as before.
If now dN denote the number of magnetons per unit mass whose statis-
tical variables, 6, 0, P,Q have values which lie within the element of phase
do given by:

(18) dQ = de d0 dP dQ,

the equilibrium law of distribution for these variables will be :

(19) dN - oe""" an tf do,

42 PARA' AND DIAMAONETISM: WILLS

where

pi J. Q2 Q

(20) € = ^^ +-— H(P8in0 8m^ + QcoB^8in^) -mHcob^

2J 2mc

the teims involving u, v, w being ignored as before, and likewise the con-
stant term Cctf'os/2.

The constant a in formulas (12) and (19), if lequiied, may be deter-
mined in each case from the condition:

(21) Joe'^'sinddQ^N.

where the integration in each case is extended over all values of the
variables whose differentials appear in the corresponding expression,
for dfi.

The Langevin Distribution Formula.

In the theory of Langevin the magnitude of the magnetic moment of
a molecule (magneton) is supposed constant and directed along a polar
axis, contributions to its magnetic moment due to its rotations about
its equatorial axes being ignored. In effect, the Langevin magneton
may therefore be considered as an axial magneton whose rotation about
its axis is not subject to statistical variation, and for which the dynamical
and magnetic effects due to rotations about its equatorial axes may be
taken as negligibly small; the latter condition requires that: P » Q » 0.

The law of distribution in Langevin's theory of a paramagnetic gas
is simply obtained from (19) by placing P =" Q » 0 in the expression
(20), which involves the disappearance of the coordinate 0 and deleting
d0, dP, dQ in expression (18) for dQ.

The Langevin law of distribution is thus found to be :

(22) dN = oe '**' sin ^ dd, where a ^•

In accordance with this formula the paramagnetism of a body con-
stituted of Langevin molecules depends simply upon the distribution
of the axes of the magnetons with respect to the external field; it is
subject, of course, to the restriction of the general theory so far developed
that the effects of molecular fields are ignored. This restriction is
unimportant in the case of a paramagnetic gas.

For calculation of the magnetisation in Langevin^s theory of a para-
magnetic gas the spatial mean value of cos $, say cos 0, will be required.
From (22) it is easily found that:

(23) cos^ = coth a

a

PARA- AND DIAMAONETISM: WILLS 43

Modification of Langevin's Distribution Formula Introducing the

Magnetic Molecular Field.

A modification of Langevin's distribution formula for a paramagnetic
gas, depending upon the consideration of the molecular field due to
the magnetons of which the body is supposed constituted, will next be
considered.

In specifying the magnetic field at the centroid of a magneton in an
isotropic body we may proceed as follows.

Imagine a small sphere of radius s drawn about the centroid, s being
the shortest distance between the centroid of the magneton in question
and that of its nearest neighbor. Concentric with this sphere imagine
a second sphere drawn with radius s' large in comparison with s but small
in comparison with the bulk dimensions of the body.

The magnetic force at the centroid of the t3rpical magneton is then
the vector sum of the external force H, the force contributed by the part
of the body outside of the s' sphere, which is well known to be 4irI/3
where I is the intensity of magnetisation, and a force. A, due to the
magnetons contained in the zone between the spheres of radii s and s'
respectively.

It is with the determination of the field A, called the molecular field,
that we are now concerned.

If F denote the resultant field, then:

(24) F-H + -^I + A

K + A, (K = H + ^ I).

Now it is evident that, as we pass from magneton to magneton in the
vicinity of the t3rpical one under consideration, A will vary in direction
and magnitude.

Let N be the number of magnetons per unit mass at a point P in a
paramagnetic body supposed constituted of axial magnetons for which
it may be assumed that P and Q are negligible.

In accordance with the fundamental assumption, which closer exam-
ination shows to be justified, that all directions of the molecular field
A are equally probable, for a magneton selected at random that part,
say dN|t, of the total number N per unit mass which find themselves
in a molecular field A whose direction is delimited by a small cone with
vertex at P and of solid angle do) and for which the magnitude of A
lies between the limits A and A + dA, will be expressed by:

dN. = N ^ w(A)dA,

44

PARA' AND DIAMAGNETISM: WILLS

where w(A) is a probability function to be specified later. These
magnetons are designated as Group A.
Referring to Fig. 3 we may express dta by:

dia ^ sin y d^ d^,

where 7 and 0 are the co-latitude and the longitude of du with respect
to a polar axis in the direction of K; also from the figure:

(25) A« = P + K* - 2FK cos 5,

(26) P = A« + K* + 2AK cos 7

where 6 is the angle between K and F.

Fro. 3

By differentiation of (26) we find for all magnetons of Group A (for
which A is constant) :

FdF
sm 7 07 « —

AK

We mivy therefore write:

(27)

dNa-- -^"J^FdAdFd^.
4tK a

Let B be the angle which the axis of a typical magneton of Group A
makes with the direction of the field K; then a number, say dNb, of
the magnetons of Group A will make angles with the direction of K
which lie within the limits $ and B + 60. These magnetons are desig-
nated as of group B. The number dNb will obviously depend upon the
law of distribution of the axes of the magneton in Group A and we
may write appropriately:

dNb = dN«f(^)d^,
where f(B) is a function to be determined.

PARA' AND DIAMAONETISM: WILLS 45

It is evident that the number of magnetons of Group A which make
angles with the direction of K and which lie within the Umits $ and
$ + d$ will be equal to the number of the same group making angles
with the direction of the resultant field F which lie within the limits
p and p + dfi, is p designate the angle made with F by the axis of a
t3rpical magneton of Group A. The latter number is given by the
Langevin law of distribution. Consequently:

{{6)6$ = oe •«*'^ sin p dfl, where a = ^'
and therefore:

dN b = dN« a e • ** '^ sin /9 d/9 ;

or, upon substitution of the expression for dNa given above,

(28) dNb= - ,-^,^Fae'"''^8mj8dAdFd/Jd*.

4tK a

From the condition:

r

/

ae sm/9dj3»l,

o

the value of a is easily found:

(29) a = - a (sin h a)-^

The spatial mean value of cos P, say cos P, for the magnetons of Group
A will be required later. From (28) :

1

(30) cos /9 = coth a

a

The spatial mean value of cos B, say cos 0, for aU the N magnetons
will Ukewise be required later.

With a view to finding cos $ we first find an expression for co:3 $ in
terms of the distribution variables A, F and p.

Since $ ^ 6 + P, wo have :

cos $ = cos 8 cos /? — sin 5 sin p;

and from (25) :

P+K«-A» v/(2FK)«-(P+K«-A«)».

'"^'^ 2FK ' ''''' ^FK '

therefore:

eoB « = y=.{(P+K»-A«) co8/S-V(2FK)»-(F+K»-A*)« sin p\.

46 PARA' AND DIAMAGNBTISM: WILLS

An expression for cos 0 is obtained by multiplying the right hand mem-
ber of (28) by this expression for cos $, integrating over all values of the
distribution variables and dividing by N. It is tiius found that:

_. 1 ?w(A) * ; 5« (F + K*- A»)dF J. .CO. (I . ''r^

COB 0 = — J — T- dA J i^i Je sm/9co8/9d/3jd^

8» •J A ±(X_K)

(31)

J_ f w(A)^ r« V (2FK)« - (F« -K*~ A*)*dF
X J*e sin /S cosf ^ - pjdfifd^,

0

where the + sign in the lower limit of the integrab with respect to the
variable F is to be taken if A > K and the — sign if A < K.
Now the integral

/aeo«/l X

e sin /3 cos (- - j8) d/9

0 2

is proportional to the magnetisation of the magnetons of Group A in a
direction perpendicular to that of the resultant field F and this mag-
netisation must, on grounds of symmetry, vanish.

Consequently from (31), after integration with respect to 0 and P
and the introduction of the value of a from (29), we have finally:

(32) coe(?- J-^'dA/(cotha..-)( ) dF,

mF
where (a = -^.

No further progress toward the evaluation of cos B can be made until
the probability function w(A) has been determined.

The statistical problem here presented has been solved by Gans.^
For the argument the reader is referred to the original paper; it is some-
what lengthy and only the result will be given here.

It is found that:

4tA» --^
(33) w(A)= ==e ^^,

V^T Ao'

where Ao is a constant representing the most probable value of the
molecular field A.

^Gans: Ann, <2. Phya. 50, p. 163; 1916.

PARA' AND DIAMAGNETISM: WILLS 47

Under the assumption that there is one magneton per molecule:

where ti is the magnetic moment of a magneton, M the molecular
weight, No the Loschmidt number, p the density and s the nearest
distance of approach of two magnetons.

Inserting in (32) the expression for w(A) given by (33) we obtain for
the mean value of cos $ the following expression:

00 A+K

A«

(35)

oos0=:^Je---MAJiooth^-l) ^ + ^ ^')dF,

o MA-K)

where ^^"'kT^'

and the + sign in the lower limit of the integral involving F is to be
used if A > K and the — sign if A < K.

The Distribution Function in Quantiun Theories of Paramagnetism.

The general law of distribution for the statistical variables of a system
of similar units, which is expressed by equation (1) of the present section
is a result of classical statistical theory which presupposes that the
energy associated with any degree of freedom of a unit is capable of
continuous variation.

It will appear, however, in Sect. VII of this review that to arrive at a
satisfactory theory of paramagnetism which will account for experimen-
tal results at low temperatures it is necessary to replace the assimiption
that the energy associated with the various degrees of freedom of the
rotating magnetons is capable of continuous variation by one which
requires the energy to vary in accordance with Planck's quantiun
relation, e = hu.

It becomes necessary, therefore, to modify appropriately the law of
distribution furnished by classical statistical mechanics in order to
take account of Planck's quantum specifications relating to the energy
associated with any degree of freedom of the rotating magnetons.

The problem of quantitization here presented is quite similar to that
worked out by Planck in the derivation of his law of black body radia-
tion but is considerably more complicated, as will appear in the discus-
sion given in Sect. VII.

Further consideration of this matter is deferred until that section is
reached.

The results obtained in the present section will be of service in con-
nection with the discussion of certain theories of dia- and paramagnetism
which will be considered later.

48 PARA' AND DIAMAGNETISM: WILLS

IV
EARLY ATTEMPTS AT ELECTRON THEORIES OF MAGNETISM

At the very beginning of the epoch covered by the present survey
the foundations of the modem electron theory of matter were being
rapidly laid. Investigations during the closing years of the preceding
century furnished strong support to the view that the ultimate structure
of matter is essentially electronic in nature.

In particular the assiunption of an electronic constitution of matter
was found competent to remove many outstanding diflSculties encoun-
tered by Maxwell's electromagnetic theory in attempts to explain optical
phenomena of dispersion.

Impressed with the success of the electron theory in this direction.
Professor W. Voigt,^ in 1902, was led to an investigation having for its
object the determination of how far the electronic structure assumed for
material bodies in the optical theory of dispersion could be made to
serve in the explanation of the phenomena of magnetisation.

About the same time Sir J. J. Thomson^ engaged in an investigation
having the same object in view. His results were in accord with those
found by Voigt somewhat earlier.

On account of the importance of the results foimd by both of these

investigators it seems worth while to outline the argument of one of

them.

Voigt's Attempt at an Electron Theory of Magnetism.

In the elementary theory of dispersion it is assumed that the molecules
of a material body contain a number of electrons which, in the absence
of an external electric or magnetic field, are in stable equilibrium, or in
orbital motion about equilibrium configurations, under restoring forces
of quasi-elastic nature proportional to the displacements of the electrons
from their equilibrium positions. In order to account for absorption
the assumption is made that a dissipative force acts on each electron
proportional to its velocity of displacement. In case the body is subject
to an external electric field E and an external magnetic field H an
electron will experience two additional forces: one proportional to the
electric field intensity and one proportional to the vector product of
its velocity and the magnetic field intensity.

If (, 17, f be the rectangular coordinates of an electron with respect
to its equilibrium position as origin its equations of motion will be:

mf = -hf-k{ + eEi + -(iyH, - fH,);

c

(1) mi = - hi? - ki; + eE, + - (fHi - fH,),

c

mf hf - kf + eE, + ? «H, - i^H,),

c

1 W. Voigt: Ann, d. Phys., 9, p. 115; 1902.

• J. J. Thomson: PhU. Mag, 6, Ser. 6, p. 673; 1903.

PARA" AND DIAMAONETISM: WILLS 49^

where h and k are constants, m the mass of the electron and e its charge.
These are the f imdamental equations of the elementary electron theory
of dispersion, in which, however, the mutual effects of displacements of
the electrons are not taken into account.

Professor Voigt introduces at this point the following assumptions:
I. The external electric field shall be zero.

II. The external magnetic field shall be constant and chosen parallel
to the z-axis.
III. The dissipative constant h shall be zero, in order to correspond
to Ampere's assumption of the existence of molecular currents encoim-
tering no resistance.

With these assumptions the solutions of equations (1) are respectively :

f = ai cos (pit + ai) + a2 cos (pst + as),

(2) 1? = ai sin (pit + ai) - a2 sin (pjt + 02),

r = b sin (pt + /3),

where ai, a2, ai, as, bi, fi, pi, p2, p are constants and :

/k eH eH

(3) P=V-, Pi = P-^> P'^P + ^c'

the values for pi and p2 being approximate, in accordance with the
assmnption that the square of the natural periodicity p of the electron
is large in comparison with the quantity (eH/2mc)'.

As regards the initial conditions, the interval of time required for the
establishment of the external magnetic field is supposed to be extremely
small and its establishment is supposed to occur in such a way that the
effects of the electric field, necessarily present during the period of
establishment of the magnetic field, may be ignored.^ To the order of
approximation specified in the previous paragraph it may then be
assumed that the configuration and the velocity of an electron is un-
changed during the period of establishment of the external magnetic
field.

Accordingly, we shall have, from (2), for the initial component dis-
placements of the electron:

fo = ai cos ai -|- aa cos aj,

(4) 71^ = ai sin ai — as sin at,

fo = b sin /3;

and for the component initial velocities:

f o = ~ Pi *i sin ai — P2 as sin at,

(5) i;^ = pi ai cos ai — P2 aa cos as,

f o == pb cos p.

1 It will appear later that the effects thus ignored are of fundamental importance ii>
Langevin's theory of diamagnetism.

50 PARA' AND DIAMAGNETISM: WILLS

We now suppose the electron under consideration to be contained in a
small element of volume, dr, of a material body and that the origin O
of our system of coordinates is also contained within the element.

In order to test the magnetisation of the body we shall inquire as to
the magnetic force due to this element at a point P on the Z-axLs in
the neighborhood of the element. We first need to find an expression
for the magnetic force at P due to a single electron. It will, in fact,
suffice to confine ourselves to the consideration of the z-component of
this force.

Denoting OP by D, and supposing D large in comparison with OQ,
this component force, to second order approximation in the small quan-
tity f/D, from (11), Sect. I, may be expressed by:

<6) ^•(^f"^^^^^ + D^-

If Z denote the mean value in time of this expression, we find, with
the aid of (2), that:

Z = ^ (Pi »i* - Pt at*)-
or, after substituting the values of ai' and at' obtained from (4) and (5) :

(7) Z =^{(Pi - P2)tto* - O- PiPatto + %) + 4ppi(iiot-i)i|o}

Noting the values of pi, pt and p given by (3) this equation is seen to
reduce to:

(8) z - ^ [(.,,-,4o- ^{ f (e.«+v) -^ (e.«+o }]•

For brevity let:

e * '

(9) *.o=^(C + 0,

m *

Here, evidently, Z^ is the value of Z before the application of the mag-
netic field H; ^so is the potential energy of the electron due to its dis-
placement perpendicular to the z-axis at the instant (t = 0) the field is
applied; and "^^ is its kinetic energy at the same instant due to its
motion perpendicular to the z-axis. At any time later the corresponding

PARA- AND DIAMAGNETISM: WILLS 51

potential and kinetic energies will be denoted by ^^ and ^|. Using the
abbreviations given by (9) equation (8) may be written:

This equation expresses the difference between the z-components of
the mean value in time of the magnetic force at P due to the motion of
the electron at Q before and after the application of the magnetic field.
For present purposes what is required is the mean value of Z — Z^ due
to the spatial distribution of the electrons in an element of volume
dr at 0, of which electrons the one considered above is typical; and to
find this, the mean value in space of "9^ — ^^ for the electrons in the
element dr is required. These electrons are assumed to be originally
quite uncoordinated in configuration and motion.

Under no magnetic field the orbit of the typical electron will be
elliptical, and-the equations of the path of its projected motion on the
xy-plane will be:

(11) f = a cos pt, v = fi sin pt;

so that:

f = — pa sin pt, i; = p/3 cos pt;

and hence:

m k Ic

It follows, then, that at the instant the magnetic field is applied:

(12) ^«, - *«, = 2^ - a') cos 2pt.

Equation (11) is typical for a large number, N, of electrons in the
volume element dt. Let dt be the time of description by the typical
electron of an element of its orbit of which the projection on the xy-plane
is ds. At any instant the probability that the electron will be on this
element of its orbit will be dt/T where T is the periodic time in which
the electron describes its orbit. The "expectation," then, for the num-
ber of the N electrons which will be in the same element of phase in
their respective orbits as that defined in position and magnitude by
the element ds in the case of the typical electron will be Ndt/T; and
hence the mean value, q, in space of any conmion quantity, q, associated
with each of the N electrons will be expressed by:

1 1

1 /• Ndt If,,

52 PARA' AND DIAMAGNETISM: WILLS

Therefore, if the left hand member of (12) be taken for q:

T

^.o-^.o=^/(^.o-*Jdt.

o

The right hand member of this equation vanishes by virtue of (12),
and hence the expression on the left also vanishes. From (10) it now
follows that:

<13) Z = Zo.

Consequently, in accordance with the present theory, if the body in
question were originally unmagnetised, it would remain so upon the
application of a magnetic field.

A medium with the electron structure assumed in the elementary
theory of dispersion thus fails to account for either para- or diamagnetism
when the electrons are supposed to move in their orbits without dis-
-sipation and without collisions. If dissipation be assumed it is necessary
to the existence of a steady state that the electrons receive through
collisions accessions of energy. The question then arises as to whether
under these conditions the magnetisation due to the motion of the
electrons will be different with, and without a magnetic field.

As far as the answer to this question is concerned dissipation in a
sense may be ignored. For the effect of dissipation on the motion of
the electron will be compensated by the continually recurring collisions,
which for simplicity are supposed instantaneous. Now in the theory
of dispersion the time of description of its orbit by an electron is very
-small and it is therefore here assumed that an electron will describe
its orbit many times between successive collisions.

In the discussion which precedes it was shown that the difference,
Z — 2m, between the z-components of the mean value in time of the
magnetic force at P due to the motion of the typical electron with and
without the magnetic field H depends simply upon the difference
^M ~ ^M of its potential and kinetic energy due to its displacement and
nM)tion perpendicular to H at the instant the magnetic field is applied.
In the case now under consideration, where collisions are taken into
account, it is therefore easily seen that the effect of a collision of the
typical electron moving in the constant field of strength H is to change
the value of Z — 25o given by (10) to a new value given by:

^''^ z - Zo = - 2-^^ (*.. - *.o.

where ^n and "^zi are respectively the potential and kinetic energy at
an instant just after the collision due to the displacement and motion
-of the electron perpendicular to H.

PARA- AND DIAMAGNETISM: WILLS 53

If the collisions are quite at random, then, in the case of isotropic
bodies at any rate, the mean value of 2m due to the motions of the
electrons in an element of volume dr at O must vanish, since magnetisa-
tion would require the presence of a magnetic field. Hence to obtain
the mean value of Z we have only to ignore Zo in (14) and find the mean
value in space of the right hand member of this equation. Hence,
if n denote the number of electrons per unit volume of the type con-
sidered:

— e*Hndr -

^'^^ ^ ' - 2S^^D» ^*-> - *•»>•

Assuming completely uncoordinated configurations and motions of

the electrons ^.i is two thirds of the mean potential energy, and ^.i is
two thirds of the mean kinetic energy of the electrons reckoned for
configurations and motions just after collisions. If ^i and ^i denote
respectively the mean potential — ^and the mean kinetic energy per unit
volume for these configurations and motions, then:

,7 e«Hdr ,

But this is equal to the magnetic force which would be produced at the
point P by a small magnet at O with its axis in the direction OP and
with a moment

e'Hdr

Hence, if M be the magnetic moment per unit volume:

^''^ ^ = ^<*'-*>>'

and, if «c be the volume magnetic susceptibility:
(17) K = -^— (*i - ^0.

It appears, from the result expressed by equation (17), that with
the assumptions of the present argument it is possible to account for
both para- and diamagnetism in a medium having the same electronic
structure as that which serves so well in the optical theory of dispersion.
Moreover, the present theory does not require for the explanation of
para- and diamagnetism two essentially different fundamental assump-
tions, as is the case in the older theories of Ampere and of Weber.

The theory of Voigt leaves open the way to explanation of the ex-
perimentally well known variations of magnetic susceptibility with
changes in the physical state of the medium, through the variations in
the circumstance of collision which such changes of state entail. Our
knowledge, however, of what excites and maintains the motion of the

54 PARA- AND DIAMAGNETISM: WILLS

electroDfi is far too scant to enable the theory to predict how any par-
ticular medium will behave under the action of a magnetic field.

Sir J. J. Thompson, in his theoretical investigation of the magnetic
properties of a material witii a molecular structure in which electrons
are supposed to be grouped in rings with the electrons in any ring
spaced at equal distances around the ring and rotating with a conunon
angular velocity in a plane about an axis through its center, arrived
at the result that, unless the electrons were subject to loss of energy
through dissipation, the material would show neither dia- nor paramag-
netic quality. This is in agreement with the negative result foimd by
Voigt with the method outlined above. In the case for which dissipation
is assumed it was found that paramagnetism would result.

The difference between the magnetic properties of electrons describing
free orbits with no dissipation, in accordance with the analysis of Thom-
son, and the constant molecular currents assumed by Ampere, appears
from the analysis of Thomson to be due to the fact that in the case of
the electrons describing free orbits with no dissipation dia- and para-
magnetic effects just cancel each other.

Having been led to the negative result stated above, Thomson, in
the same paper (1903), suggested that the magnetic properties of a
substance may depend upon the properties of aggregations of largQ
numbers of molecules. In the light of the trend of ideas in the subse-
quent development of theories of magnetism a quotation is warranted:

"In the case of such aggregations, however, we may easily conceive
that the orbits of charged bodies moving within them may not be free,
but that in consequence of the forces exerted by the molecules in the
aggregate the orbit may be constrained to occupy an invariable position
with respect to the aggregate — as if, to take a rough analogy, the orbit
was a tube bored through the aggregate, so that the orbit and aggregate
move like a rigid body, and in order to deflect the orbit it is necessary
to deflect the aggregate. Under these conditions it is easy to see that
the orbits would experience forces equivalent on the average to those
on a continuous current flowing around the orbit; the aggregate and its
orbit would imder these forces act like a system of littie magnets; and
the body would exhibit magnetic properties quite analogous to those
possessed by a S3rstem of Amperean currents."

There is here a suggestion of a possible modification of the molecular
structure assumed in the optical theory of dispersion which might be
competent to account for the magnetic properties of material bodies.
The direction is indicated along which electron theories of magnetism
might naturally develop, while retaining the Amperean conception of a
magnetic molecule with currents circulating without resistance within it
in orbits which are in rigid connection with the molecule itself.

An important advance in this direction was made by Langevin in
1905.

PARA- AND DIAMAGNETISM: WILLS 55

V
THE THEORY OF LANGEVIN

The electron theory of magnetism proposed by Langevin^ in 1905 dem-
onstrated that with a suitably conceived magnetic molecule or magneton
it is possible to account satisfactorily for both dia- and paramagnetism.

The basic ideas upon which the theory of Langevin rests have been
adopted in nearly all theories of magnetism developed since 1905.
This theory is therefore reviewed below in some detail.

A magnetic molecule as conceived by Langevin contains a number of
electrons of which some are negative and some positive, the algebraic
sum of the charges on all the electrons in a molecule being ssero. Some
of the electrons are supposed to be in orbital motion within the molecule
in closed orbits and the planes of the orbits are supposed to maintain,
by virtue of internal forces, definite orientations with respect to the
molecule as a whole. The arrangement of the orbits may possess such
a degree of symmetry that the resultant magnetic moment of the mole-
cule is zero. On the other hand, if the arrangement fail of such sym-
metry, the magnetic moment of the molecule will have a finite value.

It will appear that the efifect of the application of an external mag-
netic field to a body with a structure of such magnetic molecules is to
accelerate the motions of the electrons in their orbits in a sense to produce
diamagnetism. In case the magnetio moments of the molecules are
not zero there will be superimposed upon this effect another, viz., an
orientation of the molecules tending to line up their magnetic axes in
the direction of the external field.

In the following brief review of Langevin's celebrated paper of 1905
changes in the notation have been made with the object of making it
conform more nearly with that used above and vector methods replace
cartesian.

Diamagnetism.

An examination of the properties of the molecular structure assimied
by Langevin for diamagnetic isotropic bodies will show how it is com-
petent to account for diamagnetism in such bodies.

We consider a small element of volume of such a body which for gener-
ality is supposed to be in motion. The element is supposed to contain
a large number of electrons, some of which, at any rate, are in rapid
orbital motion about the centroids of the molecules to which they belong.
Let O, Fig. 4, be the centroid of these electrons, moving with velocity v,
and let (x, y, z) be a S3rstem of rectangular axes whose directions are
fixed in space but whose origin coincides with O at the iostant under con-
sideration. Let Q(x, y, z) be the position of a typical electron, C(a, b, c)

^ Ann, de Chim, et de Phys, Ser. 8, t. V, p. 70; 1905.

56

PARA' AND DIAMAGNETISM: WILLS

the centroid of the molecule to which this electron belongs; and, with
reference to O, let r be the position vector of Q, and q that of C; while
8 is the position vector of Q with reference to C.

Assuming the element to be electrically neutral and unpolarized we
have:

(1)

Ze « 0; Zes « 0.

Since O is the centroid of the element, Zx =
Zy = Zz » 0; and, since it is isotropic:

(2)

Zxy = Zyz = Zzx * 0,
Za = Zb = Zc « 0,
Zab » Zbc " Zca » 0.

It then follows, if (, 17, f be the coordinates of
Q with reference to C, that:

(3) Zf = Ziy = Zt = Z{i, = Zi,r = Zi|f - 0,

where the summations are to be taken over
all the electrons in the element.
As far as the mean magnetic field of the electron is concerned the

electron at Q, due to its motion with velocity s about the centroid C, is,
from (12) Sect. I, equivalent to a small magnet whose moment is

Fig. 4

(4)

2c^^^'

and, if M be the magnetic moment of the element of volume due to all
the electrons of a given type within it, say classical negative electrons,
then:

(5)

M = — Zsxs,
2c •

where the summation is over all the electrons of the type considered.

By differentiation with respect to the time we find for the time rate of
change of this quantity:

(6)

e

M =* —Zsxs.
2c

If F be the force on the typical electron due to the action upon it
of the rest of the molecule in which it is situated, B and H the electric
and magnetic force, respectively, of external origin, we shall have from
the equation of motion of the typical electron:

(7)

ms = F+eE + - (v + s) xH - mq - mv,

where e is the charge of the electron and m its mass.

PARA' AND DIAMAONETISM: WILLS 57

Here the quantities F, E, and H all refer to the point at which the
typical electron is situated, but, since the element of volume is small,
they may be expressed as follows:

F =Fo + (s.VF)o + ....,

(8) E =E, + (s.VE), + ....,

H = H„+(s.VH)^ +

where the zero subscript indicates that the corresponding quantity
is to be evaluated at O, the centroid of the element.

From (6), (8), (2), and (3) we obtain, upon neglecting terms of higher
order than the first in the small quantity s, writing

(9) Z? = Zi,« = 2f« = -'
and dropping the zero subscripts:

(10) M = ;^2)sxF+;^y(curlE + - vdivH- -vVH)-- — H>»
^ ^ 2mc 4mc I c c c dt j

in which the vectors and their space derivatives refer to the point O.
Now, from Maxwell's field equations:

J. rx .X i« l^H 1/ „„ dH\

divH = 0; curlE=-^-«.(vVH--j;

and consequently the preceding equation reduces to:
<11) M = 4ssxF-£^|(IH).

The first term on the right represents a time variation in M due to
the action of the internal forces of the molecules; this vanishes if , as in
Langevin's theory of diamagnetism, each molecule has no initial mag-
netic moment.

If AM denote the change in the magnetic moment of the element due
to the establishment of the external field within it, then, by integration
of the last equation:

(12) AM=--^,IH.

4mc*

Owing to the creation of the external field within it the element thus
acquires a diamagnetic moment.

It may be noticed that in the expression for AM the charge of an elec-
tron appears as a square. Consequently, if positive as well as negative
electrons are in orbital motion within the molecules, they, too, will
give rise to diamagnetism in accordance with (12). On accoimt of the
large mass of the positive electron, however, the contribution of the

68 PARA' AND DIAMAGNETISM: WILLS

positive electrons to diamagnetism would probably be very small in
comparison with that of the negative.

For the quantity I we may write nk^, n being the number of electrons
in the element of volume and k' the square of the radius of gyration of
the mean configuration of the electrons in a molecide with respect to
an axis through their centroid. From (12) we then have:

(13) AM = --^k«H.

4mc*

From (14) Sect. I, the mean absolute value of the components of the
magnetic moments of the n orbits in the direction of H, say Mh» ^^ ^
given by:

eS

(14) Mh = ~ cos ^,

CT

where cos B denotes the mean value of cos 9, 9 being the angle between
n and H.

From (13) the change in m^i ^ay Am^, due to the creation of the mag-
netic field H will be given by:

e*

It follows from (14) and (15) that the ratio of A^h to Mh niust be
very smaU for aU attainable field strengths; in fact less than 10^^ if
T be assumed of the order of the period of light vibrations, say 10""
seconds.

If «c be the magnetic susceptibility per unit volume and N the number
of electrons per unit volume, then, from (13):

where p is the mass density of electrons per imit volume.

In accordance with the argument advanced here all substances will
possess the diamagnetic property. If the magnetic molecules of any
substance possess initial magnetic moment of their own, they will
possess paramagnetic as well as diamagnetic quality. If the initial
moment be zero, no external action upon the molecule will produce one.

The argument has left out of account any explicit reference to the
effect of collisions among the molecules upon the diamagnetic state of
the substance. It will be recalled, however, that there has been nothing
assumed in the argument to prevent motion of the most general kind of
the element of volume containing the ensemble of electrons; and hence,
whatever be the motion of the ensemble, its diamagnetic state is at

PARA- AND DIAMAGNETISM: WILLS 59

each instant determined simply by its actual configuration with reference
to the external magnetic field, and therefore is independent of collisions
among the molecules.

Again the argument does not take account of the interior forces of a
molecule which may result from the diamagnetic action itself. But
it will be seen presently that, in the mean, a diamagnetic modification
implies only a change of velocity of an electron in its orbit without
deformation of the orbit, and the absence of a mean deformation of
the molecule due to a diamagnetic modification implies that the cones-
ponding interior reactions due to it must be negligible.

The fijdty of spectral lines lends important support to the view that
the intramolecular motions of a substance depend but slightly upon
the temperature; the comparatively slow thermal motions can therefore
modify but very little the intramolecular motions giving rise to diamag-
netism on the present theory. The diamagnetic property is thus
practically independent of temperature, according to the experimental
law of Curie. There are, however, many exceptions to this law.

An important question is that relating to a possible change in the
area of the orbit of an electron due to the action of an external magnetic
field. Let f (r) be the central force which holds an electron in its orbit,
supposed circular.

In the absence of an external field :

(17) mcA = f(r),

where w is the angular velocity of an electron in its orbit.

If H^ denote the component of the external field perpendicular to
the plane of the orbit, then, after the field is applied:

m (o) + Aw)* (r + Ar) +-H„ea)r =f(r + Ar),

c

where Aoy and Ar are the variations in o) and r respectively due to the
action of the field. Retaining terms of the first order only in the small
quantities, Aciy and Ar, we therefore have:

f' (r)Ar = 2mra)Aa) + ma9'Ar+ -HnCwr;

c

and hence :

(18) (f' (r) - m«*} Ar » 2ma)rAa> + -H„e«r.

c

Now, if r be the orbital period and S the orbital area, then, using (14)
and (17), Sect. I:

Awr* c eS c e er*

=^«-A- «-Am= — A(HScos^) = - — H„:

2 e cr e ** 4irtnc ^ ^ 4mo^'

60 PARA' AND DIAMAONETISM: WILLS

and, therefore:

H.e
2mc '

(19) — 4ina)*Ar = 2ma)rAa) + -H^etfr,

c

From (18) and (19):

(f'(r) + 3m«»}Ar-0.
Thus, either:

H.e

(a) Ar - 0; Aw - "^

2mc

or:

(b) f'(r) - -3m«»- --*

r

If condition (b) ie satisfied,

f _3

r " r '
andhenoe:

(20) f » ^

where K is a constant.

Except, then, in the very special case that the central force holding
the electron in its orbit varies inversely as the cube of the radius of the
orbit, condition (a) will be satisfied, the effect of the magnetic field being
simply to cause a variation of the angular velocity of the electron by
the amoimt — Ho e/2mc.

It is evident that the component of the magnetic force in the plane
of the orbit will not operate to change the area of the orbit, since the
displacements to which it gives rise are perpendicular to the plane of
the orbit.

The change of period, giving rise to diamagnetism, in the orbital
motions of electrons within the atoms corresponds to the simple Zeeman
effect in magneto-optics.

It is of some little interest to compare the formulas (16) found for
diamagnetism by Langevin with that which holds for a substance which
is constituted of the spherical magnetons discussed in Sect. II.

It was there shown that the effect of establishing an external magnetic
field H within such a magneton was to change its magnetic moment by
an amount:

where Q is the moment of inertia of charge of the magneton and J its

PARA' AND DIAMAGNETISM: WILLS 61

ordinary moment of inertia with respect to an axis through its centroid»
If the magneton be constituted of electrons of a single type, of mass m
and charge e, symmetrically spaced about a positive nucleus then Q»
ek* and q = mk^ where k is the radius of gyration of the electrons in the
magneton.

If, then, K be the volume magnetic susceptibiUty of the body con-
stituted of such magnetons it follows from (21) that:

Ne«
jc = — -— - k^
4mc^

where N denotes the number of electrons per imit volume and p the
mass density^ of the electrons. This result agrees with that expressed
by (16).

Paramagnetism.

A body will exhibit paramagnetic quality in the presence of a mag-
netic field in addition to the diamagnetism considered above when its
magnetic molecules have individually other than zero magnetic moment.
The theory appropriate to a paramagnetic gas was developed by Lange-
vin in his paper of 1905. Later this theory was made the basis of a
theory of ferromagnetism by Weiss.

In Langevin's theory of a paramagnetic gas the magnitude of the
magnetic moment of a molecule is supposed to be invariable under all
conditions, the slight diamagnetic changes in its moment being ignored.

It is of interest to examine first, as regards its general nature, the
process whereby the paramagnetic state is set up when a gas is sub-
jected to an external magnetic field. At the instant the field is appUed
the diamagnetic state discussed above will be established immediately.
The paramagnetic state, on the other hand, will require an appreciable
time for its establishment.

At the instant the magnetic field H is appUed a magnetic molecule
acquires potential energy with respect to the field of amount

-H dv

where if is its magnetic moment. This increase in the potential energy
of a molecule is derived initially from its kinetic energy of rotation,
just as the potential energy of a molecule of a gas subjected to a gravita-
tional field is acquired from its kinetic energy while it is rising in the
field. Now the result of this partition of kinetic energy among the
various degrees of freedom (translation and rotation) of the molecules ia
incompatible with thermal equilibrium. It is in the process of the
establishment of thermal equilibrium through collisions that para^
magnetism makes its appearance. In this process magnetic energy is
derived from the energy of thermal agitation of amount

- Hdv.

62 PARA- AND DIAMAQNETI8M: WILLS

If the molecules have no other potential energy relative to their
orientation, as in the case of a gas and probably a liquid, in order to
maintain the medium at a constant temperature it would be necessary
at each instant to furnish to it an amount of heat energy per unit volume
equal to — H.dl, if I denote intensity of magnetisation. In the case
of a solid where the molecules have a potential energy of orientation
it is only for the case of a closed cycle that a similar conclusion may be
drawn.

\^th the aid of the laws of thermod3mamics it is easQy shown that
the magnetic moment M of a given mass of a paramagnetic substance
in an external field of strength H must, in the case of a gas or a liquid,
be a function of H/T:

M

-<f)

where T is the absolute temperature.

For a small reversible modification in which H changes by dH, and
T by dT, the heat evolved, say dQ, which depends upon H, will be given
by:

dQ.H(-dH + -dT).

Since the modification is reversible dQ/T must be a perfect
and hence:

A/1 ^^ . ^/l ^\
dT\T dH/ dH\T dT / '

from which it follows directly that:

dT dH

The integral of this equation is given by :
(22) M = f (ly

which is the result it was desired to prove. The argument is readily
extended to show that this result will also hold for a solid body, pro-
vided its internal energy does not depend appreciably upon H.

Thermodynamics alone will not permit of the determination of the
function f . For many substances experiment shows M to be directly
proportional to H and this, with the result expressed by (22), if the
conditions stated are satisfied, leads to the result:

(23) M = -H

where A is a constant independent of T.

PARA- AND DIAMAGNETISM: WILLff 63

In the particular case of a paramagnetic gas such as oxygen the form
of the function f may easily be determined.

Theory of a Paramagnetic Gas.

In his theory of a paramagnetic gas Langevin assumes each of the
magnetic molecules to have a magnetic moment, Mj the magnitude of
which is the same for all molecules. The direction of the magnetic
axis of the molecule is then that of the vector m- Elffects due to the
rotations of a molecule about axes perpendicular to its magnetic axis
are ignored. The molecular magnetic field is also ignored, since it will
certainly be very small for gases under ordinary conditions.

The appropriate distribution function for this case, as Langevin
showed, is given by (22) Sect. Ill; and from (23) of the same section
the mean vdue, cos 6, of the cosines of the angles made by the magnetic
axes of the molecules with the external magnetic field H is expressed by:

1 ^H

(24) cos ^ = coth a ; (a = — ),

a ic X

where k is Boltzmann's constant and T the absolute temperature.

On grounds of Efymmetry I, the intensity of magnetisation, must be
in the direction of the external field H; and it must be equal in magnitude
to the sum of the projections of the moments of the individual molecules
in imit volume in this direction. Accordingly I will be given by:

I « MU cos ^
(^^^ = Mn (coth a - i)

where n denotes the number of magnetic molecules per unit volume.

From this result it appears that I is a function of H/T as re-
quired by thermodynamics and, moreover, owing to the factor n, that
it is directly proportional to the pressure of the gas.

When the bracketed expression on the right of the expression for I
takes on the value unity I will assume its maximum vsdue, ^n, De*
noting the maximum value of I by lo, from (25) :

(26) I = lo (cosh a - -).

a

The curve in Fig. 5 shows the manner in which I/Io varies with a»
It will be seen presently that imder ordinary conditions of experiment
a will be quite small for oxygen; in fact of the order 10~*.

When a is small I/Io will vary directly with H. At low temperatures,
however, and for powerful fields a may become so large that the relation
between I/Io and a becomes non-linear.

64

PARA" AND DIAMAONETISM: WILLS

From (26), by development of coth a in ascending powers of a, neg-
lecting powers of higher order than the first:

(27)

I - I.|.

with sufficient approximation under ordinary conditions of experiment;
and, if K be the coefficient of volume magnetic susceptibility:

(28)

3kT'

•• «.

Fio. 5

showing that k varies inversely with the temperature in accordance
with what is known as Curie's law of paramagnetism.

The preceding theory may
Also be valid for a medium other \r
than a polyatomic gas, such as
oxygen, when the energy of ro-
tation of the molecules is known
to be a function of the tempera-
ture, in accordance with thermo-
d3mamic theory. In all such
cases it is only necessary that
the energy of rotation shall be
proportional to the absolute
temperature in order that the
theory may be applicable; the
•quantity k only will have to be modified.

All magnetic substances for which the mutual actions among the
molecules are negligible, such as solutions of paramagnetic salts, should
have magnetisation curves exactly similar.

The expression for k given by (28) may be written:

since lo = n/i, and p == nkT, p being the pressure of the gas.

At normal pressure, and at the temperature O^C, Curie found for
oxygen:

K = 1.43 X 10-^

It follows that the maximimi intensity of magnetisation for oxygen
will be given by:

lo « (3 X 10» X 1.43 X 10-^* = 0.65.
For liquid oxygen, therefore, with a density 500 times greater, a value
of I > 325 might be expected.

PARA- AND DIAMAQNETISM: WILLS d6

The order of a for oxygen under ordinary conditions of experiment
may now be found. We have:

a ^ _ ___ ™^ •
nkT p

The value of I found above for oxygen under nonnal conditions was 0.65.
Hence:

a « 0.65 X 10-*H;

for a fairly powerful field, H » 10,000 say, and then:

a = 0.66 X 10-«.

If it be admitted that the magnetic moment m for a molecule of oxygen
is due to a single electron with a charge equal to that of an atom of
hydrogen in electrolysis rotating in a circular orbit of radius r equal to
1.5 X 10~* cm., the velocity of the electron may be calculated as follows.
Since, from (14), Sect. I,

_ eS _ evr
'*■" ci^ " 2c'

where S is the area of the orbit, r the periodic time and v the velocity
of the election, we shall have:

1^ = 11/4= ^°®^-

Now e, being expressed in electrostatic units,

ne

- « 0.40;

c

and since under normal conditions, as found above, lo » 0.65, it follows
that:

^ «. ^ .^ 1-5 X 10^
0.65 - 0.40 X — X V,

from which:

V = 2 X 10* cm/sec.

This velocity is of the same order as that which an electron would
have in stable circular orbital motion about a positive charge of equal
magnitude placed at the center of the orbit. For in this case:

mv* e* . e*

=-- , V* = — -.

r r* mr

66 PARA' AND DIAMAGNETISM: WILLS

from which:

V = 10" cms/sec.

It is worthy of note that the resultant magnetic moment of a molecule
of oxygen may be accounted for by the orbital motion of a single electron;
this would also be true for a molecule of iron, for which the maximum
magnetisation per molecule is of the same order as that for oxygen.

In the case of the magnetisation of a paramagnetic gas such as oxygen,
we have seen that the kinetic energy of the molecules furnishes per
unit volume during the period of rearrangement (which results in the
appearance of paramagnetism) an amoimt of energy

- /H.dl;

so that the energy per unit volume of the medium is augmented by an
amount

2

The gas must therefore be heated by an amount which may be cal-
culated as follows.

Suppose the volume to remain constant and let AT be the rise in
temperature due to magnetisation, then:

CAT ^Kw.
2

C being the specific heat at constant volume. Now, approximately,
in C. G. S. units:

and, therefore,

C = 10^ K « 1.43 X 10-^

AT = 0.8 X 10-"H«.

From this result, for H = 10,000, AT « 10r*C; while for H =
40,000, AT - 10"* C**. This elevation of temperature would vary
directly with the susceptibility «, and therefore inversely with the abso-
lute temperature.

In concluding this somewhat brief review of Langevin's theory the
following remarks may prove to be of interest later.

His theory of paramagnetism is what may be termed an equipartition
theory ; for it is based on classical statistical theory that leads to equipar-
tition of energy among the statistical coordinates of a system which
appear only as the sum of squares in the energy function of a statistical
unit.

PARA' AND DIAMAGNETISM: WILLS 67

The property of pennanancy is given to the magnetic moments of
the molecules; for example, these moments are not subject to variation
with temperature.

The effects of intra-molecular forces have been ignored, thus restricting
the range of application of the theory to paramagnetic gases.

By ignoring the effects due to rotations of the molecules about axes
perpendicular to their magnetic axes they are deprived of gyroscopic
properties which, as we shall see, may play an important role in magneti-
sation.

68 PARA" AND DIAMAGNETISM: WILLS

VI

MODIFICATIONS OF THE THEORY OF LANGEVIN INDEPENDENT

OF QUANTA HYPOTHESES

The theory of Langevin, as we have seen, leads in the case of diamag-
netism to the result that the diamagnetic susceptibility of all bodies
should be independent of the temperature and the field strength; and in
the case of paramagnetism to Curie's law, which requires the suscepti-
bility to vary inversely with the absolute temperature.

Now many of the experimental facts found since the time (1905) of
publication of Langevin's theory are not in accord with these results.
Consequently various attempts at modification of the theory have been
made. In the present section we shall consider modifications of the
Lang^vin theory which do not invoke the aid of quantum h3rpothe8e8.

Theory of Honda.

Eotaro Honda^ in 1914 proposed a modification based upon the follow-
ing two assumptions:

(a) — The magnetic moments of molecules are not constant but depend
upon the temperature.

(b) — The molecules exert mutual forces upon one another, the ten-
dency of which is to prevent their lining up in the direction of the
external field.

A magnetic molecule in the case of a soUd is supposed by Honda to
consist in general of an aggregate of a number of actual molecules,
such aggregates being subject, however, to the usual laws of thermal
molecular motion. In accordance with assumption (a) the form of a
molecule is supposed to depend upon the temperature; a change in form
involving at the same time a change in the value of the magnetic moment
of the molecule. Thus the form of the molecule of a body in the ferro-
magnetic state is assumed to be spherical, so that it shall not be subject
to orientation through thermal impacts. In the ferromagnetic range
of temperatures the small mutual forces only will be operative in opposing
the tendency of the magnetic molecules to hne up with their axes along
the direction of the external magnetic field, and in consequence a large
magnetisation results in this case. In the passage from the ferromagnetic
state to the paramagnetic the magnetic molecule is supposed to pass
from the spherical to an elongated form, with the result that a large
thermal action opposing the lining up of the molecules becomes operative,
and consequently the body passes from the ferro- to the paramagentic
state. The energy of deformation of the molecules in this process
together with that required by the new degrees of freedom is supposed
to account for the heat absorbed in the process of transition.

1 K. Honda, Tokio, Sci. Rep. 3. p. 171; 1914.

PARA' AND DIAMAGNETISM: WILLS 69

The distribution function proposed by Honda, incorporating the
assumptions (a) and (b), is:

Mof(T)H ^
a e r-zi— — cos 0,
kT + «

where /iof(T) represents the magnetic moment of a molecule, /io being
the value of this quantity at absolute zero, and 0 is a constant or a
function of the temperature expressing the mutual action of the molecules
upon one another. This distribution function of Honda reduces to
that of Langevin if f (T) = const., and 0 = 0.

The symbols other than fiJl{T)j and 4> have the same significance as
in Langevin's theory.

The function 4> which expresses the mutual action of the molecules
represents an effect which, like thermal action, tends to hinder the
lining up of the molecules with their axes in the direction of the external
magnetic force, and hence is added to ihe temperature factor kT.
In paramagnetic bodies 0 is in general small in comparison with kT
and only becomes of importance at low temperatures.

The modified distribution function leads to an expression for the
magnetic susceptibility which is in good agreement in many cases with
the experimental results of K. Onnes and A. Perrier at low temperatures
and also with other experimental results at higher temperatures, when
appropriate choice of the temperature functions f and 0 are made.

The theory is also applied with some success to the explanation of the
paramagnetic behaviour of ferromagnetic substances at temperatures
above the critical temperature.

The functions f and 0 are not capable of determination from theoretical
considerations, and the theory suffers chiefly from this deficiency.

Theories of R. Gans.

In a series of papers beginning in 1910 R. Gans^ * ' ^ has made suc-
cessive attempts toward the improvement of theories of dia- and para-
magnetism beyond the point reached by Langevin in his paper of 1905.

The progress made by Gans in this connection may perhaps be satis-
factorily estimated from a brief review of two of his papers which ap-
peared in 1916, entitled respectively "Theorie des Dia-, Para-, und
Metamagnetismus,"* and "Uber Paramagnetismus."*

In the former of these two papers he considers a material body
supposed constituted of axial magnetons. The magneton itself is
supposed to consist of a rigid system of classical negative electrons

1 R. Gans: Oott, Naehr,, p. 197; 1910.
tR. Gans: OoU, Nadir., p. 118; 1911.
• R. Gana: Ann, d, Phya. 49, p. 149; 1916.
« R. Gans: Ann. d. Phya. 50. p. 103; 1916.

70 PARA' AND DIAMAQNETI8M: WILLS

within a unifonnly charged positive sphere, the center of which
coincides with the centroid of the system of negative electrons. The
equatorial moments of inertia of the magneton, A and B, are supposed
equal and the polar axis for which C is the moment of inertia is called
simply the axis of the magneton.

The angular velocities of rotation of the magneton are supposed so
small that the resultant magnetic fields due to these rotations nuty be
considered as linear functions of the corresponding angular velocities;
and the accelerations giving rise to radiation to be so small that the
energy radiated may be neglected. The magneton system may then be
considered as quasiHstationary. Furthermore the inertia mass of a
magneton is supposed to be entirely of electromagnetic origin. Finally,
the molecular magnetic field is ignored on the present theory.

For a body constituted of magnetons of the type here contemplated,
either one or the other of the laws of distribution given respectively by
(12) or (19) of Sect. Ill is applicable, depending upon whether or not
the rotations of the magnetons about their individual axes of symmetiy
are dependent upon or independent of thermal agitation. In the
former case the law of distribution leads to a theory of diamagnetism
and in the latter to a theory of paramagnetism. We consider the
former case first.

From (12) Sect. Ill, the appropriate law of distribution for this case is:

__ «

(1) dN »ae ^''^sintfdQ,

where:

(2) €«-^+~+r— (Psin*sin^ + Qcos«sin^ + Rco8^),

2J ZL* 2mc

and

(3) dQ « dtfd^d^dPdQdR;

the significance of all the quantities here involved is given in Sect. IIL
The expression (1) gives the number of magnetons per unit mass
whose statistical variables are delimited by the elementary phase domain
do. Each of these will contribute to the magnetisation per unit mass
an amoimt |^.H/H, |^ being the magnetic moment of any one of the
magnetons of this group.

If M denote the scalar value of the magnetisation per unit mass,
then:

(4) U'^nf^e'^BinedQ^fe^^mnedQ,

PARA' AND DIAMAGNETISM: WILLS 71

where the integratioD is to be taken over all values of the variables
whose differentials occur in dl2 and is supposed performed after t^ .H/H
is expressed in terms of these variables.
From (60), Sect. II, with the aid of (7) and (8), Sect. Ill:

~=- = b (P sin 0 sin ^ + Q cos 0 sin ^ + R cos ^), (b = r-—).
H 2mc

If we now let:

/e-«*

(6) Z = /e "^sin^dft,

the expression for M may be put in the simple form:

NkT d log Z.

(6) M = -

H dlogb

By division of this expression by H we obtain for the susceptibility
per unit mass, Xt the following expression:

^^^ ^ H* dlogb

For the case in which the rotations of the magnetons about their
axes of symmetry are supposed independent of thermal agitation the
appropriate law of distribution is given by (19), Sect. Ill:

(8) dN -ae^sin^dO,

where

ps J. ^ e

(9) e r=. /^ + -— H (P sin 0 sin ^ + Q cos 0 sin ^) — m H cos 9,

2 J 2mc

(10) do = d^ d^ d« dP dQ.

Proceeding in a similar way to that foUowed in the case just considered
the following expression is found for the susceptibility per unit mass:

^ NkT/d log Z' d log Z'\
^^^^ ^ ^ tf V(d log M d log b /'

where

(12) Z' = fe" " sin ^ dQ,

€ and do being given by (9) and (10) respectively.

n tAMU- JL%:^ LiAM^^a^m^M WILL^

IZ) Z»4i*v"2ykT//=c/e *** ■»#<».

TUsnMnwmxknmammUjwomierzJ >C;J «C;a^J<C

C«e 7-4 > C.

U for btetitj wt

spot:

04;

1

y-

tPV(J-C),
2kT

♦ (7)

then from (13):

4.»>/«

'•i'c

(15) Z -

r V(2»kD»PC-

# — C

T

Unog thk value for Z fonnulft (7) fmniriies for the magnetic auaoep-
per unit

X- -m/m +

\ ^ 2J VV'-r^T) y/j

This expreation aawimen a aunider fcMin if we let:

(16)

2 12

°^^^"Vi7e^»(T)"7i'*'3'

2 2 + C/J'
whereupon we obtam:

(17) X - - NW?^?^ {1 + h0(7)|.

The quantity y defined above is at constant temperature proportional
to the field strength, H, while I/t* at constant field strength is propor-
tional to the absolute temperature, T.

The susooptibility x> lus shown by the expression just found, depends
upon the function tl(y) which may be calcidated with the aid of a table

. i

PARA- AND DIAMAGNETISM: WILLS

73

for the probability integral for assumed values of y and l/y. The
variation of 0(7) with 7 (proportional to H) and of 0(7) with I/t*
(proportional to the absolute temperature) are shown in Fig. 6 and Fig. 7
respectively.

From formula (17) it is seen that the susceptibility, Xoi ^i" very weak
fields is given by:

xrw2J+C

and hence:

X — Xo

Xoh

- «(7).

In general, then, x depends upon the field strength and investigation
brings out the fact that the curve showing the relation between the
susceptibility and field strength is of the t3rpe shown in Fig, 8.

In Fig. 9 is shown the type of curve obtained experimentally by

Fig, 8

H

V.

Fig. 9

Honda for many diamagnetic substances. The experiments of Honda
were not sufficiently extended in the direction of small field strengths
to show whether or not his curves, if continued, would be of the type
called for by the present theory.

If the present theory in its main features is correct suitable quanti-
tative measurements of the susceptibility would make possible the

74 PARA' AND DIAMAGNETISM: WILLS

derivation of valuable information as to the constitution of the magneton,
as regards its size, shape, and moments of inertia.

Case II— J - C.

In this case all the principal moments of inertia of the magneton
are equal, and hence h " 0, and 7 "» 0, so that:

X - Xo - - Nb»J;

the susceptibility is thus independent of both field strength and
temperature. This is found experimentally to be the case with many
substances.

It is important to remember in connection with this case that although
the principal moments of inertia of the magneton are assumed equal,
this does not imply that the magneton is to be considered as a geometrical
sphere. If this were the case the statistical method would be no longer
applicable and the problem would become one of electromagnetism
simply.

Case III— J < C.

The results found for this case are quite similar to those found for
Case I and it is therefore not worth while to consider it in detail.

Moments of Intertia of Diamagnetic Magnetons.

As a result of an extensive series of experiments, H. Isnardi* reached
the conclusion that diamagnetic susceptibility in general is quite inde-
pendent of the field strength. If this be so the assumptions of Case II
are warranted. The principal moments of inertia of the magneton
may then be considered equal and the formula found for the suscepti-
bility for this case may be used.

Upon substituting for b its value e/2mc in this formula we obtain:

where N is the number of magnetons per gram, e/mc « 1.77 x 10*
electromagnetic units and J the moment of inertia of the magneton.

Assuming one magneton to the atom, if No be Avogadro's number,
and A the atomic weight.

No = NA « 6.176 X 10";

and we obtain from the formula for x:

J « - 2.067 X 10-«AX.

> H. lanardi, Contribuci6n ai estudio de las eieneiaav Uniy. Naol. de La Plata.-^Aiifi. d.
Phu9, 61, p. 685; 1920.

PARA' AND DIAMAGNETISM: WILLS

76

From the experimental results foimd by Owen^ the values of J have
been calculated by Cans' for various elements with the aid of this
formula. These values together with the corresponding values for
A and x are given in:

Table I

£3emeot

A

-xXlO»

JX10«
in g. ems'.

Element

A

-xX10»

JX10-
in g. cms*.

Be

B

C(Dia) . .

8

P

Sa

Cw

Zn

Ga

Ge

As

Se

Br

Sr

9.1
11.0
12.0
28.3
31.0
32.07
63.57
66.37
69.9
72.6
75.0
79.2
79.92
87.62

1.00

0.7

0.49

0.13

0.90

0.49

0.086

0.166

0.24

0.12

0.31

0.32

0.40

0.2?

1.88

1.69

1.22

0.761

6.77

3.26

1.12

2.09

3.46

1.80

4.81

6.24

6.61

3.62?

Zr

^:::::

In

Sn(gray)

Sb

Te

I

Cs

Pb

Bi

90.6
107.9
112.4
114.8
119.0
120.2
127 6
126.9
132.8
197.2
200.0
204.0
207.1
208.0

0.46

0.20

0.18

0.11

0.35ap.

0.82

0.32

0.36

0.10

0.16

0.19

0.24

0.12

1.40

8^43

4.46

4.18

2.61

8.61ap.
20.4

8.43

9.46

2.76

6.12

7.86
10.1

6.14
60.2

It appears that the values for the moments of inertia for the various
substances are all of the same order of magnetude. These values are
considerably less than those found for paramagnetic substances, as
will appear later.

In this connection it should be remarked that Isnardi's conclusion
that diamagnetic susceptibilities are in general independent of field
strength, is not fully supported by the experiments of Frivold.'

Paramagnetism and Metamagnetism.

Formulas (11) and (12) are those required for the explanation^of
para- and metamagnetism.

From (12), after integration with respect to P, Q, ^, 4>, and the
substitution of x for cos 6, we obtain:

(19)
where

(20)

■*"i tt«(l-x«) ax.

Z' = 8ir»kTj/e e dx,

-1

2i_2

2 kT

a =

kT

* M. Owen, Ann, d. Phy». 37, p. 664; 1912.

> R. GanB, Ann. d. Phys. 61, p. 163; 1920.

« O. E. Frivold, Ann. d. Phy$. 57, p. 471; 1918.

76

PARA' AND DIAMAGNETISM: WILUS

The expression (19) after integration with respect to x may be
written in the form :

(21)

Z' » 4 ir»V5i- kTJ

e

a« +T«

{*(r + a) -*(r-.a)),

where.

T> =

a'

4a« 2b»kT'

(22)

T±€t

* (t ± a) « "7= Je-^"dX.

Observing that d logr — d log m — d log b and that d log a « d log b,
we obtain from (11) and (20) the following expression for the suscepti-
bility per unit mass:

Nb*J , 4 e"'^'* + "'>

(33) x-x-7 {4t>- 2rf + l -

2of

y/r * (r + a) ^ ^ {t -^ a)
(2 r sinh 2aT + a cosh 2 a r). {

The values of x divided by the constant N b* A f or various values
of a and T, given in Table II, were calculated by Gans from formula
(23). For a given value 6f the temperature,T, the quantity r is constant,
from (22) ; and a is directly proportional to the field strength, from (20).

Table II
x+Nb«A

a

r-0

r-H

r-1

r-2

0.0

-0.667

-0.600

0.000

+2.00

0.2

-0.670

-0.606

-0.010

+1.92

0.4

-0.681

-0.614

-0.040

+1.69

0.6

-0.698

-0.636

-0.082

+1.348

0.8

-0.719

-0.660

-0.136

+1.093

1.0

-0.746

-0.691

-0.196

+1.900

1.6

-0.8189

-0.682

-0.339

+1.636

2.0

-0.8802

-0.773

-0.488

+1.278

3.0

-0.9444

-0.889

-0.723

-0.143

4.0

-0.9687

-0.937

-0.844

-0.471

6.0

-0.9800

-0.960

-0.900

-0.660

10.0

-0.9960

-0.990

-0.976

-0.916

00

1

-1.0000

1.000

-1.000

-1.000

The table shows that, for all values of r equal to unity or less, x is
negative for all values of a, except that when r == 1 and a = 0, it
vanishes; and that for r » 2, x niay be positive for values of a suf-

PARA' AND DIAMAGNETISM: WILLS 77

ficiently low, and negative for higher values of a; the susceptibility
thus depending upon the field strength.

A substance whose susceptibility, as regards sign, depends upon the
field strength is called metamagnetic. Weber and Overbeck^ have
observed the phenomenon of metamagnetism in copper-zinc aUo3rs;
and Honda has observed it in the element Indium.* It is possible,
however, that the observed phenomena might have been due to the
presence of traces of iron in the specimens.

Another interesting conclusion which may be drawn from the present
theory is that, by suitable increase of temperature and field strength,
all so-called paramagnetic bodies would become diamagnetic.

The explanation of the curious results called for by the present
theory of paramagnetism is to be found in the fact that the theory
itself tacitly hypothecates two separable causes operative to produce
magnetisation; one tending to produce diamagnetism, and the other
paramagnetism.

It is not difficult to see that the cause operating to produce dia-
magnetism is the rotations, subject to thermal variation, of the
magnetons about their equatorial axes of inertia; and that the cause
tending to produce paramagnetism is the rotations, not subject to
thermal variation, of the magnetons about their axes of symmetry.

The relative strengths of these two operating causes depend, in
accordance with the theory, upon the temperature and field strength;
and, therefore, according to the values of these two quantities, the one
cause or the other may predominate.

In the second' of his papers published in 1916, entitled "Uber Para-
magnetismus," Gans developed a theory of paramagnetism in which
the molecular magnetic field is taken into account, this field having
been ignored in his paper on dia-, para-, and metamagnetism just
reviewed.

It will be recalled that on the latter theory paramagnetism cannot
exist by itself, but always occurs accompanied by diamagnetism, caused
by the effects of thermal variations in the rotations of the magnetons
about their equatorial axes of inertia; and that, with sufficiently high
temperatures and external fields, the diamagnetism due to this cause
may predominate over the paramagnetism due to the rotations with
constant angular velocity of the magnetons about their axes of sym-
metry.

For temperatures which are attainable, however, in the case of almost
all paramagnetic substances, the paramagnetic effect predominates

1 K. Overbeck: Ann. d. Phys. 46. p. 677; 1915.
s K. Honda: Ann. d. Phy». 32. p. 1043; 1910.
» I.e. — p. 69, note 4.

78 PARA'^ AND DIAMAGNETISM: WILLS

strongly over the diamagnetic effect, which may consequently be ig-
nored and each magneton considered to have a constant magnetic
moment m due to its rotation with constant angular velocity about its
axis of symmetry; it is assumed that this moment is the same for all
magnetons. The magneton thus considered is the equivalent of the
magnetic molecule of Lang^vin.

With the assumptions relating to the magneton here made formula
(85) y Sect. Ill, is applicable for the calculation of the magnetic moment

per unit mass. This formula gives the mean value, cos ^, of cos 9, 9
being the angle between the direction of the axis of a magneton and the
field K, whose relation to the external field H and the intensity of
magnetisation, I, \b expressed by the equation:

(24) ^"^+f''

The magnetic moment per unit mass is obtained by multplying

cos ^ by the product of the number of magnetons per unit mass, N,
and the constant magnetic moment, Mi of a magneton. We thus ob-
tain, from the formula for cos $ in question, the following expression
for the magnetic moment per unit mass:

(25) M«Nmcos9

' "^AdA I (cotha — ) ( — ^— ) dF,

> O :I:(A-K)

V^K*i ^,l^.Ks a K*

where

As regards the significance of the s3rmbols, it will be recalled that A
is the scalar value of the molecular field, A^ the most probable value of
A, F the scalar value of the resultant magnetic field, k Boltzmann's
constant and T the absolute temperature.

From the expression (25) for the magnetisation per unit mass, we
now derive an expression for x, the susceptibility per unit mass.
By definition:

VdH/ VdK dH/

For isotropic sobetanoes, with which the theory is concerned, K and H
will be oollinear, and from (24) we find:

PARA' AND DIAMAGNETISM: WILLS 79

dK , . 4ir dl
dH ^ 3 dH

and, ance for paramagnetic substances the second term on the right
will be very small in comparison with unity, it may be neglected. We
may therefore write:

dM

X=Lt K-«

dK

In the evaluation of the ri^t hand member of this expression the
+ sign in the lower limit of the integral involving F in expression (25)
for M is to be used, since in the limit E will be less than A. It is found
after easy calculation that:

(26) x= z^f'^/^J 1 L (a) + -ga L' (a) j e " *' A dA,

where

L(a) = cotha- - ; a^i^;

a kT

and L'(a) is the differential coefficient of L(a) with respect to a.
For brevity we now write:

(27) z - ^„

kT

^"mA.'

4Nm

by (34) Sec. Ill:

(28)

^ 3*^^ Ms*'

where No is Lioschmidt's number, M the molecular weight, p the
density and s the smallest distance of approach between two magnetons.
Upon introducing the abbreviations into (26) we finally obtain:

(29)

-«.]"{M^)+i^L',f)}e-d..

This formula implies a dependency of the susceptibility upon the
temperature, since t is proportional to T; and also upon the density,
since r and x© are each inversely proportional to the square root of the
density.

For liquids and solids, however, variations of the density with tem-
perature may be disregarded.

80 PARA' AND DIAMAGNETISM: WILLS

For brevity kt:

(80) ^ - e ;

then, from (27) :

(31) r - -| .

Upon intiodueing the temperature function:

we obtain from (29) :

(83) --*(^) = *(r),

a formula involving two disposable constants, Xo ^^^ ®- This formula
implies that, with the exception of gases, all paramagnetic bodies obey
a law of corresponding states.

The value of the temperature function ^ (r) is now required. It is
convenient to derive expressions for ^(r) for two cases; vis., when r is
small, and when r is large. In the first case it is to be understood that
r is not so small as to take the theory out of the equipartition range.

Case 1. r small.

For details of the calculation the reader may refer to the original
paper. The result of the calculation is to show that:

(84) *w=i-I^'+|(„)*+^V)»+^V)' ;

the B's represent Bernoulli numbers a few of which are:

_ 1 _ 1 _ 1 _ 1.

' 6' 30' 42' 30

Case 2. r large.

The details of the calculation are also omitted in this case. It is
found that:

(36)

^' 2 I 1! T 21 T» 3! T» /

B Sf

For very high temperatures, terms after thejfirst on the right of (35)
may be neglected; it is then foimd from (33)^and (31) that:

^ ^ X- ^ ^«T"3kT

which is the Curie-^Lang^vin law for paramagnetism.

PARA- AND DIAMA0NETJ8M: WILLS

81

This result was to be expected, sinoe at high temperautres the influ-
ence of the molecular field upon the niagnetons is small in comparison
with the disorganizing effects of thermal agitation.

Experimental Test of Theory.

The theory is compared by Gans with experimental determinations
of the susceptibility by K. Onnes, Oosterhuis, Perrier and Honda.

For Crystalline Gadolinium Sulphate (Gds(S04)a HsO), and for Ferric
Ammonium Sulphate (FesS04(NH4)sS04+24HtO, the Curie-Langevin
law is found to be well obeyed down to the respective temperatures
T=20.1'*K, and T^U.T'K. On the present theory, for these two
substances, and in fact for all for which xT is constant, the molecular
field Aq is so small that G will also be small, so that T/G will still be a
large number. The inference is that here the mutual action of the
magnetons may be ignored.

The substances listed in Table III, with the values assigned to the
disposable constants Xo ^^^ ^ show, as regards their susceptibilities,
agreement with the present theory which leaves little to be desired for
temperatures as low as 14.7^K.

Table III

Substance

Formula

Xo

e

GryBtfiUine ferroua sulphate

Ciystalline manganous sulphate. .
Water-free ferric sulphate

FeSOi .7H,0
M11SO4 .4HtO
Fe,(S04),

2212X10-*

4837X10-*

302X10-*

12.64

9.90

120.00

Molecular constants, — The theory furnishes, with the aid of experi-
mental results for the substances above considered, values for the fol-
lowing constants:

The nimiber of Weiss magnetons per molecule.
The most probable value for the molecular field A.
The smallest distance of approach, s, between two magnetons.
For very high values of T we have, from (36) :

(87)

Nm* Vt

3k

XoQ.

Now, since it has been assumed that each molecule contains only one of
the magnetons of the present theory, /aNo will be equal to the magnetic
moment per gram molecule. No being the Loschmidt number with the
value 6.175X10*'; and No = MN where M is the molecular weight.
Upon multiplying the preceding expression, (36), by MN; substituting:
No for MN, and solving the resulting equation for mNo, we find:

82

PARA' AND DIAMAONETISM: WILLS

(38)

V!

MNo-^7\/irkNoMxoe

as the magnetic moment per gram molecule.

If the molecule contain q magnetic atoms, then, in accordance with
Weiss, mNo/q is an integer multiple, p, of 1123.5. Thus:

(39)

V

1123.6 p -V-v^kNoMXoS^ q, (kNo- 8.316 X10»).

We denote by p' the nearest whole number to the value for p calcu-
lated from this equation.

Weiss usually assumes q » 1 f or salts, such, for example, as Fei (S04)s,
containing more than one metal atom.

The most probable molecular field is calculated from the second of
equations (27) as follows :

(40)

^kT kN^G

kNoS

fjLT mNo 1123.6pq

The smallest possible distance of approach, s, between two magnetons
is obtained from (28) : <^

(41)

8« =

16ir M*NoP 16ir 1123.6* qVP.

9 MAJ 9

N,

Using the values of the constants Xo ^^^ ^ given in Table III, the
results given in Table IV are obtained for Cr3r8talline Ferrous Sulphate,
Crystalline Manganous Sulphate and Water-free Ferric Sulphate.

Table IV

Substance

M

p

xoXlO*

e

P'

P

26
29
36

AoXlO-*
in Gauss

sXlO*
in cm.

MnS04 .4H|0

Fe(804).(q-1)

278.0
223.1
390.9

1.90
2.11
3.10

2212
4387
302.0

12.64
9.90
120.

20.09
29.13
35.63

0.3587
0.2516
2.494

3.46
5.25
1.22

Remarks —

^

It will be noticed that the values for p' do not approximate very
closely to integer niunbers; and the Weiss magneton theory here fails
of any very substantial support. This circiunstance is, however, with-
out influence upon the other molecular constants concerned.

The molecular fields are seen to be quite large. Water of crystalli-
zation appears to have the effect of decreasing the molecular field,
owing probably to increase in the smallest possible distance of approach
of neighboring magnetons.

PARA' AND DJAMAONETISM: WILLS 83

The Bmallest distance of approach, s, is of the order of one tenth the
diameter of a molecule. This may be explained by supposing the
magneton excentrically placed in the molecule.

Although the present theory is in good agreement with experiment
down to very low temperatures for the substances considered above it
breaks down (at very low temperatures) for many others. Gans has
therefore proposed a modification based upon a quantum hypothesis.
This modification will be considered in Section VII, deaUng with
quantum theories of magnetism.

Theory of Honda and Okubo.

In a paper entitled "On a Kinetic Theory of Magnetism in General'^
Honda and Okubo^ have attempted a modification of Langevin's
theory for a paramagnetic gas, in which, effects due to the rotations
of a magnetic molecule about axes perpendicular to the magnetic
axis are taken into account.

The vector magnetic moment of a molecule is considered as made up
of two parts: an axial component in the direction of its axis of rotation^
and a transverse component perpendicular to this axis.

In accordance with the argument advanced in the paper cited the
axial components of the magnetic moments of the molecules of a body
subject' to an external magnetic field would, due to the motions of the
molecules induced by the field, give rise to paramagnetism; and the trans-
verse components to diamagnetism.

The theory has much in conmion, as regards its fundamental assump-
tions, with Cans' theory of dia^, para-, and metamagnetism which has
been reviewed in some detail above.

The arguments of Honda and Okubo have been subjected to rather
severe criticism by Weaver.*

Theory of Oxley.

In an extended series of very interesting papers entitled "On the
Influence of Molecular Constitution and Temperature on Magnetic
Susceptibility," A. E. Oxle}^^ has introduced a modification of Langevin's
theory, in which the molecular field plays a leading role in diamagnetic
substances, as well as in para,- and ferromagnetic substances.

The theory of Oxley, bringing into prominence, as it does, the mole-
cular field, is analogous in many respects to the theory of ferromagnetism
developed by Weiss upon Langevin's theory of a paramagnetic gas as a
basis, supplemented by the assumption of the existence within ferro-
magnetic substances of enormous internal fields.

1 Honda and Okubo: Phy. Rev. 13, p. 6; 1919.

* W. Weaver: Phy. Rev. 16. p. 438; 1920.

• A. E. Ozley. Roy. 8oc. Pha. Trana. 214. A. p. 109; 1913-14.— 215 A, p. 79; 1914-16.
—220 A. p. 247; 1919-20.

84 PARA' AND DJAMAGNETJSM: WILLS

It therefore appeared appropriate to treat the work of Chdey and of
Weiss together in a separate contribution. This has been done by Pro-
fessor E. M. Terry in the part of this report dealing with ferromagnetism.^

Theory of Frivold.

In a paper entitled "Zur Theorie des Ferro- und Paramagnetismus
O. E. Frivold^ has developed a theory of ferro- and paramagnetism,
consisting in a modification of Langevin's theory for a paramagnetic
gas, in which the molecular magnetic field is taken into account.

In this theory the elementary magnets or magnetons are identified
with the atoms whose centers are supposed fixed at the comers of a
cubic space lattice, and capable of rotation about their respective
-centers.

Statistical theory is applied to this system of magnetons, and results
found from which the magnetisation curve may be obtained. Com-
parison of this curve with the corresponding one which results from
the Langevin theory furnishes a measure of the efifect of the mter-action
of the magnetons, and permits the calculation of the magnetic molecular
field.

A more detailed account of this theory is given by Professor Terry
in the section of this report referred to above.

While other attempts toward the improvement of Langevin's equipar-
tition theory of magnetism have been made, it is hoped that the considera-
tion of those which have been presented here in more or less detail will
serve to enable the reader to form a fair idea of the trend of attempted
improvements on this justly celebrated theory.

^ cf. p. 154 of this report.

* O. E. Frivold. Ann. d. Phys. 65. p. 1 : 1921. cf. p. 132 of this report.

PARA- AND DIAMAGNETISM: WILLS 85

VII
THEORIES OF PARAMAGNETISM BASED ON QUANTUM HYPOTHESES

In 1911 Nemst^ showed, in contradiction to the laws of classical
statistical mechanics, that the specific heats of polyatomic gases appear
to decrease with decreasing temperature. This was confirmed later
by the investigations of Scheel and Heuse,' and their results ascribed to
the behavior of that portion of the specific heat which depends upon
the rotation of the molecules.

There then appeared a series of investigations having to do with the
rotatory energy of molecules. Of these some were of a theoretical
nature in which attempts were made at quantiticing the rotatory energy.

Meanwhile the experimental investigations of Onnes, Oosterhuis,
Perrier, du Bois, Honda and Owen on the variation with temperature
of the susceptibility of paramagnetic substances gave results which
were in opposition to equipartition theories of paramagnetism. The
theory of magnetism was thus in a similar dilemma to that in which
the theory of specific heats found itself.

Modifications of existing theories of magnetism through the intro-
duction of quantum hypotheses were, of course, in order. The earlier
theorists in this field were faced with a fundamental difficulty, shared
by some of those working at the improvement of the theory of specific
heats, which had its origin in the attempted quantitization of the rotary
energy of the molecules.

Poincard at the Solvay Congress in 1911 called attention to the
difficulty as follows:

''Imagine an oscillator with three degrees of freedom, isotropic and
capable of vibration in such manner that the periods of vibration are
the same with respect to three axes. Thus, for motions parallel to the
(x, y, z) axes, let the corresponding energies be respectively ohu, 0hv
and 7hi;, where a, /3, y are all integers, h is Planck's constant, and u, the
common frequency. Let the axes now be changed: with respect to the
new axes the energies will be ahv, fi^hv, and 7'hu, where a', jS', / are
integers. This is impossible."

In reply Planck said:

"An hypothesis of quanta for plural degrees of freedom has not yet
been formulated, but I believe it to be nowise impossible of achieve-
ment."

In 1916 Planck,' through the publication of his paper on ''Die Ph3rsi-
kalische Structur des Phasenraiunes," demonstrated the correctness of
his view here expressed.

> W. NenuBt: ZeiUdir. /. EUktroihem, 17, p. 015; 1911.

• K. Scheel u. W. Reuse: Berl, Ber. p. 44; 1913; Ann, d. Phyw. 40, p. 473; 1913.

> M. Planck: Ann. d. Phya. 50, p. 385; 1910.

86 PARA' AND DJAMAGNETJSM: WILLS

Prior to the publication of Planck's paper writers attempting to
improve magnetic theories through the introduction of quanta hypoth-
eses were forced to make such assiunptions as seemed plausible, yet not
firmly based.

We shall therefore pass over with but brief mention the earlier at-
tempts at quantum theories of paramagnetism.

Theory of Oosterhuis.

Among the first in this field was Oosterhuis^ who proposed a modifi-
cation of Langevin's equipartition formula for the susceptibility per
unit mass:

^"skT

where N is Avogadros's nimiber, m the magnetic moment of a molecule,
and k Boltxmann's constant. Here kT represents the mean energy
per degree of freedom of a molecule, and Oosterhuis simply replaces
this by the expression

1 / hw hiK

ekT _i

representing the mean energy of rotation of the molecules for one degree
of freedom on the quantum hypothesis of Einstein and Stem, which
assumes all molecules to rotate, at a g^ven temperature, with the same
angular velocity, v being the common frequency of rotation and h
Planck's constant.

Theory of Keesom.

Keesom' does not assume with Oosterhuis that all molecules at a
given temperature in a substance rotate with a conmion angular velocity,
but considers the motions of molecular rotation to be resolved into a
system of standing elastic waves, after the manner of Debye in his
theory of specific heats. Owing to the discrete structure of matter,
waves with a length shorter than a certain minimum determined by the
structure are not possible of existence, and consequently the number of
possible frequencies for the standing waves will be finite and all below
a certain maximum, vm say. The magnetic molecule, as with Oosterhuis,
is supposed to have a negUgible moment of inertia about its magnetic
axis, while its other principal moments of inertia are supposed equal.

The mean rotational energy corresponding to a single degree of free-
dom is then f oimd to be

I E. Oosterhuis: Phy. ZeOeehr. 14, p. 682; 1913.
I W. H. Keeoom: Phy, ZeiUehr. IS, p. 8; 19U.

PARA' AND DIAMAGNETISM: WILLS 87

•m

L f h«^ .1

and this expression on Keesom's theory replaces kT in Langevin's
formula for the susceptibility of paramagnetic substances.

The theory of Oosterhius shows fairly good agreement with experi-
ment, in fact about as good as that of Keesom, and as it is foimded upon
far simpler assumptions is to be preferred.

The Theory of Gans.

In his paper ''Uber Paramagnetismus/'^ which appeared in 1916,
and which has been reviewed in Section V as far as the part which
deals with the equipartition portion of the theory is concerned, Gans
proposes a quantum modification, in order to obtain a theory which
will be applicable to all paramagnetic substances at very low tem-
peratures.

As was stated in Section V, his equipartition theory is in good agree-
ment with experiments in the case of some substances down to very
low temperatures. But susceptibility curves, (x-T), of observations on
Uranium, Magnesium, Aluminium, Molybdenum, Mobium, Tantalum,
and Wolfram all show a tendency at some point to become parallel to
the T-axis; in fact this tendency in the case of some of these substances
is evident at room temperatures; and in the case of Molybdenum and
Wolfram at temperatures of 1200®C and 1100*'C, respectively.

These experimental results cannot be accoimted for on his equipar-
tition theory; and Gans was thus led to modify it through the intro-
duction of a quantum hypothesis relating to the distribution of the
rotatory energy of the magnetons. As in the case of his equipartition
theory, Gans takes the molecular field into accoimt in his modification.

It is important to remember that the quantum theory of Gans is
only applicable for very low temperatures, where by the term low
temperatures is meant temperatures at and below which the equiparti-
tion theory is no longer valid; thus in the case of Molybdenum and
Wolfram temperatm^s below llOO^C are considered as low tempera-
tures.

At very low temperatures it may safely be assumed that temperature
agitation is so slight that the magnetons perform but small vibrations
about their positions of equilibrium, which are determined for any
magneton, in the absence of an external magnetic field, by the molecular
field A at that magneton. In fact the vibration frequency, v, for the

>l.c.

88 PARA' AND DJAMAQNBTJSM: WILLS

magneton, and the most probable value of v, say v^, are respectively
given by:

where J represents the moment of inertia of the magneton about any
axis through its oentroid perpendicular to its magnetic axis.

The quantum assumption now made is, that the energy distribution
for the two degrees of freedom of the magneton about two perpendicular
axes in its equatorial plane is the same as that which would obtain if
each degree of freedom be treated as though it were that for a simple
oscillator with this one degree of freedom.

To give precision, then, to the fundamental assumptions now intro-
duced, it is supposed that the typical magneton with moment m finds
itself in a magnetic field F, and that the temperature is so low as to
allow it to perform infinitely small vibrations about its equilibrium
position determined by the direction of this field.

Let 01 and 0s be the angular displacements of the magneton about
two perpendicular axes, then the total energy, e, of the magneton will
be g^ven by:

(2) e =y («!«+«,«) + ^ W+«.«) = ^ (Ci«+C«),

where Ci and C% are the maximum amplitudes of 0i and 0s, respectively.
If j8 denote the angle which the magnetic axis of the magneton makes
with the resultant field F in which it finds itself, then:

cos^-1-- =1 ^— ;

and the mean value in time of cos j8 will therefore be given by:

^ 2 4

and hence, with the aid of (2) :

cos/J « 1 —

2mF

The spatial mean value of this expression over all the N^magnetons
in a unit mass will be expressed by:

where e is the mean energy of a magneton.

PARA' AND DJAMAGNETISM: WILLS 89

In accordance with the quantum hypothesis made by Gans:

2hu

(8) €= -h7 ,

kT

e - 1

the expression on the right being twice the mean energy assigned to
each degree of freedom of the magneton, conforming with Planck's
original theory of radiation which implies no zero-point energy.
From the last two equations it follows that:

hv 1

(4) ^^ ^ "" ^ "■ i^ "E

V

kT

e -1

This expression corresponds on the equipartition theory to Formula
(30), Sect. Ill viz. :

kT / kT*

cos P == ooth

The equipartition theory is therefore modified in accordance with
Gans's quantum hypothesis by replacing in (25), Sect. VI,

_ mF . /mF, . hu 1

^*^kT-VkT^yi-,-F-ir=i-

e

The subsequent development, taking account of the molecular field,
is along lines closely analogous to those followed in the equipartition
theory. For the details the reader may consult the original paper.

The theory furnishes an expression for the susceptibility which
contains three arbitrary constants: Xoi the susceptibility at absolute
zero; $ {^hv^/k); and 0 (^Mo/k).

In the case of Platinum and Water-free Manganous Sulphate, with
the values of the disposable constants given below, the theory is foimd
to be in good agreement with experiment:

Xo G $

Platinum 1.189xl0-« 2097.^ 60.0^

Water-free Manganous Sulphate 670. x lO"* 84.94® 23.5®.

Molecular constants. — From these experimental results interesting
information as to the following molecular constants may be obained:

(a) The most probable vibration frequency, u^, of the magnetons
in the molecular field.

(b) The equatorial moment of inertia, J, of a magneton.

90 PARA' AND DIAMAGNETJSM: WILLS

The most probable vibration frequency, v^, for a magneton in the
molecular field is given by:

From the second of equations (1) we have for the equatorial moment
of inertia of a magneton:

4ir«wo^

where A^ may be calculated as in (40), Sect. VI. It is foimd that for
Platinum A^b 1243X10*, and for Water-free Manganous Sulphate
A^«2.292X10».
Thus the following values are obtained :

u^XlO"" JXIO**

For Platinum 1.30 67.7

For Water-free Manganese Sulphate . 0.483 12.4.

Theory of von Weyssenhofif.

Jan von Weyssenhoff,^ in a paper which appeared in 1916, appears to
have been the first to evolve a quantum theory of paramagnetiBm in
which the method operates explicitly with quanta from the beginning.
This author avoids the difficulty brought forward by Poincare through
the introduction of a simplified model to represent the structure of
paramagnetic bodies.

In this simplified model the magnetic molecules (magnetons) are sup-
posed capable of rotation only about axes parallel to a given plane,
(the x-y plane), and also perpendicular to their own magnetic axes. The
angle between the z-axis and the magnetic axis of a magneton is denoted
by $. The position of a magneton is then uniquely determined by some
value of B between — r and r . It may reasonably be expected that such
a model will show, as regards its magnetic properties, a behaviour
similar to a more general one in which the magneton may turn freely
about a fixed point.

An external field of strength H is supposed to act in the direction of
the z-axis.

The potential energy, U, of a magneton with magnetic moment m
is expressed by:

(7) U= mH (1 - cos ^)« A« sin* ^ , where A« - 2mH;

> J. yon Weyaienholf : Ann, d. Phya. 51, p. 285; 1916.

PARA' AND DIAMAGNETJSM: WILLS 91

and the kinetic energy, E, by:

(8) E = ^J*« = ~^, where^ = M,

and J 18 the moment of inertia about the fixed axis of the magneton.

In the present theory the mutual magnetic inter-action of the mole-
cules is not taken into account. Hence when A^O a magneton may
turn freely about its fixed axis. For very large values of A all the
axes of the magnetons will deviate but little from the direction of the
external field H, and they will then behave in a manner quite similar to a
system of Planck linear oscillators. For, the total energy of a magneton,
6, which in the general case is given by:

(9) e= 2 J^+A« 8in«^ = ^ + A« sin* |
will in this case be expressed by:

^^ 2 4 2J 4

which is an expression identical in t3rpe with that for the energy of one of
Planck's linear oscillators.

It is now proposed to apply to this model the second quantum theory
of Planck, or rather, that portion of it which is termed by him thermo-
dynamic.

To this end it is first necessary to consider the phase domain appro-
priate to the model. This consists of a strip of the ^— ^ plane of breadth
2ir, parallel to the ^-axis. Here 6 and ^, already defined above,
may be designated respectively as the generalized coordinate 0 and the
generalised momentmn ^:

(11) ^"^'^^2J^>=J^-

The method of Planck now requires the calculation of the magnitude
and form of the elementary domains in the O—^f plane of equal proba-
bility.

In accordance with Planck's ideas these elementary domains of equal
probability must be bounded by curves e » const. For large values
of A these curves must be ellipses, as is evident from equation (lO).

The magnitude of each of the elementary domains must be the same
and equal to Planck's constant, h, since for large values of A the mole-
cules of the model are equivalent to a system of linear oscillators for
which, as shown by Planck, the magnitude, h, of an elementary domain
is independent of u, and hence of A.

92 PARA' AND DIAMAGNETISM: WILLS

The family of bounding curves, c » const., for the elementary dom-
ains is given by equations of the type:

(12) ^+A«8in«-«C^,

where C is a constant for any given curve.
The area bounded by any such curve will be given by:

(13)

/ ^ d ^=4 /\/2J(C?- A« 8in«| d 6,

the limit of integration, g, depending upon the value of C.
It is now required to find a series of values for C:

such that the area of the elementary domain between the (n — 1) st
curve and the n'th curve shall be equal to h for all values of n; or, what
is the same thing, that the area enclosed by the n'th curve shall be
equal to nh.
We have, with the aid of (13) :

(15) 4/v^ VCa*- A«sin«-da - nh,

o ^

^f2sin-iC./A forC„ < A
^ \ IT for C„ > A.

The curves on the ^^ plane represented by equation (12) for
different values of C are separated into two distinct classes; one class
lying within the curve G, shown in Fig. 10, for which the external field
H is such that C«A; and the other class l3ang without this curve.
The values of C„ for the first class will all be less than A, while the
values of Ca for the second class will all be greater than A. For the
requirements of a theory of paramagnetism it will appear presently
that only the second class need be considered.

The case when the external field H is such that C^ »: A is interesting
as representing the case in which the pendulous motion of a magnetic
molecule is about to pass into rotary motion. The area of the curve
G for this case is easily seen from (13) to be expressed by:

(16) 4AV2J/co6- dd = 8A\/2J.

o ^

If it were possible to express quite generally Cb as a function of n
by means of (15), a formula for the mean energy of the magnetic
molecules could be at once derived; also it would be possible to derive
an expression for the orientation of the axes of the magnetic molecules
as a fimction of the temperatm^ for a given field strength. Un-

PARA- AND DIAMAONETISM: WILLS

93^

fortunately, this general procedure is not possible, and the argument

has to be restricted to special cases. It will appear, however, that one

of these special cases is broad enough to furnish a basis for an explanation

of paramagnetism.
For the case in which the external field H is so large that the area

of the curve given by (13) for 0^= A, viz., 8A\/2jr is much greater

than h, all the elementary domains com-
ing into consideration will lie within the
curve G, and quite near the origin; and
since d may now replace sin $, the
bounding curves of the elementary do-
I mains will become ellipses, one of which
is shown by the dotted line in Fig. 10.
This corresponds exactly to the case of
Planck's linear oscillators. The attain-
ment of this case, however, would require
external fields far greater than can be ob-
tained in practice.
We now consider the special case in which the external field H is such

that the area of the curve G, viz., 8A\/2J is far smaller than the

quantum h.
In this case:

(17)

H <

y

256/* J *

As regards order of magnitude, m = 10""^°, J = 10"*° and h=6.55
X 10"*'. Hence the order of magnitude of the right hand member of
the inequality (17) will be 10*. This number represents a field con-
siderably greater than any that can be obtained in practice and we
may conclude that a theory of paramagnetism may be foimded upon
this special case.

Now if, for the moment, we consider the external field to be such
that Ci = A, then the area of the curve G will be such that 8A\/2J =
h; and it follows that the elementary domains coming into considera-
tion in the present case, where 8ir\/2J is very smaU in comparison
with h, will all lie outside the curve G. One of these is shown by the
shaded area in Fig. 10. The upper limit of the integral in (15) will
therefore be v, and the integral itself will therefore be a complete
elliptic integral of Legendre.

For the case of paramagnetism we have, therefore :

(18)

h = 4V2j| VC„* - A*sin»^dfl.

94 PARA' AND DIAMAONETJSM: WILLS

Writing:

^ c.'

the integral can be put in the fonn of a series:

<19) n h - 4x>/2J C»|l - (i/ k««-(-J4)'-^ ... I

From (19), 0^ has now to be found as a function of n and A. We pass
over the details of the calculation which may be found on page 301
of the paper under review. The calculation is simplified by the fact
that A may be considered as a small qhantity. The result shows that:

where

4irv/2J

n n

Now let:
N be the total number of magnetic molecules per imit mass;
N <a^ the number of magnetic molecules per unit mass with energies
between the limits specified by the boimdaries of the n'th elementaiy
domain;

e^ the mean value of the total energy for the N ta^ magnetic molecules.
Also let:

(22) *n » C. V2J y 1 - V sin« ^ , where k„ - p ,

2 ^«

express the value of ^ for any point on the n'th boundary curve, ob-
tained from (12).
Then:

(23) ^-^//(|j + A«Bm«^)d*d*,

where the integration is over the nHh elementary domain. The result
of the evaluation of the integral in (23) is to show that:

kA A* 'T^J 1 1

(24) ...ke(n-n)+f. + - + -^n^— ^-i).

where

h*

(25) ke « -^^.

The constant 8 has the dimensions of temperature.

PARA' AND DIAMAGNETISM: WILLS 95

From here on the calculation follows the lines laid down by Planck
in the development of his second radiation formula in which the oscil-
lators are supposed to absorb energy continuously and to emit it in
quanta.

The total energy, W, of the N magnetons considered is given by:

(26) W = N 2 «„ €„,

and this being supposed specified, the well known thermodynamic
method of Planck^ leads to the law of distribution of energy:

(27) Nwn = ae"^T « ofje ,

where ai is a constant which depends upon A and T but not upon n.

Equation (26) gives the law of distribution of the magnetic molecules
as regards their energy, that is, the number of molecules per unit mass
with energies lying between the limits specified by the boundaries of
the n'th elementary domain.

The results so far found are capable of direct application in the
theory of rotatory specific heats, and of paramagnetism. We pass over
the part of the paper having to do with the theory of specific heats
and consider now the application of the results found to a theory of
paramagnetism.

The potential energy of a magneton, from (7), is given by:

(28) u = ~ (1 - cos e) = A' sin* ^.

^ 2

If X he the magnetic susceptibility per unit mass, then, as on Lan-
gevin's theory:

Nu

(29) ^ " II ^^® ^'

where cos 0 is the spatial mean value of cos d, whose value on the present
quantum theory will, of course, be different in general from that found
on the equipartition theory of Langevin.

Flt>m (28), if U denote the spatial mean value of U:

2-

(30) cos^ = 1 - -jU.

A

Now if Uq denote the mean potential energy of a magneton whose
total mean energy, e^, is specified as being within the boundary limits of
the n'th elementary domain (whose area on the ^ plane equals
h), then:

1 M. Planck: Vorlesungen Qber die Theorie der W&rmBirmhlanc — ^Dritter AbBchnitt.

M PARA- AND DIAMAQNBTISM: WILLS

U. -^/A»8m»^f.-f._.)d*.

This equation, after the evaluatioii of the integral, with tlie aid of (22),
and taking note of (20) and (21), gives:

A* 2ii*JA*
Therefore the mean values of cos 9 in the n domains will be given bjr:

^i ,% 1

(31)

(co8tf).--^A'-— mH, (n-l).

4**J »t , 1 1

(COS*).- -rr A' (--—-)
n n n— 1

A* 1 mH

8ken(n-l) 4ken(n-l)

These equations, with the aid of the distribution function given by
(27), enable us to derive directly the following expression for cos di

• 1 -|(n«-n)

1 — S e T

mH 2 n (n-1)

(32) cos e

4k0 ; e^„, . „,

From (29) and (32) we obtain the following expression for die mag-
netic susceptibility per unit mass:

_ 1

?i??^-.T"~t n(n-l)
h»

(33) X - ^^^ir*J

2e
I

where

(34) T-^-

- « (n> - o)

T 32««JkT
From (33) it follows that at suffidently high temperatures:

'^ "2kT

PARA^ AND DIAMAGNETISM: WILLS 97

which agrees with the Langevin f onnula except that there here appears
in the denominator a factor 2 instead of a 3, as in the Langevin formula.
The model for the molecular structiu^ here adopted allows, however,
but one degree of rotary freedom for the magnetic molecule and, if
the Langevin calculation be carried out under the assumption of but
one degree of freedom for the magnetic molecule, it turns out that the
numerical factor in the denominator would be 2 instead of 3. There-
fore the author introduces the factor 2/3 on the right of formula (33).
The final formula for the magnetic susceptibility then becomes:

A 1-

M g_#(n«-n)

(35) x^'4^^1 ^iil^)-

1

This formula gives for the mass susceptibility at absolute zero:

16 Nm*

(36) ^ '^ ^ "p" *" ^*

A test of the theory is made through comparison of values of x>
calculated (with appropriate values of the disposable constants Xo &^<1
6) from the experimental values of x determined by Onnes and Ooster-
huis for crystalline- and for water-free manganese sulphate, with re-
sults given in Tables (V) and (VI) below.

Theories Based on Planck's Method of Quantitization.

Following the appearance in 1916 of Planck's paper^ on "Die physika-
lishe Structur des Phasenraumes," which set forth the procedure to
be followed in quantitizing the energy of an oscillator with plural de-
grees of freedom, the time was ripe for fiuther improvements in the
theories of rotatory specific heats and of paramagnetism.

As mentioned above the point had previously been reached in the
development of theories in both of these subjects where a method was
required for the quantitization of the rotatory energy of a molecule,
or magneton, with plural degrees of freedom of rotation.

In theories of magnetism the magneton commonly hypothecated was
supposed to have a constant magnetic moment due to its rotation
about an axis of sjrmmetry, and to possess dynamic S3rmmetry about
axes through its centroid perpendicular to the axis of symmetry; and,
since the requirement of constancy for the magnetic moment of the
magneton about this axis demands that its motion about it be inde-
pendent of thermal agitation, only two degrees of freedom were assigned
to it.

> I.e., p. 85.

98 PARA- AND DIAMAONETISM: WILLS

The definite problem up for iolution before satiflfactoiy progresB
could be made was:

To quantitize properly the rotatory energy of a magneton with two
degrees of freedom of rotation.

In Planck's quantum theory of radiation the quantum difficulty of
Poincare, stated above, does not arise, since the linear oscillator in-
voked by Planck for the purpose of effecting interchange of energy of
different frequencies in black body radiation has but a single
degree of freedom. The probability elementary phase domains for a
linear oscillator were shown by Planck to be the areas included between
consecutive ellipses similar and similarly placed in the ^ plane,
each area on his quantum hypothesis being equal to the imiversal
constant h; ^ being the generalized coordinate of the oscillator rep-
resenting its electric moment and ^ the corresponding generalized
momentum, viz., the partial derivative of the kinetic energy of the

oscillator with respect to the generalized velocity ^.

Now from the viewpoint of Planck the quantum difficulty of Poincare
may be stated as that of correctly delimiting the elementary proba-
bility domain in the specific problem under consideration. If this
delimitation be accomplished, the remaining difficulties are simply
those of formal anal3rsi8.

In cases where the statistical element or molecule has but a single
degree of freedom the proper delimitation of the elementary proba-
bility domains is generally a fairly simple matter, as in the case of
Planck's linear oscillators, or again, in the case of the constrained
motion of the magnetons in the model of molecular structure assumed
by V. Weyssenhoff in his theory of paramagnetism.

We shall now notice briefly some quantum theories of paramagnetism
based on Planck's method of quantitization.

Theory of Reiche.

Fritz Reiche^ in 1917 published a very interesting paper entitled
"Zur Quantentheorie des Paramagnetismus" in which he generalizes
the assimiptions of v. Weyssenhoff as regards molecular structure by
considering it to be such that each magnetic molecule (magneton with
fixed magnetic moment) should be capable of free rotation about a fixed
point. The rotation of the magneton about its magnetic axis (axis
of synmietry) is supposed independent of thermal agitation and its
moment of inertia about any equatorial axes through its centroid,
denoted by J, is assumed to be the same for all such axes. ^^

From what has been said above it will be dear that the problem^bf
Reiche differs essentially from that of v. Weyssenhoff only in that

>F. Reiohe: Ann, d. Phy. 54, p. 401; 1917.

PARA' AND DIAMAONETISM: WILLS 99

part which has to do with the delimitation of the appropriate elementary
phase domains. To go into the details of the anal3rsi8 whereby this is
effected, following the method of Planck, would carry us beyond the
scope of the present review and the reader who is interested is referred
to the original paper; also to a paper by Adams. ^

The author finds an expression for the mean value of cos 6 for the
magnetic molecules, where d is the angle between the magnetic axis
of such a molecule and the direction of the external field H; and then
substitutes this in the following expression giving the magnetic sus-
ceptibility per unit mass:

N/i

X = ^ cos ^ ,

where N is the number of magnetic molecules per unit mass, /i is the
magnetic moment of a molecule, and cos ^ is the mean value of cos 9.
It is thus found that:

(37) ^e-'+^^ •*"'■'

V = Nm» j4 3„..n(n«-l).

1

where

(38) <r -

8ir*JkT

The corresponding expression for x found by v. Weyssenhoff is
given by (35), and it should be noticed that c in the theory of v.
Weyssenhoff has a value equal to one fourth of that given by (38).

For very low temperatures (o large) formula (37) gives:

6 Nil*
(39) x = 7-rr*'J;

4 h*

while (35) reduces to:

(40) ^'J^^^-

For high temperatures (a small) both (37) and (35) give:

^'■3kT'
the equipartition expression of Curie-Langevin.

> E. P. Adams: BuU, Nai. Ru, Caun. I, 5, p. 301.

100 PARA' AND DIAMAGNETISM: WILLS

Other Theories.

Sophie Rotssajn^ treated the same problem as that oonaidered by
Reiche, using, however, a very di£ferent method of analyaia. The
final formula found for the misceptibility is, as was to be expected,
precisely the same as that arrived at by Reiche.

The procedure followed by both Reiche and Rotszajn as regards
quanta hypotheses presupposes the validity of what is commonly known
as Planck's second theory of radiation, which assumes the absorption
of energy by his linear oscillators to be continuous and the emission
to be in quanta; and which predicts the existence of a zero-point energy.
It will be recalled that on this theory the distribution of energy, for
the stationary state, as regards frequency u, is, if c denote the mean
energy of an oscillator of frequency v, given by:

"fe*^)

where h is Planck's and k Boltsmann's constant.
In the first form of Planck's theory, eventually discarded by him:

hi;

kT

e" - 1

and thus does not predict the existence of a zero-point energy.

A. Smekal,* in spite of the fact that the second fonn of Planck's
theory is now conmionly preferred to the first, thought it worth while
to develop a quantum theory of paramagnetism based on the assump-
tions of the first form of the theory, using the same magneton model as
that assumed by Reiche and Rotszajn. He was led to a fonnula for
the susceptibility which shows by no means so good an agreement with
the experimental facts as that found by them on the basis of the second
form of Planck's theory. His result, then, adds another argument in
favor of the second form of the theory and, therefore, for the existence
of a zero-point energy.

Comparison of Theories with Experiments.

Of the various quantum theories which have been considered probably
the most satisfactory is that of Reiche which, as far as the fundamental
assumptions and the final results are concerned, is the same as that of
Rotszajn.

The theory proposed by v. Weyssenhoff is also satisfactory from the
standpoint of its development from his fundamental assumptions; but

1 8. Botssajn: Ann, d, Phu%, 57, p. 81; 1918.
• A. Smekal: Ann, d. Phut. 57, p. 376; 1918.

PARA- AND DIAMAGNETISM: WILLS

101

these are more artificial than those of Reiche and Rotszajn, including
as they do the restriction of the movement of the magneton (apart
from its rotation about its axis of symmetry) to motion in two dimensions.

The theory of Gans, while based upon an incorrect quantmn hypoth-
esis, takes account of the consequences of the presence of the ''mo-
lecular field" which is ignored on other theories. Comparison of this
theory with experiment has already been made (see p. 96 of M. S.)

The theory of Oosterhuis may be taken as representative' of those
theories, other than that of Gans, based upon quanta hypotheses which
were developed before Planck in 1916 published his general method
whereby quantitization may be effected in a statistical system whose
elements have plural degrees of freedom.

Remarks —

Table V

Wateb-fbee manganese sulphate — ^MnSOi
Reiche ...J = 1.99X lO""**;

k-41

-21

V. Wey. ... J = 4.44 X 10""; m = 4.35 X lO'^^ Xo = 6-577 X 10

k-41

-20.

Oost. . . . J = .87 X 10""; /i = 1.80 X lO"""; Xo = 6-89 X 10

k-6

•TK

xxl0*cal.

xX10*cal.

xXW

xXlO«ca

Reiche

V. Wey.

obs.

Ost.

14.4

637.9

646

636

628

17.8

614.9

617

627

619

20.1

697.8

694.3

603

603

64.9

316.1

313.4

314.5

326.7

77.4

277.6

276.1

274.8

284.0

169.6

142.7

142.2

144.2

146.4

293.9

86.8

88.9

87.8

86.3

Table VI
Cbtbtalline manganese sulphate — MnS04+4HsO

Reiche ... J
V. Wey.. . . J
Oost. ... J

3.14 X 10-~;

1.1 X 10"~; M = 3.65 X 10"^'; ^o =

1.09 X lO"**;/! = 1.69 X 10"'

Xo »

7.294 XIO"*.
3.1000 X lO"*.

VK

xX10*cal.

xXWcal.

XX10«

xXWcaL

Reiche

V. Wey.

obs.

Ost.

HA

1233

1249

1233

1231

17.8

1014

1019

1021

1015

20.1

905

905.8

914

904

64.9

293.2

290.3

292

291

70.6

270.6

267.7

270

268

77.4

247.3

244.1

247

245

169.6

114.6

112.7

111.5

112.6

288.7

67.6

66.5

66.3

66.3

102

PARA' AND DIAMAGNETISM: WILLS

It must be remembered that none of the theories here mentioned
takes cognisance of the mutual action of the molecules, except that
of Gans, and in this respect is therefore deficient.

Tables V to VIII enable one to judge as to bow far the theories
are in accord with experiments. In this connection it should be noted
that in each of them there are two disposable constants.

Table VII
Watbr-hubs rBRRic bitlphatb — FetCSOOi

k-40

-21

Reiche ... J = 1.40 X 10"*; m « 3.42 X 10"^'; Xo = 286.4 X 10

k-«

TK

xXlO*caL

xX10*ob8.

64.0

70.5

77.6

169.6

989.8

177.6

167.6

156.7

85.1

53.3

177.1

167.3

157.2

85.6

53.3

Table VIII
Cbtbtaiximb ncRBo sulphate: FeS04+7HsO

Reiche ...J = 2.23 X 10 *;m « 2.94 X 10*";Xo = 3.365 X 10 *.

T*K

xXlO*cal.

xX10*ob8.

14.7
20.3
64.6
77.3
292.3

760.5
568.7
189.8
159.5
42.4

756
571
191
160
42.4

PARA' AND DIAMAONETISM: WILLS 103

vni

DIAMAONETISM IN METALS

DUE TO

MOTIONS OF FREE ELECTRONS

In accordance with views on the nature of electric conduction in
metals brought forward by Lorentz, Drude and others, there are
present in metals large nmnbers of free electrons which move about
among the atoms in a manner similar to that of the molecules of a
gas; and, moreover, the thermal properties of metals also lend support
to the assumption that free electrons are present in them in large
niunber. Although there are outstanding difficulties in the attempt to
ascribe to free electrons many observed electric and thermal properties
of metals there is yet strong evidence in favor of this assumption.

If the free electrons are present and moving about in metals like
the molecules of a gas, it is evident that in the presence of a magnetic
field the free paths of the electrons wiU be curved, and with a curvature
in such sens^ as to furnish diamagnetic quality to the metal. Super-
imposed upon the diamagnetism due to the motion of the free electrons
there will be, of course, the dia-, and perhaps the paramagnetism, of
Langevin.

Erwin Schrodinger^ in 1912 and H. A. Wilson* in 1920 have given
theories of the diamagnetism in metals due to the motions of free
electrons, arriving at quite similar conclusions by very different methods.
For the purposes of the present review it wiU suffice to outline the
argument presented by Schrodinger.

Theory of Schrodinger.

Stmcturdl Assumptions — ^The fundamental assumptions made as re-
gards the structure of a metal are precisely those made by Lorentz in
his theory of the motions of electrons in metals.

Two distinct species of particles are supposed to be present in the
metal: —

(a) Electrons with mass m and charge e moving freely among the atoms.

(b) The atoms of the metal, some of which carry a charge, while others
do not.

The electrons and the atoms are supposed to share in the thermal
motion of the metal, the particles of each type in the case of thermal
equilibrium having a mean kinetic energy equal to the mean kinetic
energy of the motion of translation of a molecule of a gas at the same
temperature as that of the metal.

Due to the small mass of an electron as compared with that of an
atom the velocities of the atoms are assumed negligible in comparison
with those of the electrons.

> E. SchrAdinger, Wien. Ber, 66, p. 1305; 1912.

• H. A. Wilson: Roy, Soc. Proc, Land, 97, p. 321; 1020.

104

PARA- AND DIAMA0NBTI8M: WILLS

The mutual action among the particles, io far as the electrons are
concerned, is supposed to occur through collisions only and as if the
colliding particles were perfectly smooth elastic spheres.

Owing to their small size the collisions of the electrons among them-
selves are ignored, and collisions only of electrons with atoms are
considered. Accordingly the mean free paths of the electrons are not
determined by their own number and size but by the number and size
of the atoms.

The Diamagnetism of Free EHectrons.

When such a mediimi is subjected to the action of a magnetic field
the free paths of the free electrons between collisions are no longer
straight, but curved, due to the action of the field. The motion of
the electrons along these curved free paths must act to produce dia-
magnetism in the mediimi.

It is now required to calculate the magnetic
moment resulting; from the curvature of the
free paths under the action of an external
magnetic field.

Referring to Fig. 11, dr is a small element of
volume of the medium; (, 17, f the coordinates
of an electron at a point Q within dr with re-
spect to an origin 0, also within dr. P is a
point on the z-axis at a distance r from 0, large
in comparison with the dimensions of dr. The
external magnetic field H is supposed in the
direction of the z-axis.

The magnetic force, say h, at P (0. 0. p) due to the typical electron

at Q ({, 1?, f) moving with velocity v ({ 1?, f) is, from (11), Sect. I, with
sufficient approximation expressed by:

h - ^vx (r-B);
cr*

r is the position vector of P and s that of Q. The scalar z-component of
this force is expressed by:

h,- 4i (yf^ - ^)-

cr*

The expression in brackets is the z-component of twice the areal
velocity of the typical electron with respect to 0 and e/r* is constant
for all the electrons in the volimie element.

The mean value, hi, of hi, is to be found through summation of
this expression for h| over all the electrons in dr, followed by integration
over a sufficiently long time T, and division by T. The order of sum-

Fio. 11

PARA' AND DIAMAONETISM: WILLS 105

mation and integration is, of course, indifferent and, sinoe the time
integral of the areal velocity of the typical electron is equal to the area
swept out by its radius vector, the time integral required is equal to
the sum of the areas swept out in the time T on the x-y plane by the
projections on this plane of the radii vectorii to all the electrons in
the element dr. If F denote the sam of these areas, then:

- 2e F
cHT

Calculation of F — ^The problem is thus reduced to the calculation of
F. This requires a knowledge of the law of distribution of the veloci-
ties of the electrons.

Before the establishment of the external field Maxwell's law may
plausibly be assumed; but with the field present the question arises as
to whether this assiunption is still plausible. The following con-
siderations show this to be the case.

Following Boltzmann^ let us consider the case of a mixture of two
gases, and let :

i, ri, t he the coordinates of a molecule of the first gas;

■ ■

(, 17, f be the component velocities of a molecule of the first gas;

X, Y, Z be the component accelerations of a molecule of the first
gas, due to the actions of external forces supposed dependent only upon
the coordinates, (, 17, T;

m be the mass of a molecule of the first gas;

• ■

f (() Vf ti ii Vt D be the velocity distribution function for the first gas.

Boltzmann showed that in the case of equilibrium (— » 0) :

dt

(2) f = foe-'*'°<^ + '*+^>,

where fo and h are such functions of the coordinates (, 17, f that for all
values of (, 17, f :

• af • df • df df dt dt

(3) f 7- + 1? - + f I- + X -' + Y?-' + Z^ « 0.

dx dy dz af dri df

Now it is assumed that this result may be applied to the present
case, where the electrons play the role of the first gas and the atoms
of the metal that of the second.

The components of the force on an electron due to the external field
H, say X, Y, Z, are given by:

1 Boltsmann: Gas Theorie, I. pp. 08-134.

106 PARA' AND DIAMAGNETISM: WILLS

mc

(4) Y- ~ (>Hi-fH.),

mc

Z-— (fHt-'nH,).
mc

Now it is noted that a violation of the Boltanann aflsmnptions is here
met with, since X, Y, Z depend upon the velocities. Scrutiny of the
Boltzmann proof shows, however, that it is still valid if it be amply

assumed that X does not depend upon (, Y does not depend upon •

and Z does not depend upon f . The above equations show that the
X, Y, Z of the present problem are such as to satisfy these conditions.
If now the value of f, from (2), be inserted in (3) we find, with the
aid of (4), that the terms in X, Y, Z all vanish and hence that the
equation

dh • dh • dh. "df* • Bt^ ' d£k
ox ay dz ox oy oz

• • m

must be satisfied identically by (, n, f . But this requires that fo and
h shaU be independent of the coordinates and therefore constant. In
this case the distribution function given by (2) is Maxwell's; and the
conclusion is reached that the presence of a magnetic field does not
alter the distribution of the free electron velocities in a metal.

Proceeding with the calculation of F, let X^ be the mean free path
(Tait's) of an electron moving with velocity v. The probability that
an electron moving with the velocity v shall proceed without collision
over a path with a length between a and a + da will be^

1 •
(6) - e^da.

It is here assumed that \y is independent of the velocity v and that
for all electrons:

1

X=

nirP

where n is the number of atoms per imit voliune and 5 is the radius of
an atom.

The number of electrons per unit volume in dr with velocities between
V and v+dv may be taken to be Vydv. Then the number of collisions

t CV. JauiB— Kin. Th. of Gaw 3rd Ed., p. 256.

PARA- AND DIAMAGNETISM: WILLS 107

of such electrons per unit volume in time T will be

-.vdv.

The fraction of these collisions for which the velocities afterward
have directions included within the solid angle da and, by virtue of
(5), which are such that the colliding electrons after collision shall have
free paths of lengths between a and a+da will be

e ^ da.

4ir X

Therefore :

vT -■

(6) 4;^*^ ^u^dvdcoda

will be the number of collisions per unit voliune in time T of electrons
with velocities between v and v+dv and for which:

(1) the velocities after collision shall be directed within the solid

angle do).

(2) the free paths after collision shall have lengths between a and

a+da.
Such collisions are denoted as of class A.

The volume element dr is now supposed subdivided into prismatic
columns, dn, parallel to the z-axis with sectional dimensions small in
comparison with those of dr, and also small in comparison with all
ordinary free paths. The nimiber of collisions of class A within a prism
of volume dn, by virtue of (6) will be expressed by:

(7) vT - ?

-— - e ^ Vy dv do) da dn.
4irA*

The areas described on the xy-plane by projections of the radii
vectorii of the electrons concerned in these collisions, as they describe
their free paths following the collisions, will all be appreciably the same.

The process of the calculation of F now requires the finding of the
sum of these areas for electrons of all classes of which class A is t3rpical,
followed by integration over all prisms of which dn is t3rpical and,
finally, by integration over all velocities of which v is typical.

The area described on the xy-plane by the projection on this plane
of the radius vector of a typical electron of class A is found as follow&
Referring to Fig. 12, let A be the position with reference to the xy-plane
of the electron at the time of a collision, B its position at the time of
its next collision. The shaded area bounded by OA, OB and the arc

AB is that required, the arc AB representing the free path of the electron.

106

PARA' AND DIAMAGNETI8M: WILLS

From tbe equations of motaon of an eketton in a magnetie field, tbe

path AB 18 easily shown to be an are of a drcle idiose length, b, is pven
by:

(8) b-aan^.

and idioee radios, p, is given by:

/AX mcv sin $

if 9 be the angle between the positive
s-axiB (direction of H) and that of the
axis of the cone corresponding to the
soUd angle d«.

If ^ be the angle between the tan-
gent to the path at A and the radios vector OA, then:

Fio. 12

(10)

Area OAB-Area OAB+Area CBA-Area CBA.

In order to obtain F, the area given by (10) has now to be multiplied
by the number of collisions pven by (7), and the appropriate integra*
tions made. If we write dtf^sin B d$ d^, the following expression for
F is thus obtained:

+ -p^sm- + -/>bf d*,
2 p 2 '

where b and p are given by (8) and (9) as functions of v and $, and OA
depends only upon the position of the prism dll.

All the integrations called for with the exception of that with respect
to v are easily made and the expression for F reduces to:

mcTdr
3e

f"vi — ,r„, -v«[dv.

Vmc/

Now IV the number of electrons per unit volume with velodtieB
between v and v+dv, is given by Maxwell's law:

iv-av«e""'"^dv,
where a and h are constants to be determined in tenns ci the total

PARA' AND DIAMAGNETISM: WILLS 109

«

number n of electrons per unit volume, and the mean squares of their

velocities v*, so that:

4 '

^n(hm)^

V'

hm = -=;.
2v»

The expression last given for F may now be put in the form:

s
4 mc n T dr (hm)*

where

OB -

\mc/
J, « J e vMv - ^(hm)* .

o

It may be shown that with sufiScient approximation^ that :

where

(12) a

^2mev/v*

Inserting the expressions found for Ji and Jt in (11), the final expression
for F is obtained:

(13) F - - J — X»nTH (1 - 2a«)dr.

omc

The mean field strength, ht, at A, due to the motions of the electrons
in dr, is then found with the aid of (1) and (12) to be given by:

2 e* Hdr,

h,= -- — Vn— -(l-2a«).

3 mc* r*

The form of this expression shows that the element of volume dr is
magnetically equivalent to a doublet of magnetic moment

1 e*

3mc*

X«nH(l-2a«)dr.

I Cf. SchK^dinger, I.e., p. 1328 and p. 1315.

no

PAR\- AND DIAMA0NETI8M: WILLS

It Kf be the magnetic Busoeptibility per unit volume it foUowB then
that:

(M)

Discussion of Results.

The fonnula (14), owing to the presence of the tenn in cf, shows that
in general jc, depends upon the field strength, but it will appear presently
that at ordinary temperatures a* will be negligibly small in comparison
with unity so that with sufficient approximation we may take:

(16)

1 ^ X.

Now the values of e and e/mc are well known, and nX and X may be
estimated from electrical conductivity measurements and plausible
assumptions concerning the true atomic volume. It is thus possible
to calculate approximately the values of jk, for different metals. This
has been done for Bi, Pb, Cu and Ag, chosen with particular reference
to the wide range of electrical conductivity exhibited by this series of
metals. The results are given in Table VIII together with the suscep-
tibilities observed for these metals. The electrical conductivities a
are expressed in C. G. S. electromagnetic units. The values given all
refer to the temperature 18°C.

Table VIII

Metal

aXW

nXXlO-"

nX10-»

XX 10"

-«iXlO»caL

-KX10*ob8.

Bi

0.84

0.046

0.8

5.54

2.37

13.7

Pb

4.84

0.267

4.8

5.56

13.8

1.36

Cu

67.2

3.174

52.5

6.04

178.

.076

Ag

61.4

3.405

53.4

6.38

202.

2.10

Comparison of the values calculated for jk,, the diamagnetic suscepti-
bility due to the free electrons, with the experimental values, k, shows
great differences to exist. It appears then that other sources of mag-
netisation than that of the free electrons are contributory in an impor-
tant way to the true magnetic susceptibility. The latter is probably due
to the combined effect of:

(a) The diamagnetism of Langevin; due to the induction effect during
the establishment of the external field upon the bound circulating elec-
trons within the atom. This effect is independent of the temperature.

PARA' AND DIAMAGNETISM: WILLS 111

(b) The paramagnetism of Langevin; due to the directive action
of the external field upon the magnetically polarized atoms or mole-
cules. This effect varies inversely with the absolute temperature.

(c) The diamagnetism of Schrodinger; due to the curvature of the
paths of the free electrons under the action of the external field. This
effect depends upon the temperature in rather a complicated way.

In the case of good conductors it may happen that the order of the
effect (c) is the same as that of the effect (b) in strongly paramagnetic
bodies. The dependency of effect (c) upon temperature, which appears
through the factor nX' ,is, however, by no means so simple as that called
for by Curie's law for paramagnetism which makes Uie susceptibility
of paramagnetic bodies vary inversely as the absolute temperature.
Therefore in all cases where effects (b) and (c) are in opposition and of
the same order of magnitude any simple law of variation of « with
temperature is not to be expected. Hereby is explained the failure
of the experimental curves between susceptibility and temperature for
metals obtained by Honda, and Owens, to exhibit any simple law of
variation of susceptibility with temperature, and in particular why
the susceptibilities of metals are so at variance with Curie's law.

The connection between the magnetic susceptibility, jc^, and the
electrical conductivity a, is obtained through comparison of formula

(15) for Kf, with the following formula for the electrical conductivity
obtained by Lorents on the same constitutive assumptions as those
adopted by Schrddinger:

(16) a-2Jl-^Xn —

^ Sirmc* Vy*

By division of (16) by (15) we find:

(17) JL._2J«_i^.

Under the same conditions X will not vary greatly from metal to metal,
so that at the same temperature «, will vary, approximately, directly
with a.

The Dependence of the Susceptibility Kt upon the Field Strength.
The formula found above (14) :

is a closer approximation for «c, than (15). The approximation is close
in either case only when cf is small in comparison with unity. Now a
is directly proportional to the field strength, since, from (12) :

112 PARA' AND DIAMAQNBTISM: WILLS

V

3 XeH

and the quantity

eH

represents the radius of the free path of an electron moving with the
velocity v^ perpendicular to the field. If, then, a* is to be small in

comparison with unity, the mean free path X must be small in comparison
with this radius.

Calculation shows that with an external magnetic field of 5X10^
gauss, the largest practically obtainable, the order of magnitude of
2af at 18^C is 10"^. Therefore any effect due to the variation of the
external field could hardly be detected. But calculation also shows
that at very low temperatures a marked decrease of susceptibility with
increasing field strength should be detected.

Note. — ^Professor Langevin has recently informed the writer of an
interesting result found by N. Bohr in his dissertation. In accordance
with the argument advanced by him it appears that the free electrons
in a metal, subject to Maxwell's Law of distribution for a simple gas,
should, on the whole, contribute nothing to its diamagnetic quality,
owing to the behaviour of the electrons at the boundary whereby they
produce an equal and opposite effect to that of the electrons in the
interior. Unfortunately this information reached the writer too late
to allow of the incorporation in the report of an outline of Bohr's argu-
ment.

FBRR0MAGNETI8M— INTRINSIC FIELDS: TERRY 115

THEORIES OF FERROMAGNETISM— INTRINSIC FIELDS

Bt Earlb M. Tbbrt
AflBociate Professor of Physics, Uniyersity of Wisconsin

HISTORICAL STATEMENT

In the early attempts to ac count for the phenomena of f erromagne-
tism, two rival theories were offered, — one by Poisson and the other
by Weber. Both regarded magnetism as a molecular property, but
they differed essentially in this, that while Poisson assumed the mole-
cules possess magnetic properties only when the substance is magnetized^
Weber considered that they have constant magnetic moments, and that
gross magnetism depends upon alignment. The fact that ferromag-
netic bodies all show saturation was taken as evidence in favor of
Weber's theory, for it is difficult to see why on the Poisson theory
magnetism should not be increased without limit. Again, the effecto
of vibrations in augmenting susceptibility were readily accounted for^
because of the greater freedom thus given to the molecules to fall in
line with the magnetizing force. The experiment of Beetz^ in which
he found that iron deposited electroUtically in a magnetic field pos-
sesses strong magnetic properties, furnished further evidence in favor
of the Weber theory.

The fact that ferromagnetic bodies do not show saturation for very
weak fields and the phenomenon of hysteresis are evidences that there
must be some form of constraint acting upon the molecular magnets.
Weber' assumed a restoring force equivalent to that of a constant mag-
netic field acting upon each molecular magnet in the direction of its
axis in the unmagnetized state. This assimiption, however, offers no
explanation of residual magnetism or of the other phenomena of hys-
teresis. In attempting to correct this defect in the Weber theory. Max-
well suggested a further assumption based upon the analogy of magne-
tization to elastic fatigue. He supposed that after a molecule has been
deflected from its original position by a magnetizing force, it returns only
partly if the deflection exceeds a certain value. While explaining reten-
tivity and some of the other phenomena of hysteresis, this theory fails
to account for certain facts observed in repeated magnetization. It
was suggested by Wiedemann and others that the deflection of the
Weber magnets might be opposed by a frictional resistance which not
only opposes alignment, but also holds the molecules in their deflected
positions after magnetization. If, however, the molecules were held
by friction until the appUed force is large enough to start them, the-

iPogo. Ann, 140, 1860, p. 107.

s Pogg- Ann, 88, 1852, p. 167, cf. p. 9 of this report.

114 FBRROMAONBTISM'-INTBINSIC FIELDS: TBRRY

flUBoeptibility for very weak fields would be zero, whereas it has ini-
tially a small constant value.

THE THEORY OF EWING

In contrast to the arbitrary constraints mentioned above, Bwing
proposed the theory that the molecular magnets are entirely free to
turn about their centers, and that the only constraints acting are the
fields due to neighboring magnets. This idea he developed in great
detail and, in fact, laid the foundation for much of the work which has
since been carried out. From a mathematical consideration of the
simple case of a 2 magnet group acted upon by an external field, he
obtained a ciu^e in which the three stages of magnetization are clearly
indicated and by an experimental study of a model in which 130 snudl
pivoted magnets were used, he obtained magnetization and hysteresis
curves which approximated the observed curves for ferromagnetic
bodies with surprising accuracy. He gave a theoretical treatment of
the case of a ferromagnetic body made up of rhombic crystals with
molecular magnets placed at the comers of their space lattices, where
the crystals are placed with all possible orientations. By a statistical
method, which has been the basis for the subsequent work of Langevin,
Weiss, Honda, and others, he showed that the percentage retentivity
should be .8927, and deduced a number of other important results.

THE WEISS MOLECULAR FIELD HYPOTHESIS
Statement of Langevin's Theory.

It was pointed out in a preceding part of this report that Langevin^
by an application of the method of statistical dynamics, has arrived at
an expression for the intensity of magnetization of a paramagnetic gas
in terms of the electron theory. For this purpose he supposed that the
state of magnetization depends upon two factors only; first, the external
field which tends to produce alignment in a given direction, and second
the thennal agitation, which acts for disorganization. By an applica-
tion of the Maxwell-Boltzmann distribution law, in which the number
of magnetic molecular axes pointing in a given direction corresponds
to the density of a gas, and the angle with the external field to height,
he arrived at the foUowing expression for the intensity of magnetiza-
tion of a paramagnetic gas at a temperature T under the influence of
a field H :

(1) — = coth a — , where

' <^mo a

*^^ ^m H

■ Langevin, Ann, de Chem, et de Phya., Ser. S, 5, 1905, p. 70. cf. p. 56 of this report.

FSRR0MA0NETI8M— INTRINSIC FIELDS: TERRY

115

In these equations,

c^ = Magnetic moment per gram molecule;

(Tm = Magnetic moment per gram molecule at saturation;

H = External field ;

T = Absolute temperature ;

RT = Twice the kinetic energy for one degree of freedom of a

molecule;
R =Gas constant for a perfect gas referred to the molecular

mass (R=83. 15X10^ ergs per degree).

Langevin's equation, plotted as C in Fig. 1, gives the percentage
saturation for a paramagnetic gas at any temperature as a function of

a

Fig. 1

the apphed field. In weak fields, the intensity, crm, is proportional to
the field, but the slope becomes less with increasing field and finally
approaches assymptotically to the saturation value crmo. By a simple
calculation he showed, for the case of oxygen, that a field of 100,000
gauss would be necessary, at ordinary temperatures, to produce an
appreciable departure from the linear law.

Langevin showed also that the well known experimental law of
Curie, i. e. the inverse proportionaUty of the susceptibility to the
absolute temperature for paramagnetic substances follows directly from
bis formula. Developing the right hand member of equation (1) in
a series, there results:

(3)

^m a

— =:.-;::: aM

'm.

3 90 45.42

aH

116 PBRROMAONBTISM—INTRINSIC FIELDS: TBRBY

Taldog, as an approximation which holds over the range of fidds
experimentany realixable, the first term only in this devdopment,
we have:

^^ ^m^ 3 SRT'

Letting X«, » - » the molecular susceptibility, there results:

C^ is called the "molecular" constant of Curie, i. e. the proportion-
ality factor when the susceptibility is referred to the gram molecule.
Curie's law, as expressed by equation (5) holds for a large number cS
paramagnetic substances over wide ranges of temperature. Assuming
it to hold at absolute sero, ^^m^, the saturation value of the intensity
may be determined for a substance by measuring its susceptibility at
A known temperature T. This is the hypothesis which has been made
by Weiss in his theory of the ''Magneton'' to be discussed later.

The Molecular Field.

By postulating a "Molecular Field," Weiss^ has extended the ideas
of Langevin to the phenomena of ferromagnetism. In this he was
guided by the method which Van der Waals used to develop a kinetic
theory of liquids by extending the ideas which Bernoulli had applied
to a perfect gas. Just as in the case of a gas, to account for the transition
to the liquid state, there must be added to the external pressure an
internal one due to the mutual attractions between the molecules, so
in the case of a ferromagnetic substance, as it is cooled in a magnetic
field from a temperature which has rendered it paramagnetic, the
transition to the ferromagnetic state is explained by assuming that, due
to the overlapping of the fields of the individual molecules, there comes
into existence an internal or molecular field, which added to the external
field, accounts for the very large intensity characteristic of this state.

Weiss assumes that the overlapping of the fields of the molecules
existing in a given region is equivalent to a imif orm field proportional
to the intensity of magnetization and directed parallel to it. Thus:

H„-NI,

! de Phya., 4th Series, Vol. 6, 1907, p. 061. Arch, des Scieneea Phyt. et Nat.,
1.31. 1911, p. 401.

I

FERROMAONETISM—INTRINSIC FIELDS: TERRY 117

where Hm is the molecular field, I the intensity of magnetization and
N, a constant characteristic of the substance. The molecules con-
tributing to this internal field are contained in a definite sphere of
action. He assumes, moreover, that the forces due to the magnetic
fields are the only ones which act upon the molecules of a ferromagnetic
substance and, except for them, the molecules are as free to rotate as
in the case of a perfect gas.

Spontaneous Magnetization.

Weiss further supposes that it is not necessary for an external field
to be acting in order that the individual parts of a body may be mag-
netized. On the contrary, he assumes that throughout the body the
molecular field alone maintains the intensity of magnetization of the
elementary units of volume at a magnitude very near the saturation
value for the particular temperature at which the body exists in the
same way that a fluid, by virtue of the internal attractive forces, main-
tains its liquid state in the absence of an external pressure. The
volumes throughout which this spontaneous magnetization exists in an
uninterrupted manner are very small, limited perhaps to the individual
crystals. In a finite body with resultant intensity zero, the directions
of magnetization of the individual elements are distributed entirely at
random, and the fimction of the external field, in giving a resultant
intensity to the body, is to produce an alignment of the individual
group intensities, but not to change their magnitudes. In other words,
if one could examine with sufficient minuteness, he would find an im-
magnetized body to possess the same intensity as one grossly magnetized
in the most powerful fields available.

The magnitude of the spontaneous intensity of magnetization may
be obtained in the following manner. Equation (1) gives the value of
the intensity of magnetization at any temperature T in terms of the
saturation value by means of the auxiliary variable a. It is then
merely necessary to replace H in Equation (2) by H„ = NI and sub-
stitute in Equation (1). This may be effected most easily by means
of a graphical elimination of a between the two equations. In equation
(6), I is defined as the magnetic moment per unit volume, while a„ is
the magnetic moment per gram molecule. It is therefore necessary to

replace I by its value ^ , where D is the density of the substance.

m

Accordingly:

^""o ND
(7) a'^ ^ p rpXcr^, and

m K 1

^ ^ «^«. <^«*„ N D ^ *

118 FBRROMAGNBTISM— INTRINSIC FIELDS: TERRY

The last equation g^ives the straight line of Fig. 1, which intersects
the former curve in two points. It is easy to show that the intersection
at 0 corresponds to a state of unstable equilibrium and that the one
at A is the one concerned. Since the parameter a contains T, the
spontaneous magnetization as a function of the absolute temperature
may be readily deduced.

The Magnitude of the Molecular Field.

Anticipating for the moment what is to be shown presently, it may
be stated that the molecular field is very large compared to fields
available in the laboratory. However, in the temperature interval
between the ferro- and paramagnetic states, there is a small region in
which the molecular field is of the same order as realizable fields, and
by measurements made in this transition region the constant N of
equation (6) may be determined; and from it the value of H„, the
molecular field may be computed.

For this region equation (2) may be written:

,^, <^m, (He+NI) <^m, (H.+ — O

(9) a^ rY— = ^-7^—'

RT

where He is the external field. At the transformation temperature 0,

(10) L«_ _ »' and a - ^=^^^^^
^ ' "m, 3 IWm

Eliminating a from these two equations,

»»« N D.

(11) 9 =

3Rm

(12)

Combining (11) and (9) and reducing, there results:

T - ^ He m .

^ (T^ND

<^m,

Letting x^ ^ ^^ where x., is the molecular susceptibility, there

He

results:

An

(13) (T - «) X«

ND

FERROMAGNETISM— INTRINSIC FIELDS: TERRY 119

This is a modified form of Curie's Law and states that the suscepti-
bility is inversely proportional to the excess of the temperature above
the transformation point. This law has been found to hold for this
region with very good accuracy and from it the value of N has been
deduced.

The following values have thus been obtained:

Substance N Hn

Iron 3,850 6,6d0,000

Nickel 12.700 6.360.000

Magnetite 33.200 14.300.000

Cobalt 6,180 8.870,000

Experimental Evidence Regarding the Ebdstence of the Molecular Field.

1. The law of Corresponding States and the Variation of {he Saturatum
Intensity with Temperature. — ^As noted above, an unmagnetized body
consists of minute crystals all magnetized to the saturation value for
that temperature but having their magnetic axes distributed at randonicic
The process of magnetization consists in lining them up, and if we could,
apply an external field sufficient to produce gross saturation, we shou^e^
be able to measure the molecular intensity, since it would then be the J
same as the gross saturation. Further, a study of the variation of tl^.
saturation intensity with temperature should furnish a direct test ojf
the concept of the molecular field as given in equations (1 ) and (6)
This test may be facilitated by a general equation applicable to all
substances analogous to that for corresponding states in the kinetic
theory of gases. Such an equation may be obtained in the following
manner.

The slope of the straight line of Fig. 1, is proportional to the tempera-
ture T. Accordingly, by giving successive values to T and determining
the intersections with the curve C, the law of variation of intensity
with temperature may be derived. The limiting case is that in which
the straight line coincides with the tangent to the curve at the origin,
and corresponds to the temperature 6 at which spontaneous ferro-
magnetism disappears. This transformation temperature may be ex-
pressed in terms of the constants of the medium by noting that at $,
equation (1) may be written with sufficient accuracy by using only the
first term of the development of equation (3); that is:

(14) ^ = ?•

Also equation (8) becomes:

120

FBRROMAGNBTISM'-INTRINSIC FIELDS: TERRY

(16)

(15) by (14) there reeuks:

e

3mR '

(16) by (8), and mmplifymg, one obtains:

(17)

1* « ? ??L

This equation, together with equation (1) g;ives the complete law of
the thermal variation of spontaneous ferromagnetism, and when ex-

pressed in terms of the variables — and -^, is the same for all substances

6 ^mo

(T,

dm.

The full line of Fig. 2, taken from the original paper of Weiss shows
the calculated curve, and the crosses, the values obtained for magnetite.
The work was carried out in a field of 8300 gauss although previous
experiments had shown that this material is practically saturated in a
field of 550 gauss. The agreement is satisfactory except in the low

FBRBOMAONBTISM'-INTRINSIC FIELDS: TERRY 121

temperature region where marked departm^ occur. For pyrrhotite
and the alloy Fes Ni the agreement is more satisfactory than for
magnetite; but for iron, nickel and cobalt, the agreement is less satis-
factory in that larger systematic departures are found.

2. The Dependence of Specific Heat upon Oie Molecular Fidd. — If
ferromagnetic substances are the seats of molecular fields of the mag-
nitudes stated above, a considerable amount of energy must be supplied
as the temperature is raised from absolute zero to the transformation
point in order to break up the alignment of the molecular magnets
within the crystals. We should expect, then, that in this region, the
specific heat would be greater than it would be, if by some means the
substance could be deprived of its magnetic properties. This effect
should show itself as an additive term to the true specific heat of a
corresponding fictitious substance having no magnetic properties. The
amount of this additional heat may be computed from the theory of
the molecular field in the following manner.

The mutual potential energy E of a group of magnets of moment m is

(18) E = - S M H cos a,

where H is the resultant field due to the group at the point where an
individual magnet is located, and a is the angle between H and this
magnet. When the summation is extended to all the magnets con-
tained in a centimeter cube, there results:

(19) E = ^IH„=-^NP,

where I is the magnetic moment per unit volume, H^y the molecular
field, and N the constant of equation (6). The negative sign indicates
that it is necessary to supply heat to demagnetize the substance. The
intensity decreases in a continuous manner from absolute zero to the
temperature at which the disappearance of ferromagnetism occurs.
Accordingly, the amount of additional heat that must be supplied in
raising a ferromagnetic body from a temperature at which the inten-
sity is I, to the Curie point 6, is

(20) Q-^^P-i5=I
^^^ '^ 2JD* 2JD'

where J is the mechanical equivalent of heat and D, the density. The
mean specific heat accordingly, is:

dq IN dp 1 do*

^ ^ "^ dT 2J D dT 2J dT

122 PERROMAGNETISM--INTRINSIC FIELDS: TERRY

where a is the magnetic moment per imit mass. This quantity is small
at low temperatures, but increases as the temperature is raised and
disappears abruptly at 6. At this point, it has the nature, not of a
latent heat of allotropic transformation, but of a discontinuity in the
true specific heat.

The magnitude of this discontinuity has been calculated by H.^A.
Lorentz.^ Developing in a series the theoretical law of the variation
of magnetization at saturation as a function of the temperature, he
found at 6:

do* 5 (T *

where <r^ is the saturation value of c.
Taking into account the relations:

(23) ^=CND, and

(24) a, ,

3R^C

m

where C is the Curie constant referred to unit mass, R the gas constant
for a single molecule, and m the molecular mass, he obtained:

On substitution of the nimierical values for R and J there results:

(26) aC„ - — .

m

Weiss^ and his co-workers have tested this theory in a series of ex-
periments extending over a period of several years. In the early work,
equation (21) was used as the form in which to make the test and the
results seemed to check the theory within the limits of accuracy of the
experiment. In the later work, however, where greater care was
taken, the check is less satisfactory. The results in which the Lorents
equation (26) was used are summed up in the following table:

> H. A. Lorents, Reoue ScierUifique, 1912, 50 aiin6e« p. 1.

* Weiaa and Beok, Joum, de Phyt., 4th Series, 7, 1908, p. 249. A. Dumas, Zurich
Thesis, 1909. Weiss. Piccard and Carrasd, Arth. des Sei. Phyt. et Nat,, 42, 1916, p. 379,
%l80 43, 1917, p. 113, and 43, 1917, p. 199.

FERROMAGNETISM— INTRINSIC FIELDS: TERRY

123

Table II

Substance

Nickel

Magnetite (Artificial)
Magnetite (Natural) .

Iron (Pure)

Iron (Swedish)

ACm observed

0285

0790
0736

120
124

ACm computed

.0282
.0644

.089

Corresponding
mag. molecule

Ni,
H(Fe,04)

Fe

3. Magnetic Properties of CrysUda and the Hysteresis Curve. Weiss^
and his group have examined a number of iron minerals and found that
some of them possess marked magneto-Ksrystalline properties. One of
the best examples is Pyrrhotite, a sulphide of iron. These crystals
are usually in the form of hexagonal plates bounded at their edges by
faces of a hexagonal prism and are deeply striated parallel to the base.
If one examines their magnetic properties in planes parallel to the base,
he finds that there is one direction in which they are very easily magne-
tized, while at right angles to this direction it is difficult to produce
saturation. Further, in the direction normal to the base, saturation
is still more difficult. Weiss found that the fields necessary for satura-
tion in these three directions are 15, 7300, and 150,000 gauss respec-
tively. After an extended examination, he concluded that the complex
crystalline structure consists of a juxtaposition of elementary crystals
of which the magnetic planes are parallel, that each crystal possesses
a direction of easy and difficult magnetization at right angles to each
other, and that the crystals are grouped in the magnetic plane with
their axes making angles of 60^ with each other.

The direction of easy magnetization is further characterized by the
fact that the intensity of magnetization can be changed in sense but
not in magnitude. For example, if one acts upon a crystal in this
direction with a large field, and then gradually reduces it, carrying
it through zero to negative values, he finds that the intensity remains
constant down to a value of — 15 gauss when it suddenly reverses and
takes a negative value of equal magnitude. In other words the h3rste-
resis curve is a rectangle with lines parallel to the H axis extending
out from the upper right and lower left hand comers. The magnetic
properties of Hematite have been studied by Kunz' who found it to
be similar to pyrrhotite in that it possesses directions of easy and
difficult magnetization, though the effect is less marked, and that the
coercive field is somewhat larger. It is ferromagnetic in some direc-
tions and paramagnetic in others.

> Weifls, Joum. de Phya., 3rd Series, 8, 1899, p. 642.
* Kuns, Areft. det Sci„ 23, 1907.

124

FERROMAGNBTISM^INTRINSIC FIELDS: TERRY

Weiss has attempted to explain h3rsteresis phenomena in pure metals
by assuming that their individual crystals possess properties similar to
pyrrhotite; that is, directions of easy and difficult magnetization, and
that each crystal is magnetized by its own intrinsic molecular field to
the saturation value for its existing temperature. In gross matter, in
the immagnetized state, the directions of easy magnetization will be
arranged entirely at random. The process of magnetization in a given
direction consists then simply in reversing the direction of magnetiza-
tion of those elementary crystals whose intensities have components
opposite to the external field. For a given cr3n3tal, this reversal occurs
when the component of the external field in the direction of its axis
equals He, the coercive field.

The form of the hysteresis ciu've to be expected on the basis of this

Fio. 8

assumption may be obtained in the following way. Let M be the mag-
netic moment of each crystal, and N the nimiber of crystals per unit
volume. Since the distribution of directions is entirely at random, the
end points of the vectors M will be uniformly distributed over the sur-
face of a imit sphere. Let the external field H act in the direction OX
of Fig. 3, and let H^ be the magnitude of the coercive field. Elemen-

FERROMAQNETISM— INTRINSIC FIELDS: TERRY 125

taiy magnets having axes l3ring within the cone of semi angle ^ vertioal
to the one indicated in the figure will be swung into this cone. The
angle ^ is determined by the expression H cos^—Hg.

The number of vectors ending in the zone determined by d ^ is given
by:

/«^x 2irr*sin*d*^^ N .

(27) 4irr' ^="^«^ * ^ *•

The magnetic moment of these magnets in the direction OX is:

N

(28) M,=Mcos0 — sin0d0.

The moment due to aU the magnets reversed into the cone is

f*MN .

(29) M,= / — — sin 0 cos 0 d 0

Jo 2

MN Im . ,^

= — — sm*0=--sm*0.
4 4

The total magnetic moment due to the magnets in the cone is then:

(30) ^''-;^''^-k['-{fJ\

In this discussion it has been assumed that the elementary crystals
can be magnetized, only in the direction of easy magnetization, while
if they resemble pyrrhotite, they are paramagnetic at right angles to
this direction. Weiss has computed the appropriate correction and
has matched a set of h3rsteresi8 curves taken from the results of Ewing
as shown in Fig. 4.

The Elementary Magnets of the Ferromagnetic Substances.

In his original study of a paramagnetic gas, Langevin expressed the
intensity of magnetization as the magnetic moment per unit volume
instead of per gram molecule as Weiss has done in his later work. For
this quantity he used the letter I and his equation was:

(31) r- * cosh a — , where

(32) a«^;
^ ^ RT'

126

FBRROMAONBTISM-'INTRINSIC FIELDS: TERRY

Feq.4

FEKROMAQNETISM'-INTRINSIC FIELDS: TERRY 127

fi is the magnetic moment for a single molecule and the other quan-
tities have the same meaning as before. In the neighborhood of the
transformation temperature 0, these equations become:

la mH

(33) r"~3» *^^ * ~RT' ''^spectively.

Putting H = N I and eliminating a there results:

3R

(34) M =

NC

an expression by means of which ^ may be determined for those sub-
stances for which the quantities 6, N, and Im have been determined.
This calculation has been carried out by Kunz.^ R is the gas constant
for a single molecule and may be obtained from the equation:

(36) p=NiRT,

where Ni is the nimiber of molecules per cc. Substituting the values
for the quantities involved, at one atmosphere and Oo C,

p = 1.01 X 10* dynes per cm^
T =273
andNi=:2.7 XW\
there results: R =1.36Xl(ri«.

Taking, for iron, Im = 1950, the value obtained by extrapolating the
results of Curie, for N = 3860, the value given by Weiss and Beck,*
and substituting in equation (34) there results:

(36) M = 4.445 X l(r*® absolute electromagnetic units.

As a check on the reasonableness of this result, a calculation of the
mass of the hydrogen atom was carried out using it and the known
density and molecular weight for hydrogen. Let No be the number of
molecular magnets per cc. in iron at absolute zero. Then:

(37) NoM=1960,

whence No = — = ^^ = 4.386 X 10".

° u 4.445 Xl(r*» "^^^^

» Kuns, Phy; Rev., 30, No. 3, March, 1910.

* Weias and Beck, Joum. de Phy:, 7, 1008, p. 249.

128

FBRBOMAGNBTISM— INTRINSIC FIELDS: TBRBY

If it 18 assumed that each molecule possesses one elementary magnet
of moment m> then this is also the number of molecules per cc. If m
is the mass of one molecule of iron, and D its density, then:

Nom = D=7.36,

(38)

whence m

7.36

4.386X10"

1.792X10-** grams.

If Mh is the mass of the hydrogen molecule and it is assumed that
the molecule of iron has two atoms, then:

(39)

Ml

1.792X10-"
111.8

1.603X10-** grams.

A recent value of this quantity deduced by Rutherford from radio-
active phenomena is:

Mh = 1.61 X 10^* grams.

If the corresponding calculations are carried out for nickel and cobalt
using the best available data the results given in the following table
are obtained.

Substance

Fc

Ni,
Co

Im

1950

570

1435

766

376

1075

N

3,850

12.700

6,180

NI-HM

6,540,000
6,350,000
8.870,000

mX10»*

4.445

3.65

6.21

M,xia-«*

1.603
1.603
1.61

N

2

6

4

It is to be noted that, in order that the computed mass of the hydrogen
atom should have the values given, it is necessary to assume that the
molecule of nickel has six atoms and that of cobalt four.

The Nature of the Molecular Field.

The hypothesis of the molecular field as introduced by Weiss is a
useful concept in the theory of ferromagnetism and has served a num-
ber of useful purposes. For example, by adding to the external field
the molecular field it is possible to explain many of the complicated
phenomena of ferromagnetism by the laws of paramagnetism. It
gives a theoretical law for the variation of the saturation value of the
intensity with temperature through the ferromagnetic range, and leads
to a law for the intensity variation with temperature above the magnetic
transformation point. By assuming for the molecular field different

FERBOMAGNETISM—INTRINSIC FIELDS: TERRY 129

values in different directions it is possible to account for many of the
complicated phenomena of crystals, and by taking into account the
energy associated with the molecular field an explanation for the dis-
continuity in the specific heat at the transformation point is obtained.
The phenomena of the molecular field, moreover, are' not confined to
ferromagnetic substances, as there are many instances of its evidence
in the case of paramagnetic and diamagnetic substances as well. One
may cite, for example, the work of Kammerliegh Onnes and Perrier^ on
the magnetic properties of mixtures of liquid oxygen and nitrogen; that
of KammerUegh Onnes and Oosterhuis' on paramagnetic substances at
low temperatures; Weiss and Foex on paramagnetism of crystalline
substances, Foex on concentrated sahne substances, and Oxley on dia-
magnetic substances, to be discussed later. It is true that in many
instances, the check is only qualitative and indicates that the theory
in its simple form is insufficient, and that the molecular field, instead of
being proportional to the intensity of magnetization should be repre-
sented by a more comphcated function such as :

(40) H„ = NiI+N3P+

While the hypothesis has thus been useful in explaining many observed
facts and directing new lines of investigation one is at once struck by
its enormous magnitude and is led to inquire by what means fields of
such intensities may be produced. For this purpose one might proceed
in the manner employed by H. A. Lorentz for dielectrics and describe a
sphere within which there exists a single molecule while on the outside
all the other molecules, in their mean effect, play the role of a homo-
geneous substance. He would then find for the coefficient N of equa-

4
tion (6) the value ~ r which falls far short of that experimentally deter-

mined on the basis of the theory.

Again, using known data, one might compute on the basis of the
inverse square law the necessary distance from a molecular magnet at
which the observed molecular field would occur and see whether it
leads to values consistent with the known densities of packing of mole-
cules. Take, for example, the molecule of iron which contains eleven
magnetons, and suppose that it has a length equal to .2X10~^ cms. the
diameter of the atom, and let m be the strength of its magnetic pole!
Then since the magnetic moment of the magneton is 16.4X10"^, there
results:

mX.2XlO-' = llXl6.4XlO-«,
or m=.9XlO-".

> Kam. Onnes and Perrier, ArM. de Chemie, 4th Series, 26, Sept., 1913.
* Kam. Onnes and Oosterhuis, Comm. Leiden No. 129, p. 132.

130 FERBOMAGNBTISM—INTRINSIC FIELDS: TBRRY

The distance from such a pole at which there would exist a field of
strength equal to that of the molecular field, 7X10*, is given by:

m .9X10-"

---^^ 7X10»,

r* r*

whence r« 3.6X10"" ;

a value much less than the measured distances between molecules. It
thus appears that fields of the required magnitude can be obtained
neither by superposition of the effects of neighboring molecular magnets
using the known average values of the intensity of magnetization, nor
by assuming sufiScient closeness of packing of the individual molecules.
On the other hand one might enquire whether it is possible to obtain
such fields by allowing electrons to rotate with very high velocities in
closed orbits about the positive nucleus. For example, suppose an
electron having a charge of 1.6 XlO"*^ to rotate in a circular orbit of
diameter .2X10"^ with a frequency of 10** equal to that of ultra-violet
light. The magnetic moment of such a circuit would be:

Moment = 1.6X 10-««X 10"XirX lO""

= 5X10-«,

which is equivalent roughly to three magnetons. The strength of the
field at the center of the trajectory is:

„ 2-ir-610-» 10"- ,^

^ TIF^ ''''

which is too small by a factor of 100.

It thus appears that the molecular field can have neither a magnetic
nor an electromagnetic origin and must therefore be of a nature differ-
ent from the ordinary magnetic fields with which we are familiar.
Weiss* has suggested that the molecular field may be of the same nature
as the ''magnetizing action of contact" observed by Maurain' and others
in their study of the magnetic properties of electrolytic iron deposited
in a magnetic field. This work will be reviewed briefly. Maurain
showed that iron deposited in a field of a few gauss is much more strongly
magnetized than that deposited without the field and afterwards sub-
jected to one; also that when the field in which the deposition occurs ex-

> AnnaUa de Phy:, 1, 1914, p. 148.

* Maurain, Joum, de Phyt,, 4th Series, 1, 1002, pp. 00 and 151.

FERROMAGNBTISM--INTRINSIC FIELDS: TERRY 131

ceeds ten or twelve gauss the iron is saturated. This iron maintains its
saturation value, practically independent of the field, but suddenly
reverses under a coercive force of 20 gauss and the hysteresis curve is
practically a rectangle similar to that of P3n:rhotite in its direction of
easy magnetization. The saturation values were rather low, however,
being only about 840.

Eaufmann and Meyer^ who repeated the work of Maurain, have con-
firmed his results regarding the shape of the hysteresis curve and the
value of intensity for weak fields, but by using stronger fields they
obtained intensities as large as 1100. Schield^ has also studied iron
thus deposited and found an intensity of 080. All of these intensities
are considerably less than those for ordinary iron, i. e., 1700, and one is
led to suspect that their peculiarities may be due to the presence of a
hydride of iron. This seems all the more probable from the fact that
many of these peculiarities disappear with time but may be partially
restored by making the specimen the cathode of an electrolytic cell.
Nevertheless the results obtained have an important bearing on the
molecular field theory.

Maurain also found that the first layers of the deposit are different
from the later ones in that they are but weakly magnetic. In fact it
was only after the deposit had reached a thickness of 80mm that its
magnetic moment increased in proportion to its thickness. It thus is
evident that there are two fields acting on the molecules at the instant
of deposition; first the external field and second, that due to the polarity
of the iron already deposited. This latter he called the ''magnetic
field of contact." He tried opposing these two by reversing the external
field after a suitable thickness of deposit had been obtained. It was
found that as long as the external field did not exceed the coercive field,
usually about 20 gauss, the magnetic moment increased in the direction
of the original field for some time in proportion to the thickness and that
it was only after the thickness of the new deposit had become com-
parable to that of the original one that the magnetic moment became
zero and finally reversed. The reversal of the polarity of the original
deposit took place slowly and could be observed with the magnetometer.

He next studied the dependence of the field of contact upon distance
by depositing upon the magnetized cathode suitable layers of neutral
metals such as gold, silver, and copper of varying thickness and again
depositing iron. With the external field reversed, he found that with
a thickness of 38mm of the neutral metal the new layers of iron behaved
in the same way as those deposited on an unmagnetized cathode. In
other words at this distance the contact field just neutralizes the exter-

iPhy8. Zeitachr,, 22, 1911, p. 513.

> Shield, Ann, d. Phyt., 4th Series. 25, 1908, p. 612.

132 PERROMAGNETISM—INTRINSIC FIELDS: TERRY

nal field. On the other hand, for very thin layers of neutral metal the
contact field is very large compared to the external field. The char-
acter of this 'Afield of contact" is as yet unexplained, but it seems
probable that it is of the same nature as the ''molecular field," and in
view of the work of Oxley on diamagnetic substances is worthy of fur-
ther study.

Theory of Frivold,

As was pointed out above, Weiss concluded that the large molecular
fields required by this theory of ferromagnetism could not be of purely
magnetic origin but must arise from other magnetic forces. In order
to determine to what extent the fields of the individual atomic magnets
can ac<;ount for the molecular field of Weiss, an extended calculation
has bei3n carried out by Frivold.^ For this purpose he assimies that
the equilibrium of the elementary magnets is determined not only by
the external field and the thermal agitation, as in the Langevin theory,
but alfo by the overlapping of fields of the elementary magnets and
treats the problem from the standpoint of statistical mechanics. The
calculation is carried out for 2 cases: the unidimensional and the vol-
ume distribution.

1. The Unidimensional Problem: Elementary magnets of number N
are considered to form a long chain and to be free to turn about their
midpoints. They are in statistical equilibrium under the influence of
their undirected temperature motions, the external fields, and their own
mutual field. Let the origin of co-ordinates be located at the middle
of the chain, and let $ and 0 be the usual polar co-ordinates, and let
the axis of co-ordinates and the external field coincide with the direction
of the chain. If the magnetic moment of an individual magnet is fi,
and if their instantaneous positions are given by ^i^i, dt^^ ....
^N^N> statistical mechanics gives, for the mean magnetic moment of the
chain at a temperature T in the direction of the chain, the following
expression :

-*/• • ■ h

V

(1) Mt=A / . . . / Ai2cos ^^e " dOidOt . . . d^N.

Here U designates the potential energy of the chain, k the Boltzmann
constant (k== 1.35X10'^* ergs), and dn the solid angle formed by the
element of surface sin Odd d^ on a unit sphere. The integration is
to be taken through the 2N variables, $i ^i, ^s ^s, . . . . ^n ^-
The constant A is determined from the following consideration : The
probability of a given condition characterized by the fact that the di-

> Frivold. Ann. der Phynk, 65. p. 1. 1021.

FERROMAGNETISM^INTRINSIC FIELDS: TERRY 133

rection of the axes of the elenentary magnets lie within the solid angles
dQi, dQs> . . • . dils is

Ae ""dQidQi . . . . dON-

Integration of this expression, when the co-ordinates 0 and ^ run
throughout the values 1 to N, gives for the probabiUty the value imity.
Therefore:

kt
(2) A I ... I e dQidQs . . . dn^^l,

/•••/'

and

. . . / /* 2 cos ^N e ^T d Qj d Qj . , . d 12n

W iviT= J '' \ _^

/ . . . / e kT d 12i d Q, . . . d Qn

In order to carry out the integrations in equation (3), it is necessary
first to determine the potential energy U. This consists of two parts,
that due to the mutual potential energy of the elementary magnets, and
that due to their positions in the external magnetic field. Calling
these Ui and Uj, respectively, we have:

U = Ui+U,

(4) • =^,^ 2'[(m„ m„+x)-3(m,„?)(m„+i, ?) j-2 (m,,, ff).

The expression in brackets is an approximation found to be accurate
within 6 per cent.

The mean magnetic moment of an elementary magnet, which con-
sists of such a chain may then be evaluated. Introducing the following
abbreviations:

2 fk=0, where fk = f m^, mk+ij-sf m^, ^Vm^+i, ^^

(6)

Sgk=^, where gk=f— , gj;

1 ^ mH

-,;ikT = P, and ^ = q;

,j,_dQ, do, dQN.

^^~ 4ir ■ 4ir • • • • 4ir '

184

PBRSOMAQNBTISM—INTRINSIC FIELDS: TERRY

equation (3) becomes:

(6)

Mt-

77—7

e«*+^dS

Letting now the integral in the denominator be designated by J,
there results:

(7)

Mt d
M dq

J may be expanded in a power series in p^, that is, in pt wers of

M» ,

a»kT

and integrated term by term. Thus:

(8)

J- 1 .... ye^(l+p*+^p**«+ . . .)dS.

The approximation, given by the series development, is closer the
higher the temperature T. To evaluate A it is necessary to determine
the following mean values :

W

D^J* .... Je^dS;
Di-J .... Je^^p^dS;

When these integrations have been carried out, neglecting the quad-
ratic and higher powers in (8), there results for the mean magnetic
moment of an elementary magnet in terms of its absolute value, the
following expression:

= (cothq-^)r]

(10) ^=lcothq-- jl 1+4' , ^ ^
^ Nm \ q/L a»kT dq

(cothq — )+ .

PBRSOMAQNETISM—INTRINSIC FIELDS: TERRY

135

If the reciprocal action of one magnet on another is not taken into
account, the second term in the square bracket of (10) is zero, and there
results the well known Langevin expression. The extent of the devia-
tions from the Langevin expression, brought about by the introduction
of the mutual actions, may be seen by substitution of numerical values
in (10). If we assume that the chain consists of iron atoms, which,
according to Weiss, possess 11 mangetons, each having a moment equal
to 16X10-" C. G. S. units, and for "a" assume the value 2XlO-» cm.,
then at a temperature of 300° absolute, since k«1.35XlO-^*i there
results:

a'kT

3.7X10-^ .

In figure (5), curve (1) represents the original Langevin function,
while curve (2) is

-p-lcothq ).

dq\ q/

This last expression for the values assumed above has a maximum of
^ at a field strength of 10* gauss. The mean magnetic moment, when

M

•

9 ^^^

....^r——

1

^

■ —

HxlO

*6

Fig. 5

the mutual actions are taken into account is represented by curve (3).
Since the mf^-TiTniim value of the departure from the Langevin curve is
of the order of 10"*, the effect is here greatly exaggerated, and it must
be concluded that at this temperature the effect of the mutual actions
is quite negligible, and a magnetic body consisting of a chain of magnets
with the values given above shows only paramagnetic properties.

It is to be noted that in the above development, the integration of
the equation (8) was carried out for the first two terms. If the quadratic

136

FBRROMAGNBTISM— INTRINSIC FIELDS: TERRY

term Ls included, the calculation is much more complicated, but the
result shows that for external fields of 50,000 gauss, the magnetic mo-
ment of the chain is a linear function of H.

At low temperatures, on the other hand, the conclusions are quite
different. For small values of T, a number of simplifications in equa-
tion (8) may be introduced and the equation corresponding to (10) is
found to be:

(11)

Ml

Nm d q

log J

-^[

1+2 -, +

•]■

where n — — --;

m*2

+

H

a'kT 2kT

The results for four values of T are shown in figure (6). Smoe the
approximations do not hold for extremely weak fields, the curves are

Tg.oi

Fig. 6

shown dotted in this r^on. The chain shows, therefore, at low tem-
peratures, properties characteristic of ferromagnetic substances, but it
goes over into the paramagnetic state for temperatures of a few degrees
absolute.

PBRROMAONBTISM— INTRINSIC FIELDS: TERRY 137

2. The Three Dimensional Problem. The calculation for the case of
the space lattice is carried out in a manner similar to that of the uni-
dimensional problem. The elementary magnets are r^arded as lo-
cated at the comers of a cubical space lattice and turn about their mid-
points. The expression for J in equation (8) is evaluated as before but
is complicated by the fact that double summations must be made. If,
as a first approximation, the expression corresponding to D and Di of
equation (0) are evaluated, there results the well-known Langevin
equation:

M ^ 1

(12) rr~=icothq

N/i q

If, however, the quadratic term of equation (8) is included, an expres-
sion, in which the mutual actions appear, is obtained. Two cases have
been studied — the ordinary cubical space lattice and the centered cubic.
For the former, Frivold obtains:

+

VkT/l 4 \a»kT/ 16^ • • • -J'

and for the latter:

<»' i^.-^^?[{'-?tfx)"--}

+

/mHV 3^/_i^Y^lU 1

VkT/ 4 Va'kT/ 15/"^ . . . • J-

A comparison of these equations shows that the numerical factors in
the two cases are of the same order and, consequently, the difference in
the arrangement of the atoms plays no important role. Accordingly,
in the following discussion only the former case will be considered. If
the mutual action is left out of account, equation (13) gives for the
case, in which the external field is relatively small, the well-known Lange-
vin equation for small fields:

M,_l/iH

138 FBRROMAQNBTISM—JNTRINSJC FIELDS: TBRRY

and for the initial permeability:

M, 1 M* N

(16)

H 3kT

If, however, the mutual action is taken into account, equation (16)
becomes:

(17) M,_l^ r 6^/^« 1

^^^^ H 3kT^L 4\a>kT/^- • • • J i

which indicates that the effect of the mutual action is to reduce
permeability.

A study of equation (13) shows that for external fields of such magni-
tude that q' comes into consideration, a temperature transformaticm
point is evidenced. For example, the magnetization curve (13) lies
above, coincides with, or is below the curve of equation (17), according
as the sign of the coeflScient

VkT/

is positive, zero, or negative, or, in other words, are equal as T is greater
than, equal to, or less than

Brk V9M>

a'k I' 2.66X16

The magnitude of these departures is, however, very small, as may
be seen by the substitution of the generally accepted numerical values
for iron. For example, putting /i = 1.76X10-*® C. G. 8. units, a « 2.86
X10-« cms., and T=390** absolute,

\a? k T/

^10-^

From this it must be concluded that at ordinary temperatures, the
influence of the mutual actions upon the magnetization curve for the
case of the hypothetical magnetic substance we have here considered,
is negligible. The reason for this is the small value for the moment
of the elementary magnet. It is of interest to consider the case in
which the atoms contain, in addition to the elementary magnets, dec-

PBRROMAGNBTJSM^INTRINSIC FIELDS: TERRY 139

trie dipoleSi the electrie moments of which are of the order of those
found for SOs e. g. 10~^*; that is, 100 times larger than the magnetic
moments for iron atoms. The external magnetic field then starts the
lining up process, whereby the internal electric and magnetic fields
are brought into play. Frivold has carried through the calculation in
this case, also, and obtained the following expression:

ri8A MTlft.Hr/ 6.27/ M.' Y, \

^'^jifUiry-hh--]

where the subscripts m and e refer to magnetic and electric moments,
respectively. Much larger departiu'es from the case in which mutual
actions are neglected are thus obtained, for here

a»kT

Unfortunately, the integral of equation (8), upon which the entire
treatment rests, is developed in powers of p ^, i. e. of

a»kT

and the series is convergent only when this expression is less than unity.
The question as to whether ferromagnetism can thus be explained by
the assumption of electrical dipoles, is still left open, but interesting
possibilities are here suggested.

Theory of Gans.

A theory of ferromagnetism has been developed by Gans in which
he has attempted to take into account the effect of molecular structure
upon magnetic properties, and to make more precise the ideas concern-
ing the nature of the molecular field than was done by Weiss. He sup-
poses that an elementary magnet or "Magneton" is an electrified body
of revolution rotating rapidly about its axis of figure. An elementary
complex consists of a group of such magnetons, distributed according
to the laws of probability throughout a space which has the form of an
ellipsoid, the three axes of which are unequal. The magnetons are free
to move about within the complex in the samemanner as the molecules
of a gas. A ferromagnetic crystal is built up of such complexes placed

140 PBRROMAONBTISM— INTRINSIC FIELDS: TERRY

at the intersections of a space lattice with their corresponding axes
parallel.

By applying the laws of statistical mechanics to a system of such
complexes, relations are obtained between magnetic and thermal quan-
tities, similar in form to those of Weiss, but which are somewhat more
comprehensive. From stability considerations, he is able to deduce
the hysteresis curve to obtain a relation between coercivity and
temperature, and to determine the number of magnetons per unit
volume and the magnetic moment per magneton.

The chief difference between the Weiss theory and that of Gans,
briefly stated, is as follows: In the former, the exciting field, acting at
a definite point within a ferromagnetic body, is composed of two parts,
the external field H, and the molecular field N M, where M is the
intensity of magnetization and N is a constant characteristic of the
substance. In the Gans theory, the exciting field consists of three
parts, the external field H, the "structure field," due to the gross mag-
netization of the body, which, by a treatment similar to that of Lorents
for dielectrics, is found to be

-tM.

and a molecular field A due solely to the action of the magnetons of the
particular complex in which the magneton under consideration is lo-
cated. It is assumed that each direction for the molecular field A is
equally probable, and that its magnitude is independent of direction.
The molecular field, on account of the different directions which it
assumes, has a tendency for disorganization and acts, therefore, in the
same sense as the thermal agitation. At high temperatures, the action
of the molecular field may be neglected in comparison to the thermal
agitation, while at low temperatures, thermal agitation may be neglected
in comparison to the molecular field.

The equation for the magnetization curve for a ferromagnetic sub-
stance may be deduced in the following manner:^ The magnetic moment
of the magnetons of a particular group has, from symmetry considera-
tions, the direction of the resultant field F, which is obtained by vec-
torial addition of E and A, where E is the sum of the external and struc-
ture fields, and is equal to

H+^tM

and A is the molecular field. This is shown in Fig. 3, section III of the
report on kinetic theories of dia- and paramagnetism by Dr. Wills.

> cf . p. 46 of this report.

FERROMAONETISM— INTRINSIC FIELDS: TERRY 141

When the component of this magnetic moment is taken in the direc-
tion of the external field and the smnmation extended to all the groups
included within a unit volume of the substance, its intensity of magneti-
zation is obtained. It is shown by equation (32) of the above reference,
that the average value of cos d, where d is the angle between the axis
of a magneton and the external field, is:

1 /•• r(A+K)

(1) cosd=;^J WJA)_dA / (cotha--)(F«+K«-A«)dF.

Jo A J ±(A-K) *

where W (A) dA denotes the probability that the molecular field A
lies between the limits A and A+d A, and

kT

In carrying out this integration, the positive sign of the lower limit
should be used for K<A and the negative for K>A.

Since it is assumed that the molecular field is constant in magnitude
and that all directions are equally probable, 'equation (1) reduces to:

rCA+K)

(2) ^^d = 7^ / (coth a - -) (F

4 K« y ^^^_K) a

i+K*-A*)dF.

The magnetic moment per unit volume will, therefore, be given by:

M,

/(A+K)
(coth a—-]
±(A-K) *

(3) M = NMC0S^=7-r^ / (cotha— )(F«+K*-A«) d F,

where Mo equals N n. Since

K=H+|irM,

this relation, together with equation (3), gives M as a function of H,
and the magnetization curve may accordingly be deduced. It may be
shown that for A = 0, equation (3) reduces to that of Langevin.

To apply this formula to a ferromagnetic body and to see how the
phenomenon of hysteresis is concluded, let us think of a ferromagnetic
crystal having a rhombic space lattice such as pyrrhotite, with eUip-
soidal elementary complexes situated at the intersections with corre-
sponding axes parallel. Let the magnetization and field strength at
points within the complexes be designated by M^ and H\ respectively.

142 FBRROMAGNETISM-'INTRINSIC FIELDS: TERRY

and let M and H refer to the corresponding quantities at points within
the crystal but outside the complexes. We may then write :

H/-H.+N/M„

where N/, N't, Ni' are constants depending upon the structure.
Further: •

M = nVM'

where n is the number of elementary complex per unit volume and V is
the voliune of a single complex.
The quantity K of equation (3) is defined by:

K=H'+~tM';

but may be expressed in terms of H and M by the following relations:

K,=H,+NiM.,

K,=Hy+N,M„

K.=H.+N,M.,
where

N.=N.'+3^; N.=N,'+^; N.-W+g^.

K may be regarded as the directive part of the total force acting on
the magnetons of a complex and is made up of the external field H within
the crystal and another field having components Ni M^, Nt M,, Ni M
which depends essentially upon the form and arrangement of the
elementary complexes and which may appropriately be named the
" Structure " field.

The magnetization curve, i.e. the M, H curve for a crystal in the
direction of one of the axes of symmetry, e.g. the X axis, may be ob-
tained from the M K curve by a shearing process as follows:

Let the dotted curve of Fig. 7 be that given by equation 3, and let
S S' be the shearing line inclined to the O M axis by an angle such that
tan a — Ni.

. FERROMAGNETISM^INTRINSIC FIELDS: TERRY

143

If P is a point on the M K cnrvey then P' is the corresponding point
on the MH curve, where PP'=QR. The shearing angle a depends
upon the structure constant N. Two cases are to be considered, i.e.
a <P, and a >P, where p is the angle between the tangent to the M K

Fig. 7

curve at the origin and the O M axis. In the first case the M H curve
lies entirely in the first quadrant; but in the second it follows the path
O C D of Fig. 8 ^. From stability considerations it may be shown that

M

ik

Fio. 8

for a </3 the magnetons of the elementary complexes are in stable
equilibriimi throughout the entire range of field strengths and that the
substance is paramagnetic. - On the other hand, when a>P, between the
fields designated by the vertical tangents at the points C C\ the equili-

144

PBRROMAGNETISM—INTRINSIC FIELDS: TERRY

brium is labile. The substance is then ferroniagnetic and exhibits the
properties of hysteresis, as indicated by the curve.

By developing the integrand of equation (3) in a power series and
making certain approximations to simplify the integrations, Gans has
deduced a number of important relations between magnetic properties
and temperature. For example, he has deduced equations connecting
retentivity and temperature, coercive force and temperature, and
obtained a relation between susceptibility and temperature for a fer-
romagnetic substance above the Curie point. The first of these is
substantially the same as obtained by Weiss and is in good agreement
with the observations of Weiss' and Foex for magnetite but not for iron
or nickle. The second and third relations are in good agreement with
the results of Terry' for iron, nickel, and cobalt at high temperatures.

The Theory of Honda and OkAbo.

In contrast to the theory of Weiss, in which molecular fields of the
order of several million gauss are assumed to be acting, an attempt has

Fig. 0

been made by Honda and OkAbo,' following the ideas of Ewing, to
deduce the curves of magnetization and hysteresis and to explain the

1 Weifls and Fote, Arch, dea Sei. Phyt. el Nat., 31, p. 4. 1911.
'Terry, Phyt. Ret. 33, No. 2, 1910 and 60, N. S. No. 6. p. 394. 1917.
* Science ReporU Tohoku Univ. s, No. 3. p. 153, 1916.

PERROMAGNETISM^INTRINSIC FIELDS: TERRY 145

properties of crystals by taking account of the mutual actions of mag-
netic molecules whose poles act according to the law of inverse squares.
For this purpose they have considered a Ewing model of nine coplanar
magnets placed at the comers of a square space lattice as shown in Fig. 9.
Although the real problem is three dimensional, a study of the two
dimensional case is sufficient to indicate the degree of success to be
expected from such a theory. If no external field acts, the elementary
magnets take positions of stable equilibriiun parallel to one of the sides of
the space elattice. Under the action of a field, however, the group turn
as a whole toward the direction of the field and takes an equilibrium
position determined by it and the mutual actions of the group.

To make the problem definite, let the origin of coordinates be at the
center of the magnet P R and let an external field H act in a direction a
with respect to the Y axis, and suppose the elementary magnets to be
turned through an angle 6 in consequence. Let 2a, 2r and m be the

sides of the space lattice, the length of an elementary magnet, and the

a
pole strength respectively, and put - = k.

The position of equilibrium of one of the magnets such as P R may
be determined by equating the torque due to the external field to that of
the 16 remaining poles of the group. The restoring torque is obviously
a function of 40 since the magnets of a square space lattice are in equilib-
rium when they stand end to end; the equilibrium being stable when
they are parallel to one of the axes of coordinates, and unstable when
parallel to the diagonals. The analysis shows that the equilibrium con-
dition may be written :

(1) Hsin(a-0) = A8in40,

where A = -^(K) is a quantity depending upon the strength of the

elementary magnets and their particular arrangement within the group.
The intensity of magnetization I in the direction of the applied field is:

(2) I = 2mrnco8 (a— 0) = I^cos {a— 6),

where n is the number of elementary magnets and Iq the saturation value
of the intensity of magnetization.

I H

(3) Using r = i> and -=h,

lo A

as " reduced " values of the intensity and field respectively, we have
the relations:

(4) i= cos (a— 0),

146 PERROMAONBTISM— INTRINSIC FIELDS: TERRY

(5) and h sin (a—d) ^ mx4$,

as the equations defining the magnetization of a simple complex. If
h and a are given, equation (5) gives the value of tf, and this, when sub-
stituted in equation (4), gives the value of i. Equation (5) is, however,
of the eighth degree in sin tf or cos 6 and must therefore be solved by an
indirect process. It is necessary, first to point out the way in which an
elementary complex behaves when acted upon by external fields of
various magnitudes in different directions. As indicated by equation
(1) the restoring torque on each magnet due to the mutual actions of the

group is a function of period -. It is a maximum for angles of - with

the sides, and reverses sign at angles of -. Let us suppose that the

4

magnets are originally parallel to Y and that a field h acts at an angle a

and rotates them through $. Four cases present themselves.

Com 1. O < a < T. The component magnetization in the direction of h

4

starts with the value i » cos$, increases continuously with h and becomes

unity for h» oo.

IT W

Case 2. z<^<^' 1^® magnetization increases continuously with h

until the deflecting torque exceeds the restoring torque, when the magnets
jump to a new position of equilibrium between h and X. This new posi-
tion is the same as though the magnets had remained in their original

IT

positions and a were changed to a—-. There results a discontinuous
increase in i. For angles a in this octant, the jump occurs for values of

IT

B in excess of - . With further increase in h, i increases continuously

to unity as h approaches infinity.

Case 5. ;;<«<— -. The magnetization increases continuously with h
2 4

until the restoring torque is exceeded by the deflecting torque when a

discontinuity occurs, and the magnetization follows the same course as

IT

though a were replaced by a—-. This case is similar to case 2, except

IT

that in the new equilibriiun position t is greater than - .

3t
Case 4' — <a<T. The magnetization up to the discontinuity is the
4

same as in the above cases. The discontinuity, however, may be of

PERROMAGNETISM^INTRINSIC FIELDS: TERRY 147

two types. For directions of h somewhat greater than -7 the torque ia

4

It 5t

greater for the Tnaximum near - than for the one near -- ; the magnets

5t
jump to a position somewhat less than — and the subsequent magnetiza^

o

tion takes place as though a were replaced by a-* - as in case 3. If, on

the other hand, a lies in the neighborhood of t, the torque is greater

5t it

near -- than r and the magnets jump to a position between h and the

o o

negative Y axis, and the subsequent magnetization takes place as though
a were replaced by a— IT.

The field h„ at which the discontinuity occurs, may be obtained in
the following way. Since

sin4
sm {a— 6)

the value of 6 for which h is a maximum is given by:

dh 5 sin (a+3^)+3 sin (a-5 6)

(7)

de 2sin*(a-d)

CaUing this d^, there results:

(8) 5 sin (a-3 0+3 sin (a-6 0=0.

The field h„ is obtained by solving this equation for ^o ^^d substituting^
in (6).

We are now in a position to study the magnetization of a ferromag-
netic mass consisting of a large number of elementary complexes with
their space lattices distributed uniformly in all directions in a plane.
Let N be the number of complexes, and d N the number whose axes
make with a certain direction an angle between a and a+d a when no
external field is acting. Then:

(9) dN=^da.

Let M be the moment of a complex, a the angle between its initial
direction and that of the external field h, and let it be turned through.

148 PERROMAGNBTISM— INTRINSIC FIELDS: TERRY

an angle 6 by the action of this field. In the direction of the field its
component is M cos (a— 0). The magnetization due to all the complexes
is:

(10) 1 = 2 /^cos(a-d)da=- / cos (a-^) da,

where Io»MN is the saturation value of the magnetization. The
reduced magnetization i is given by:

1 /■•
<11) i==" / cos (a— 6) d a.

This, together with equation (80) furnishes the solution to the problem
of finding the equation for the magnetization curve. That is, for a
given value of h, 6 may be found from (6) in terms of a, and this value
when substituted in (11) gives i. Owing to the discontinuities in 6
discussed above, it is, however, necessary to consider the problem for
large and small values of h separately.

When h is small, d is also small, and we may put sin 4 0~4 0. Equa-
tion (80) then becomes:

(12) h (sin a—e cos a) =4 6;

whence

h sin a

4+h cos a

Substituting in equation (11) there results:

1 /■'
i = - / (cos a

i = - / (cos a+d sin a) da

^ V 1 f'/ h sin* a v ,

(13) =-/ (cos «+-———) d a

rJo 4+n cos a

1 /"' h C"" h

= - / cos a d a+— / sin* a (1+ - cos a) "* d a.
tJq AtcJq 4

The first integral vanishes and the second, when expanded in a power
deries and integrated, gives

FERROMAGNETISM— INTRINSIC FIELDS: TERRY 149

(14)

= .125 h+.00196 h«+.00007 h*+

The intensity i is here expressed as an odd function of h and is nearly
linear in the neighborhood of the origin with an upward concavity which
increases with h. It approximates well the experimentally determined
curves.

The solution for large values of h is complicated by the abrupt changes
in the value of d. Further, the angle at which these discontinuities
occur depends upon both the external field and the orientation of the
complex. It is therefore necessary, in evaluating (11) to divide the
integration interval into several parts. For a given h, the critical
angles may be determined from equations (8) and (6), a study of which
shows that, for reduced fields slightly in excess of unity, there will be
three such angles, giving four integration intervals. Calling these
angles ai, as, and at we have :

"« frn-ftftf.

«i "^ «« '^ as

For the complexes l3ning within the intervals of the first and fourth
integrals, the magnets remain stable since the torque due to the external
field does not exceed the restoring torque. For the complexes of the

second integral, the magnets make jumps of - as explained in cases 2 and

3 above, and the integration limits must be changed from ai and at,

IT X

to ai— - and aj— ~, respectively. For the third integral, the magnets

lie beyond the first and second positions of stable equilibrium, and jump
by an angle t. The limits accordingly must be changed to as— x and
at— IT, respectively.

The integrand of equation (11) contains 6 and the evaluation can be
effected more easily in terms of this variable than of a. The elimination
of a may be made as follows: Differentiating (6) with respect to 6,
there results:

■-At-^)

h cos (a— ^) I T-— 1 j"=4 cos 40;

da 4 cos 4 0

whence -r = , ; ^, + 1 .

d^ h cos (a— 0)

FBRBOMAONETISM— INTRINSIC FIELDS; TBRRY

. 1 fjicat*! 1 ;- — r-T—\,

The new iategratton limits coTresponding to m, a*, and ai for given
values of b may be obtained by aubetituting theae values successively
in equation (6) and solving for S. When this has been done, there
results:

From equation (90):

(18) i-l{»m«J±i/;Vl-lrin.«4
The int^ral in this equation may be written:

(19) ; j y 1 - k» sin* « dfl- y Vl - k» 8in» « dS,

where k*<=r,. These are elliptic integrals of the second kind 'with
modulus k, and may be written:

^|ECk,4^-E(k,4ff)l.

Expanding E as a power series in k and determining the appropriate
limits of 0 in equation (17) from equations (8) and (6), Honda and Okubo
have computed ihe intensities corresponding to four different values of h.
The results are given in Table I and plotted in Figure 10 which is seen to
possess, in a marked degree, the characteristics of the experimentally
determined magnetization curve for a ferromagnetic substance.

Table I

PBRROMAGNETISM^INTRINSIC FIELDS: TERRY 151

The residual magnetism to be expected on the basis of this theory
may be obtained as follows : When h has been made infinite the mag-

nets of all the complexes having orientations between d= 7 and ± -7- take

4 4

new positions of equilibrium corresponding to discontinuous rotations
of - with respect to their initial positions, while those lying between

d= -r and t jump by t. When the field is reduced to zero, all the
4

magnets then behave in the same manner as those l3ning between zero

and 7 which return reversibly to their original positions. The residual
4

magnetism R is then given by:

w

(20) R-2f M COB tfdN, where dN-^dtf

T

IT J o

coeSde^

-4Io.

x>

W2

and the reduced residual magnetism r is:

R
r=--.8927.

The portion of the hysteresis curve l3ning between the retentivity
point and maximum induction may be deduced by considering that the
magnetization process in this interval takes place reversibly and that

all the complexes have initial orientations lying between ±7 with respect

to the direction of the field. The law of magnetization is then given by
the equations :

4 r* wxAB

i= ~ / cos (a—d) d a, and h = ": — z r^-

Tj o sm (a-e)

(21) For h smaU we have:

1 r* h .

i = - / (h+4 cos a) (1+- cos a)"* d
^y o 4

= .8927+.047h-.083h«.

152

FERROMAONBTISM— INTRINSIC FIELDS: TERRY

For larger value of h, equation (19) must be used where the proper
limits of intergation are obtained from equations (8) and (6). The
portion of the hysteresis curve for negative values of h is obtained by
assuming that the case is equivalent to the magnetization by a positive

•

I
1/1

-

^

ID

J

^^

^

4

a

3&

A

T

■

[_

— ^

tf

(

J^

y

i,

r

4

3

2

i-

«

1

i

3

4

S

-y

t

— 4

— 0

t

#

^

/

__

W

f

'

u

Fig. 10

field of a group of complexes whose initial magnetic directions are

3 T 5 T
uniformly distributed between the angles — and — .

The results of calculations are shown in Table II and plotted in
Figure 10.

FERROMAGNETISM^INTRINSIC FIELDS: TERRY

153

Table II

h

•

1

h

•

1

+0.0

1.000

-1.0

0.815

3.5

0.973

-1.5

0.015

3.0

0.962

-2.0

-0.584

2.5

0.956

-2.5

-0.786

2.0

0.944

-3.0

-0.847

1.5

0.932

-5.0

-0.981

1.0

0.922

— «

-1.000

0.0

0.893

The similarity between these curves and the curves of experiment is
striking. The most important departm^ is probably the large value of
the retentivity. For the curves here deduced the remanence is 89 per
cent, while in practice one seldom finds a value greater than 60 per cent.
This discrepancy is probably due to the fact that in this theory no ac-
count is taken of thermal agitation. Hysteresis phenomena are assumed
to take place only when the molecular magnets turn abruptly through

angles of x or t, otherwise the processes are reversible. The energy

losses due to hysteresis must be accounted for by the kinetic energy
acquired by the magnets during these jumps which is then dissipated by
friction, radiation or some other process.

Honda and Okiibo have extended their study to the case of magnetic
crystals. For this purpose, the only change it is necessary to make for
those of the rectangular system, such as Magnetitie and Hematite is
that all the elementary groups are oriented in the same direction instead
of at random as in the case discussed above. For Pyrrhotite, a hexa-

gonal space lattice must be used for which F (0) has a period of '. By

this means they have deduced the results of Weiss, Quittner and Kunz
on these crystals with the same degree of accuracy as was obtained in
the case of ordinary ferromagnetic substances.

The Mean Molecular Field of Diamagnetic Substances.

In the resume of the Weiss theory it was pointed out that many of
the phenomena of ferromagnetism may be explained in terms of the
laws of paramagnetism by the introduction of an internal or molecular
field due to the presence of surrounding molecules. Langevin has indi-
cated that the origin of the magnetic properties of both para- and
diamagnetic substances is to be found in the rotation without damping
of electrons in closed orbits about the positive nuclei. If the arrange-
ment of the orbits possesses complete symmetry, the resultant magnetic
moment and hence the field at distances large compared to molecular

154 PERROMAGNBTISM^INTRINSIC FIELDS: TERRY

magnitudes ia lero, and the substance is diamagnetic. If, on the other
hand, there is a lack of symmetry in the orbital arrangement, the field
at a distance is not zero, and the substance is paramagnetic. The
pondermotive action of repulsion exhibited by diamagnetic substances
when introduced into a magnetic field is accounted for by assuming
changes in the electronic orbits in accordance with the ordinary laws of
induced currents in a manner analogous to the explanation of the
Zeeman effect given by Lorentz.

In his theory of diamagnetism, Langevin has considered the effect of
the external field only and has not taken into account the action of
neighboring molecules when the substance is polarized. The fact that the
Zeeman effect and the rotation of the plane of polarization, both closely
related to diamagnetism are, in the case of ferromagnetic substances,
proportional to the intensity of magnetization and not to the applied
external field would indicate that in diamagnetism also, the suscepti-
bility should be a function of the state of polarization. Inasmuch as
the forces of diamagnetic repulsion are small and the susceptibility is
in general independent of the temperature, the existence of an internal
or molecular field would be difiScult to prove. Nevertheless with a
change in aggregation, such ais accompanies the transition from the
liquid to the crystalline state, one should expect, if such a field exists, a
measureable change in susceptibility, due to the distortion of the
electronic orbits caused by the effects of the magnetic fields resulting
from the new state of polarization.

Oxley^ has investigated a large number of diamagnetic substances
and has found that with few exceptions there is a decrease in diamag-
netic susceptibility of about 6 per cent, when the substance passes from
the liquid to the crystalline state. On the theoretical side he has ex
tended the method of Langevin by the introduction of an internal field
depending upon the polarization to accoimt for this discontinuity at
the transition. This extension to the theory is as follows :

Instead of assuming, as Langevin did, that the force acting on any
electron of a rotating group, is simply e E, where E is the electric field
strength and e the charge on the electron, he assumes, with Lorentz,
that it is given by

e (E-hf (P))

where P is the electric polarization of the mediiun and f (P) a function
which characterizes the grouping of the molecules for a given substance.
The crystalline state may be regarded as isotropic to a first approxima-
tion since the crystab will have all possible orientations. The effect

I Oxley, PhU. TranB. Roy. Soe., 214, p. 100. 1914; 215, p. 79. 1914; 220, p. 247. 102a
Proe. Roy- Soe, A,, 95, p. 68, 1918.

PERROMAGNETISM-'INTRINSIC FIELDS: TERRY 165

due to the modification of the internal motions of an atom or molecule
by the process of crystallation will be taken into account by a change in
the value of f (P). Following the theory of Langevin, let (a, b, c,) be
the coordinates of the center of gravity of a molecule and (x, y, z,) those
of a particular electron. Also let (f , 17, f ) = (x-a, y-b, z-c) be the coor-
dinates of an electron with respect to the center of gravity of the mole-
cule in which it is situated. Since the medium is homogeneous and
isotropic,

(1) 2f=2;i7=2f=2fi7=2i7f=2ff=0.

The sectorial velocity of an electron with reference to the center of
gravity of the molecule will have a component along o z given by:

(2) Q.«-(fA-nf),

and the component of the magnetic moment of the molecule along this
axis is then:

(3) M.=2eQ..
From (2) and (3) there results:

(4) M.=|2(fA-nf).

Let X, Y, Z, be the components of the internal forces determined by
the configuration of the molecules which act upon the electron of mass
m and let E and H be the total electric and magnetic fields respectively.
If the origin moves with a velocity having components u, v, w, then
the equations of motion are:

m f =X+e [E,+f (P,)]+e H. (v+y)-e Hy (w+z)-m i-m li;
(5)

m i) = Y+e [Ey+f (Py)l+e H, (w+i)-e H. (u+x)-m b-m v.

These equations differ from those given by Langevin only in the
addition of the term f (P). Because of the smaU dimensions of the
elementary system considered, the electric force and the polarization
will be nearly constant throughout its extent, and, designating their

z

+

156 FERROMAQNETI8M— INTRINSIC FIELDS: TERRY

values at the center of the system by Eo and f (Po) respectively, we may,
by expanding and neglecting powers higher than the first, write:

Calciilating M from the above equations, there results:

«-£[*{(t)r(f)i-4tHt)}

-Kt).^t).-«-^]

where ^ =2 p=2 *i7=2 f*. The last term of (8) is zero provided each

molecule has no initial moment as Langevin's theory requires. Dropn
ping the subscript 0 and using the electromagnetic field equations:

(9) curl E=H-, and div H=0,

their results:

Intergrating from the time 0 (H>=0) to r (H«H.) their results:

(11) AM.- -|^H.A+^/;[±f (P,)- Af (P.)]dt.

where A M. is the magnetic moment produced in the molecule by the
change in field which occurred during the interval. The second term
depends upon the molecular configuration of the substance and implies a
modification of the electron circuits which will change their self induc-
tance. Any such change of self inductance may be represented by a
small change in the intensity of the applied magnetic field, and we may
then write:

(12) f(P)-aP,

FBKROMAGNETISM^INTRINSIC FIELDS: TERRY 167

where "a" characterizes the grouping of the molecules. Accordingly:

m if(P,,-|f(P.,-a(fi-a)._.|(fflO,

where a (dH,) is the elementary change in the external field during a

small interval of time r.

Therefore:

(14) AM.- -£H.A-^/;i(aH.) d.

-^['+^]-

The term q A H, is the total variation of H. caused by the distortion of
the electron orbits. If N is the number of molecules per gram, the
specific susceptibiUty may be written:

.... NAM, Ne^Ar,. AH,1

(15) x = -H^ = — 4^Ll+a^J.

An expression is thus obtained in which the susceptibiUty is shown to
depend by means of the quantity "a" upon the state of polarization of
the substance, and the term a AH, is the molecular field produced
thereby. Ifa=0, (15) reduces to the expression originally obtained by
Langevin. Calling ai and ae the polarization constants for the liquid
and crystalline states respectively, the variation of x on crystallization
may be written:

(16) «2=(a.-aO^'.

It has been shown by Larmor^ that ai is of the order ~ for most liquids*

The value of ac is large but its exact determination in any particular
case is difficult since it depends upon the actual distribution of the
molecules about which we know relatively Uttle. It is possible, how-
ever, to obtain an approximate value of its magnitude from the work of
Cbaudier' on the change of magnetic rotatory power with change of
state. He has shown that a^ must be at least of the order 10* and ia

A TT

probably larger. * is accordingly of the magnitude of 5x10"^.

1

■PAtl. Ttom. Roy, 8oc„ 1897, A, p. 213.
CampUa, Rend., 156, p. 1008, 1913.

158 FBRBOUAONETlSM-lNTRIffSlC FIELDS: TERRY

A comporiBDD o( the molecular field for dianugDetic nibstanoea with
that of ferromagQetic mibetancce according to the Weiae theory mar ^
made as followg: For a aupercooled hquid, we may write:

(17)

while i<H cryrtala at the tame temperature x. is given by equation (15).
Hence:

.-.(...^-|L-)

(19) whence H. x,-x, (H.+a, AH,).

The term a, A H, ia the mean molecular field of the diamagnetic crystals.
Since in equation (6) of the Weiss theory, the molecular field coDstant,
which we will here designate as N*, ia taken as the proportionalty factor
beween molecular field and intensity, while in equation (19), a, dH, is
itself the molecular field, it is necessary to compare N* with a,' irtioe
the latter is defined by equation:

(20) a, AH,- a,' N AM. p

in which p, the density of the aubetance, is approximately unity fw
the crystals investigated by Oxiey. By using the first relation <rf

equation (15) and putting -~— '•-5x10'* there results:

Assuming x~5 * ^0~^> *od a,- 10*, a,' is found to be ctf the ordo* 10*,
which is of the same order as the values of N* given by Weiss and Beck.

The Local Molecular Field.

In the above discussion of the mean molecular field, it was pcnnted out

that the change of susceptibihty which accompanies the transition from

the linuid to the crvHtallme state can be satisfactorily interpreted in

Id appreciable only in the crystalline state,

kgnetic&lly by a term a« A H,. The nature

1 further than to say that it ia of such a

hin the crystal a distortion or polariiation

PERROMAGNETISM—INTRINSIC FIELDS: TERRY 169

equivalent to that actually produced by the molecular forces of the
molecules of the crystalline structure.

On the theory of magnetism developed by Langevin a diamagnetic
molecule contains oppositely spinning systems of electrons which
counter balance each other at distances large compared to molecular
dimensions, but which nevertheless produce fields close to the molecules
which may be very large. Each molecule of a crsrstal is accordingly
subjected to the intense magnetic fields of its neighbors and the resulting
distortion in the electron orbits may account for the shifting of an absorp-
tion band when a liquid crystallizes, and the natural double refraction
of crystals. The direction of this local field will alternate as we pass
from molecule to molecule through the space lattice, and is distinguished
from the mean molecular field in that it exists whether an external field
is acting or not. The forces, due to these mutual magnetic actions, are
responsible for the rigidity of cr^rstals and the existence of plane of
cleavage.

To obtain an idea of the intensity of the local moleciilar field, we as-
sume that it is of such a magnitude as to produce a change in suscepti-
bility of the order of that actually observed in the crystallization experi-
ments.

From the theory of Langevin, we have:

(22) ^"-?^^

M 4rm

where A M is the change in the magnetic moment of an electron orbit
of moment M by the application of the field H; r, the period of an elec-

tron, and — the ratio of the charge to its mass. From equation (22)
m

we have:

(23) ^'"f^"'

Ml 4rm

and

AMe_Hre

M« 4 T m'

where the subscript 1 and c refer to the liquid and crystalline states
respectively. In passing from the Uquid to the cr3rstalline state the
alteration of (Mi) produced by the local molecular field H, is A M/,
where:

(24) ^M/^ erHo

Ml 4 T m'

160 FERROMAGNETISM— INTRINSIC FIELDS: TERRY

and

(25) Me=-Mi±AMi'.

^though He alternates as we pass from molecule to molecule, the sign
of A Mi^ will remain the same, for when He' changes sign, Mi reverses
also so that every molecule suffers the same distortion due to the local
molecular field. The double sign implies that the arrangement of
molecules due to their particular kind of packing will be such that in
some cases hx is positive and in others, negative. From equation (24)
and (25) we find that :

m m..m,(i±2l&);

also that:

(27)

AMc-AMi Te/ldberiH

AM

LMi_Te/ld:eTiHe\ ^

1 ri\ 4 T m /

The electrons which give rise to diamagnetism also produce the
Zeeman effect, a sUght change in frequency being responsible for both
phenomena. We may, therefore, write Te=Tii:5r, where 5r is the
change in period produced by the local molecular field H^ when crystal-
hzation sets in. From equation (15) it follows that

(28) xc= -gp A Me, xi = ^ A Ml, and

5x = Xc-X=~(AMe-AMi),

where n is the number of electrons per molecule and N the number of
molecules per gram. The change of period dr is defined by:

(29)

gr e Ti He
r 4 T m

From equations (27), (28), and (29) it follows that:

(30) »?=(i±«_:i^.)(i±llLl!)_i.

X \ 4Tm/\ 4Tm/

This equation gives the order of magnitude of the local field Ha in toms

PERROMAGNETISM— INTRINSIC FIELDS: TERRY 161

of the percentage change in x on crystallization. In all the substances
investigated this change amounts to a few per cent. Hence:

1 _e^r»He'
100'"l6ir»m«'

Taking n = 10"" seconds, and — = 2. X 10^ we get:

m

He =6X10^ gauss.

We have no data at present as to how far an absorption line is shifted
when a substance passes from the liquid to the crystalline state, but such
evidence would be a direct test of the magnitude of He. On the other
hand, it is known that the magnetic double refraction induced in a
liquid is proportional to the square of the external field. If we assume
that this law holds up to fields of the order 10^, we should expect on the
basis of the local field idea for a crystal, a double refraction about
40,000 times as great as the largest values induced in a hquid. This is
about the ratio of the double refractions of nitrobenzene subjected to a
field of 3X 10* gausses and the natural double refraction of quartz. The
fact that most uniaxial crystals have a double refraction comparable
to that of quartz, and hence, a magnitude much greater than that
induced in liquids by fields available in the laboratory would support
the idea that the intrinsic molecular field, if interpreted magnetically,
must be of an order high compared to 3X10*. These fields are even
larger than those observed for ferromagnetic substances when inter-
preted according to the Weiss theory.

The Stresses and Energy Associated with the Molecular Field.

If there exists a molecular field of the order deduced in the previous
sections, then the forces associated with the diamagnetic crystalline
structure must be very large and the potential energy of the crystallire
state will be considerable. It should, therefore, be possible to give a
rough check on the value of the local molecular field from a consideration
of the latent heat of fusion of crystals. If |i| is the local magnetic
moment which in conjunction with the local field Ho, binds one molecule
to another in the crystalline structure, and if all the elementary systems
are independent, then the energy possessed by one gram of the substance
in virtue of a particular crystalline grouping, may be written:

m E- t n

2po

162 FERROMAGNETISM—INTRINSIC FIELDS: TERRY

where n is the number of molecules per cc, p the density, and I = n |i
the aggregate of the local intensity of magnetization per cc. Here a/
IB the constant of the local molecular field as used above. The local
molecular field H^^h^' I has been shown to be of the order 10^, and
since a^' is of order 10*, it follows that I is of order 10*. Hence, the
energy per gram given by equation (31) is of order of 10*, the thermal
equivalent of which is approximately 25 calories. This represents the
energy necessary to destroy the crystalline structure, that is, the latent
heat of fusion. It is of the right order of magnitude since a large
niunber of diamagnetic crystalline substances have latent heats ranging
from 21 for aniline to 44 for acetic acid. It is also the order of mag-
nitude of the latent heat of transformation of iron from the ferro-to
the paramagnetic state as found by Weiss and Beck. It is obvious that
until we know the arrangement within the cr3rstalline structure the
value of ae must necesssarily be merely an approximation; but the fact
that it agrees even as regards the order of magnitude is good evidence
for the existence of such local molecular fields and intensities as have
been assumed.

Molecular Field and Tensile Strength.

Whatever may be the nature of the forces which hold the molecules
of a liquid together, we have in addition to them, on crystallization,
those of the intrinsic local field. If it is assumed that the only addi-
tional forces binding molecules together on crystallization are those
due to their magnetic fields, then it should be possible to predict their
tensile strengths from considerations of their local fields and intensities
of magnetization. The potential energy associated with each unit
volume of a crystalline substance in addition to that when in the liquid

form wiU be

1

2

^H« I.

This is then a measure of the mechanical stress which binds the molecules
together and determines the rigidity of the substance. In a previous
section it has been shown that for diamagnetic substances I is of the
order of 10* and since H^ is of order 10' it follows that the tensile strength
should be of the order .5X 10' dynes per squares centimeter. That this
is of the order experimentally determined in some cases may be seen by
comparing with glass Ll-LSXlV, quartz 10X10", lead .16X10*, etc.
Moreover if one uses the corresponding values of intensity and molecular
field for ferromagnetic substances as determined by Weiss, he obtains
the following values for tensile strength: iron, 5.5X10*, nickel 1.4X10*
and aubalt 4.4 X 10* which compare favorably with the observed values.

FBRROMAGNETISM—INTRINSIC FIELDS: TERRY 163

It may then be concluded that the stresses due to the local molecular
field give a satisfactory interpretation of ultimate tensile strength of
crystalline media for both dia-and ferromagnetic substances.

The Change of Density on CrystaUization Interpreted as a
Magnetostriction EifFect of the Molecular Field.

It has been shown by Larmor^ that the potential energy per gram of
a diamagnetic liquid, the molecules of which have a small mutual in*
fiuence, is

(32) ^ [Ki W+\ Ki* m ,

where Ki is the susceptibiUty per unit volume and X is a constant approxi-
mately equal to - * If now a liquid is subjected to a magnetic field a

change of volume occurs such that the internal pressure is reduced by an
amount equal to the potential energy per unit volume of the magnetic
field. Since Ki is of the order of -7X10"^, the second term of (32) is
negligible compared to the first, and if C is the compressibiUty of the
liquid, the change in volume due to the field may be written:

(33) 5V=~CKiff:

a relation which has been verified by Quincke for fields up to 50,000
Gauss. If it is assumed that this law holds for fields of the order of
the local molecular fields, i.e. 10^ gauss, then the change of volume on
crystallization may be computed by replacing the first term of equation
(32) by the second expression of equation (31). There results then:

(34) 5V=^Cae'P.

From considerations involving the determination of the quantities
ae' and I from internal stresses accompanying the change of freezing
point with pressure, Oxley deduced for the substances listed below the
following values:

ae' = 2.5X10*, and 1=400.

Since C for these substances is of the order .8X10"^", there results:

d V=i .8XlO-*«X2.5X10*X16X10i"«.16 cc.

> Lsnnor, Proc. Roy. 8oc, A., 52, p. 63, 1802.

164 FBRROMAONETISM— INTRINSIC FIELDS: TERRY

The following are observed values of 3 V for a few substances.

SvbaianeeB (V

Benxene 10

Naphthalene 14

Benxophenone 19

Formic acid 10

Di-phenylamine 10

The calculated values agree as well as could be expected with the
observed values, since, for a^' and I, we know the orders of magnitude
only, since they are unknown functions of the molecular structure and
the space lattice which are different for each substance.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 166

THEORIES OF MAGNETIC CRYSTAI5 AND THE

MAGNETON

Bt J. KUNZ

AsMciate Profesaor of Mathematical Physics, Uniyersity of Illinois

The ferromagnetic crystals, which have been investigated so far, are
pyrhotite Fe? Sg, apparently hexagonal; magnetite Fes O4, of the cubical
system; iron crystals of the cubical system; and hematite, FcaOs,
rhombohedric and hemihedric. The majority of investigations are
due to P. Weiss and his coworkers.

The simplest phenomena are offered by Pyrhotite, which has first
been investigated by P. Weiss,^ and whose studies were continued
by J. Eunz' and by M. Ziegler.'

The methods of investigation are essentiaUy the same in all measure-
ments; they have been partly introduced and widely perfected by P.
Weiss and his students: they are either methods of deflection, or bal-
listic methods. The three dimensional problem is reduced to a two
dimensional one by cutting thin plates from a crystal, parallel to a
certain crystal surface. These plates, in horizontal or vertical position,
moveable round about a vertical axis, are placed in a magnetic field of
given direction and magnitude. If the plate is placed horizontally, the
deviation D gives the component In of magnetization perpendicular to
the direction of the horizontal magnetic field H, according to the for-
mula:

D-VXHXI«,

where V is the volmne of the plate. If the magnetic field is turned
round about the crystal plate, we find readily the normal component
of magnetization for the various directions of the crystal plate.

In order to determine the component Ip parallel to the field, we
suspend the same plate in a vertical position, so that the field falls in
the surface of the plate, which is at rest, R. If we now rotate the field
by a small angle a to the right or to the left, the plate will be subject
to a moment of force

D>=IpV. HBin(a-/3).

The plate itself rotates by a small angle fi. If, moreover, the plate
has a component of magnetization Is perpendicular to the plane of the

> p. Weifls, Lea propii^tte magn^tiques de la pyrrhotine. Journal de phy9%qu€t 1906,
p. 469.

> P. Weifls and J. Euns, /. d. Phy., 1905. p. 847.

* Max Ziegler. Kristall Magnetische Eigenaohaften dea PyrrhotinB. Diasertation
ZOrich, 1916.

\

166 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

plate, it win make a contribution 1$ cos (a—fi) to the moment. In
order to reduce this part to a minimum, we choee the plates as thin as
possible. In the case of the normal pyrhotite the magnetic plane facili-
tates essentially the measurements. Morever, the demagnetizing
action of the plates can be neglected in many cases, so that the external
6eld may be used directly as magnetizing field. Because of the correc-
tions P and Is this method is cumbersome and is often replaced by the
ballistic method. A primary coil produces a uniform magnetic field in
which is placed a secondary coil, S, connected with a ballistic galvan-
ometer. A ballistic deflection arises when the crystalline plate is intro-
duced or withdrawn from the secondary cofl, expressed by:

edt-GIp,V,

where G is a constant, e the induced e.mi. At the same time, with the
normal component, we can determine the hysteresis of rotation, by turn-
ing the field first in one, and then in the opposite direction roimd about
the plate suspended in a horizontal plane. The apparatus required
has been perfected and described by Weiss and his students (for instance,
in the thesis of V. Quittner and Earl Beck.)

We proceed to the results obtained with the various crystals, among
which the normal pyrhotite is distinguished by the possession of a
magnetic plane and rather simple magnetic properties.

PYRHOTITE

The chemical composition corresponds approximately to FerSi;
it crystallizes apparently in the hexagonal system, and its magnetic
properties correspond at most to the rhombic system.

A. Streng^ made in 1882 the important discovery of the magnetic
plane of the pyrhotite, at least for the permanent magnetism. These
measurements were made complete by Abt' and later by the detailed
measurements of Weiss, and Weiss and Eunz. We must distinguish
between two t3rpes of pyrhotite: the crystals from Morro-Velho in
Brazil, without cleavage> and with uneven fracture. The magnetic
properties are very simple. Weiss calls these crystals normal pyrhotites.
The abnormal pyrhotites are widely spread; leaf -like; with badly defined
magnetic properties; and with great thermomagnetic irregularities,
especially with respect to hysteresis.

The plane of base of the normal pyrhotite is the magnetic plane, in
which the crystal is much easier magnetizable than in the perpendicular
direction. The magnetic properties repeat themselves three times in

1 A. strong. Neu€9 Jahthuek der Mintraioaie, 1, p. 185. 1882.
> Abt. ITftfdemann'c AnnaUn, 1896, p. 135.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 167

angular distances of 60^ in the magnetic plane, but in various magni-
tudes. It looks as if the crystals were made up of three elementary
crystals (crystal components) placed side by side so that the magnetic
planes are parallel to each other and that the directions of easy magneti-
zation are inclined mutually by 60^. In order to obtain the properties of
the simple or elementary cr3rstal, we have to correct the measurements
by a graphical method of successive approximation. We chose such
samples in which one of the components predominates strongly. For
the purified crystal a curve arises of rhombic symmetry, where every
elementary crystal plate shows a distinguished direction, in which
saturation is reached by very weak fields; while in the perpendicular
direction up to 13400 Gauss are required for saturation to take place.

MEASUREMENTS AT ORDINARY TEMPERATURES

Fig. 1 gives the curves of the couple in the magnetic plane for 5550
Gauss. I represents the component In of magnetization perpendicular

1. Principal Component. (100%).

2. Second Component. (14.1%).

3. Third ComponeAt (2.5%).

Fio. 1

to the field. Fig. 2 gives In for the various directions and dif*
ferent fields. Fig. 3 gives the corresponding components Ip parallel
to the field. In passes twice through zero in the interval
from 0....180°, while Ip in the same interval shows only maxima
and minima; this is a common property of the two components for all
plates of all crystals. It is easy to construct the resultant I by means
of the two components. The result is shown in Fig. 4. If the
end point of the vector H covers the whole magnetic plane, the end
point of I, the resultant magnetization, remains within a certain circle
which Weiss called the circle of magnetization. If the vector H of the
field rotates with constant velocity round about the point O, then,

168 THEORIES OP MAGNETIC CRYSTALS AND MAGNETON: KUNZ

Fio.a

^^

r

v

f

/'H'^/eaacPa^ss

^-/y* '¥000

//

»?-//« 7J/0

//

^'N'/Z/^O

a

Fio. S

starting from the direction
of easy magnetization Ox,
the vector I of magnetiza-
tion foUowB at firat veiy
slowly the field ; its end point
remains on the circle of mag-
netization mitil H has nearly
reached the direction Oy
of diflScult magnetization.
Then I leaves the circle of
magnetization and curves
rapidly on a flat curve be-
hind the field, in order to
reach it in the direction Oy.
The larger the magnetic
field, the more the curve of
magnetization will approach
the circle of saturation. For
sufficiently high fields (13400
Gauss) the circle of satura-
tion will be described by I
with sufficient approxima-

THSORIBS OP MAGNETIC CRYSTALS AND MAGNETON: KUNZi IW

tion. (Between 30^ and
60^ deviations of about
1% occur). P. Weias
assumes that for an in-
finitely large, perfectly
homogeneous crystal
in the direction of easy
magnetization Ox satura-
tion is reached even in
the weakest magnetic
fields; in the other two
principal directions the
same would occur, if
there would not exist an
internal demagnetizing
field of magnitude N I,
where N is a constant
coefficient.

A-H«<1992 GauBB, B-H«<4000 Gauss

G-H»7310 Gauss, D-H-10275 Gauss

£-H» 11140 Gauss

FiQ.4

HYPOTHESIS OF WEISS

Intrinsic Molecular Field Hi

In order to represent the properties so far described of the normal
pyrhotite to a first approximation, P. Weiss makes the following assump-
tion: in the directions of the three principal axes of the crystal there
exists an intrinsic molecular field proportional to I in that direction
and proportional to a certain coefficient having a special value for each
axis. With respect to the sum of the external and the molecular mag-
netic fields H the crystal behaves like an isotropic medium. Let
X| Y, Z be the three perpendicular principal axes of the crystal, H, the
external field, with the components Hx, H„ H., the intensity of magneti-
zation I with the components I,, I„ lai the constant coefficients of the
molecular field Ni, Ns, Ns respectively; then the components of the
molecular field are equal to:

H,„=NiI,; H^ = N,I^, H^-N,I..

In general the molecular field has not the direction of I, except in the
direction of the three axes. The resultant components of the magnetic
force are equal to:

H,+Nil,; H,-hN,Iy, H.-hN,I..

170 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

If in a certain direction the resultant magnetic force coincides with the
resultant intensity I of magnetization, the following equations will hold:

(1)

H,+Ni I. _ Hy+Nt I, _ H5+ N, I. _
. n,

J.X Xy X,

where n is the reciprocal value of the susceptibility of the crystal which
is isotropic with respect to the total field. It foUows immediately:

(2)

Ix-

Hx

n-Ni

I,

H,

n-Nt

I.=

H.

n-N,

— — » — ' — ^are the susceptibilities in the direction of the three

n— Ni n— Nj n— Nj

axes with respect to the external field alone. Ni, Ni, N« are considered

as constants, while n must be considered as function of the sum of the

external and the internal field; for sufficiently weak fields n is constant;

therefore the curves of magnetization, according to (2), in the direction

of the 3 axes for small fields, are straight lines through the origin ot the

system of coordinates; the curve of saturation I=Ia is a line parallel to

the axis of H, and one straight line goes over into the other by a cotain

curve. If the magnetization is restricted to the plane xy, then we have:

'3)

or, considering Fig. 5:

H,+Nil, Hy+N« ly
I« " ly '

H cos a+I Ni cos ^ H sin a+Ni I sin ^

Icos ^

Isin ^

or

Fig. 6

I H sin (a-^) = (Ni-N,)
P sin ^ cos ^:

or

H sin (a— ^) = N I sin ^
cos ^

if weputNi-Ni=N. The
independence of the coeffi-
cient Ni— Ni of the mag-

netic field can be tested in the following way according to Weiss.

THBORIBS OP MAGNETIC CRYSTALS AND MAGNETON: KUNZ 171

H,=H,-H, tan /3-H,-H.??-I,(l.^-5?^,

ar, by means of (3):

H^«(Ni-N,)I^.

Hence I, aa function of H<| is a straight line, passing through the origin.
For saturation we have :

I«(Ni-N,) = 7200.

In large fields the agreement is good; in weak fields deviations from
the straight lines occur, which are not yet explained. N^ecting these
snliall deviations we may state: the crystal destroys a component H^ of
the field proportional to I,; the remaining component Hj is proportional
to the magnetization (and parallel) to I.

If for smallest fields saturation shall be obtained in the direction Ox of

easy magnetization, then _^ must be equal to oo, or n— Ni=0.

In the other two principal directions the same would be true, if it were
not for a demagnetizing field N I. The curve of magnetization in the
direction 0 x should be a straight line parallel to the axis H; in the
direction 0 y a straight line inclined toward Hi, and the deviations may
be explained by a lack of homogeneity of the crystal. This points to
the necessity in these magnetic measurements of testing at first the
crystals by the usual crystallQgraphic methods for purity and homo-
geneity. A physicist and a crystallographer ought to cooperate in
tiiese investigations. The approximate truth of the theory can be
tested by the moment of force D. Here also small deviations between
theory and experiment, amounting to about 3 per cent, occur.

P. Weiss has given the following interpretation of the law H sin
(a— «p) — (Ni— NO I sin ^ cos ^. In a state of equilibrium the molecular
magnets shall be distributed in parallel straight lines within the mag-
netic plane, so that the crystal presents saturation in the direction 0 x,
even without an external field. If now under an angle a the field H
acts, the magnets will turn away from the direction 0 x and assume a
new position of equilibrium, given through the angle p. We assume
that the adjacent magnets act in such a way upon each other that
there results upon a pole in a magnetic force A m cos ^ in a horizontal
direction, and a force B m sin ^ in a vertical direction. Then the resul-
tant X component of the magnetizing force will be:

Hx»H cosa+A^cos <p,

172 THSORISS OP MAGNETIC CRYSTALS AND MAGNETON: KUNZ

and:

H,">H8ina— B^isin ^.

But the equilibrium requires:

or:
or:
fdien weput:

H, sin ^""H, ooe ^;
(H ooe o+A |i ooe ^) sin ^--ooe ^ (H sin a— B /a sin p);

Hsin (a— ^)BNIsin ^ooe ^
(A+B)m-NL (Kg. 6)

Fig. 6

HYSTERESIS

We consider at first the ordinary alternating hysteresis in the direction

0 X of easy magnetisation. According to the molecular scheme in the
direction O x saturation occurs even without a field; if then the external
field begins to act in this direction in increasing magnitude, the magneti-
zation I remains constant and will be represented by the straight line
AB of Fig. 7. If now the field decreases, I remains unchanged until H
assumes a negative value — He; then the molecular magnets swing round
suddenly in the opposite direction and the end point of the vector

1 goes from C into C and moves then upon the straight line C B'.
If we reverse the direction of the field, the magnetization passes through
the points B' C A' A B. The loop of h3rsteresis is therefore the right
angular surface A C C A'. Per unit volume and per cycle the energy
dissipated is equal to:

4 He I«=4.15.4.47=2900 ergs; He=15.4,

while in the same crystal the demagnetizing field in the direction O y
was 730 Gauss. The connection between these two quantities has not

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 173

yet been found theoretieaQy. The experimental curve approaches the
theoretical one to a certain degree. It is similar to the hysteresis loops
of ordinary iron. The distance between two points in the same hei^t

CIA

B

remains nearly constant and equal to 2 Hq. We proceed now to the
hysteresis of rotation. If we turn the magnetic field round the crystal
plate in one and then in the opposite sense, it will describe a curve with
a loop, indicated in Fig. 8. In the neighborhood of M, which corre-

^f^

Fig. 8

sponds to the direction of easy magnetization (0 x in Fig. 9) there is
no hysteresis in a wide range of angles, which corresponds about to the
arc AB of Fig. 9. We shall assume with P. Weiss that on the arc A B

174 THB0BIB8 OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

of the circle of saturatioD there ia no hyBteresis, but that hyBteresis ap-
pears, ^en in an irreversible process the magnetisation passes from B
through C to D. The hysteresis of rotation oug^t to disappear in very
strong fields, when the magnetisation describes the circle of saturation.
For the direction 0 x the coercive field is H«, for the parallel directicm

B D we shall assume H«'->H« A= — - also a constant coercive force.

llB

If I,">Iai, then H«'-bO and He' ou^^t to diminish with I,. As a matter
of fact the hysteresis does not disappear for I,»Iai on a straight

line. Moreover, the law
depends on the manner in
which the moments of
force are measured
because of viscosity. The
energy dissipated per
cycle and per unit volume
would be equal to 4XH«'
times the length of the
cord C B« Vln'-V

FiG.s E=4H.Vl«»-V. But

here experience agrees
even worse with the theory of Weiss than before and in the case of
abnormal pyrhotite crystals the phenomena are much more complicated.

INFLUENCE OF TEMPERATURE

At first the influence of the temperature on 1^, n— Ni, n— Ns, n— Ns
of the normal elementary crystals of pyrhotite has to be investigated.
This has been done by Ziegler, by means of the method of couples and
by the following considerations:

D

-V H 1,,= V H I sin (a-^)«V P (Ni-NO sin ^cos ^.

We consider two adjacent positions of H and I. In the first position
both vectors shall coincide with the axis x, in the second position they
shall be removed by an infinitesimal amount. Then putting the last two
expressions equal to each other, we obtain:

or

H, A (a-v5') = I, A^ (Ni-N,) = H, (Aa-A ^);
Va ajr (Ni-N,) I.+H.' ^"""^ ""'

AD,-V(Ni-N,)A^,

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 175
hence

\Aa/x VdaA (Ni-N,)I,+H/

As saturation in the direction x is obtained by fields of 2000 Gauss, we
can write for the experiments made Im instead of I, and obtain:

VdaA (Ni-NOIm+H/

and in the direction y:

\da /y n— Ni ;

moreover, for the maximum couple:

d„=^\ni-no.

The experiments which required great skill, showed that Ni—Ni is
independent of the temperature, and that n — Ni remains constant nearly
up to the Curie point, and then probably increases rapidly; n— N|
increases in the neighborhood of the Curie piont, I. decreases, or the
magnetic plane becomes more pronounced at higher temperatures.
In has a distinct maximum at about 160° of the absolute scale. In the
neighborhood of about 320° Im vanishes rather abruptly. These features
do not correspond to the theory which P. Weiss has given of the curve
Im, T. If normal crystals are heated imder the simultaneous influence
of a magnetic field up to 330°, the magnetic plane remains imchanged
and the ratio of the three crystal components remains also the same.

THE ABNORMAL CRYSTALS OF PYRHOTITB

While the magnetic properties of the normal pyrhotite are compara-
tively simple, those of the abnormal pyrhotite are very manifold and
have not yet been interpreted theoretically. Already in weak fields
the magnetization is not restricted to the magnetic plane, but reaches
considerable amounts in the perpendicular direction, neither does the
law of the demagnetizing field H<i= (Ni— N2) I, hold. The phenomena
of hysteresis are much more developed in the abnormal than in the nor-
mal crystals. Hysteresis appears not only in the direction of difficult
magnetization, but also of easy magnetization, which even increases
with increasing field, sometimes even with the number of cycles de-

176 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

scribed. If we determine the coupte as function of rising and falling
temperature under the influence of a constant field, the resultant curves
show even thermic hysteresiB. After the heating and cooling of a crystal
in the magnetic field, the couple does not return to its original value.
After the first heating the couple often appears to be increased, in the
ratio 1:3 or even 1:4, 5. It may also decrease.

CRYSTALUNE GROUPS

The curves of the couples which result when the magnetic field is
turned round the horizontal plate show, as in the case of the normal
pyrhotite, in general three groups or components of crystals in various
ratios. In a given case the original composition was given by the ratios:
66.3 : 25.6 : 8. 1. After heating and cooling imder the simultaneous action
of a field of 2000 Gauss, which acted in various directions, we found the
ratios : 46.9 : 29.4 : 23.7. It looks as if the elementary magnets under the
action of high temperature and of the magnetic field distribute them-
selves in different groups among the three components. If the crystal
is heated without a magnetic field, the three components are about
equally strong, this is also true if a field is applied and rotated during the
cooling process. The new grouping requires a small interval of time,
i.e., more than 2 seconds and less than 4 minutes. The hysteresiB in
the ''magnetic plane", gives another method of determining the ratios
of the three components.

While in the normal p3Thotite all phenomena, even hysteresis, are
reversible in a thermal sense, i.e., nothing appears similar to the harden-
ing and tempering of steel; in the abnormal p3Thotites slow or rapid
cooling affects the curves of the couples as well as those of the h3r8teresis.
As in the case of steel rapid cooling increases the hysteresis.

If we start from a state of equal distribution of the three components
of the crystal and heat it up to 350° and then cool it while at the same
time a field of 500 Gauss acts in the direction of one of the three maxima
of magnetization, then this maximum acquires a dominating influence
and the (^wo\thef maxima remain equal to each other but remain inferior
to the first maximum. With increasing field the uneven distribution
of the three components becomes more conspicuous. It is not yet
known whether the directing influence of the field tends towards a
limit or not. According as a field of about 5000 Gauss coincides with
one or the other of the three maxima, that component will be increased
at the expense of the two adjacent components of crystallization, which
according to Fig. 10 amount to about 50 per cent, of the principal com-
ponent. The three maxima play the same role not only with respect to
the couples but also to the hysteresis. The varying ratio of the three
components might also be explained by the assumption that the mag-

THEORIES OP MAGNETIC CRYSTALS AND MAGNETON: KUNZ ITT

netic state of the three crystal groups is an independent variable. The
experimental answer to this question was not decisive.

We have given here a somewhat
detailed description of the magnetic
properties of normal and abnormal
pyrhotite; of the first, because it is
the only crystal whose properties
we understand approximately owing
to the theory of Weiss; of the second,
because it shows a large variety of
phenomena which are imexplained
problems. In his theory Weiss has
considered the mutual action of the
elementary magnets in a summary
way by introducing the molecular
magnetic field as proportional to the
intensity of magnetization. The
dynamics of crystal structure has to
be applied to the magnetic pheno-
mena.

In aD other magnetic cr3rstals we
have not yet a sufficient theory.
We shall only point out a few of
the manifold phenomena, as it would be useless to give a full descrip-
tion of the large variety of all magnetic properties known in these
substances. For a fuller account of the experimental facts we refer
the reader to the original researches.

Fig. 10

MAGNETITE

In an investigation of the year 1896, P. Weiss showed that magnetite,
though a crystal of the cubical system, has certain directions in which
the magnetization is either too weak or too strong according to the
cubical system. The theory of the magnetic ellipsoid of W. Thomson
rests on the assumption that the intensity of magnetization is propor-
tional to the magnetic field. But as magnetite is ferromagnetic, where
that proportionality does not exist, the magnetic theory of Thomson
does not hold. The investigation of Weiss was continued by his student

Quittner.*

SYMMETRY

The magnetic properties of many, if not all, samples of magnetite
crystal deviate more or less from those of the cubical symmetry. If a

iV. Quittner. Die magnetiaohen Eigenaohaften des Magnetitee. Thesii, ZCkrioh^
1906 und Annalen der Physik, 1909, N. F. 30.

178 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

thin plate paraOel to the cubical surface is cut out of the crystal, then ac-
cording to the cubical symmetry and the quartemary axes the magnetic
properties should repeat themselves four times within 360°. One
might expect a curve of the type (a) of Fig. 1 1. Instead of this curve we
obtain different curves of less symmetry according to the intensity of
the field. The axes are at (f , Wf, I8(f, 27(f, the diagonals of the
cubical face pass through 45^, 135^, 225°, 315°. Instead of four waves
the first curve obtained with a field of 57.3 Gauss shows only two identi-
cal waves. With a small deviation the phenomenon repeats itself from
180° to 180°. In the second curve we observe an inflexion in the diagonal
at 135°, which increases with increasing field, in order to form a new

a

^•%

^^

/"»

:r\

/'■^,

•^

/ \

JF^

/ \

Wi

1 \

yi

\^

9 \

A

1

1 \

/f

\

/ i
t

A^

1

/d

0X

\

j

/ik

///'

k '

0

1 /

\

1

1

/v

r" .'

m

ll^

1 J

ll

f

If /

1 1

%i<

1/

\

1

vi

'/

w

'J

^

w

(T SO* /dO*

270* 36C

l-H»57.3GauM 2-H-78.4 Gauss

3-H=94.5 0an»

Fi

o. 11

wave which reaches an amplitude, at 300 to 400 Gauss, almost as large
as the first wave. The greater the magnetic force, the more magnetite
approaches the cubical symmetry. The two principal axes of the cubical
face are never quite equivalent and the question arises as to the magnetic
behavior of the third axis. In order to answer this question, a plate
parallel to the octahedron has to be cut out of the crystal and to be
tested magnetieaUy; this face contains no principal axis, but makes
equal angles with aU three axes. If the symmetry is cubical, three
equal waves must appear within 180° ; if it is quadratic, two equal waves,
the third different; and if it is rhombic, the three waves are different,
one or two of them may even disappear. In Fig. 12a (63.0 Gauss) only
two different waves appear; for a stronger field (78.7 Gauss) three
waves with different amplitudes appear, and these three waves become
similar to each other only for much stronger fields; while for very weak
fields only one wave appears. According to these magnetic properties

TBSORISS OF MAONBTIC CRYSTALS AND MAGNBTON: KUNZ 179

magnetite has at most the Bymmetry of the riiombic system. The
curves a and b of Fig. 12 correspond to the motion in one and the
opposite direction of the magnetic field, their distance is a measure of
the hysteresiB of rotation of I^. The middle curve is the curve d
magnetization. The d^ree of deviation from the cubic symmetry
ia very different for the different crys-
tals, plates even have been found which
showed abnost no deviation, but they
bad a very pronounced viscosity. If
certain plates parallel to each other are
cut out of the same crystal, they show
as a rule identical properties, but in ex-
ceptional cases their magnetic properties
vary. These phenomena again show
the neceesity of testing the crystals by
etching meliiods before they ore used for
magnetic measurements.

A large number of measurements of
the normal and parallel components of
magnetization in thin plates cut parallel
to the faces of the cube and of the
octahedron have been made. For very
weak fields the material is practically o'
isotropic, and seems to approach is- i-H-63.0Gau8s

otropiam for very strong fields; the normal ~ -tbt Gaun

component !„ approaches zero in very ^

strong fields. Strange reversals in the maxima of the curves occur;
they appear for certain fields on the projections of the axes, for other

/90*

Pio. 13
fields on the sides of the triangles. On the whole the phenomena are
very complicated.

V. Quittner has discussed several theories of molecular structure at
magnetite and has reached the following conclusion : Magnetite consists

180 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

of equal parts of three elementa with magnetic planes similar to pyr-
hotite; the magnetic planes of these three elements are perpendicular
to each other. The molecules are arranged so as to form three systems ai
planes at equal distance from each other. Ttus ideal magnetite is cut
by four systems of intermediate unmagnetic planes, which are paraDd
to the faces of the octahedron. If these four planes (of cleavage) are
equal to each other then the crystal presents cubical symmetry, if they
are di£Ferent, the symmetry has a lower d^ree. The Figs. 13 illustrate
the structure assigned to magnetite by Quittner.

IRON CRYSTALS

Iron crystals have been prepared in the following way by de Freuden-
reich: Castings of the Goldschmidt process were made in a crucible
imbedded in a heap of sand and slowly cooled down. Tbe resulting
pieces of iron, of the size of a fskst, were broken and showed very beauti-
ful crystaljif surfaces up to 2 cm. of length, which were identified as
<nibical surfaces. The chemical impurities amounted to about 2 per
cent., of which silicon was the predominating part. The magnetic
investigation was carried out in the laboratory of P. Weiss^ according
to the methods of the couples (1^) and of induction (Ip). In the results

Beck gives the components of magnetization per unit mass a » -, where d

d

is the density. A number of plates were subject to magnetic measure-
ments. They were cut parallel to the faces of the cube, the octahedron
and the rhomboidal dodecahedron.

[ Figs. 14, 15 and 16 give the components of magnetization <rp paraDd
to the magnetic field of three plates cut parallel to these three faces for
different field strength. Fig. 17 gives the components of magnetization
(Tq perpendicular to the field for a cubical plate, and Fig. 18 gives the
resultant magnetization in various directions for three different fields.
We consider ( t first the curves of the Fig. 14. In weak fields <rp is nearly
constant in tL various directions, but in H the demagnetizing action
of the plate is not considered. If the field increases characteristic
differences appear in the different directions; minima appear in the
diagonals; maxima in the quartemary axes; and the differences are most
pronoimced in middle fields of about 392 Gauss, where they amount to
about 12 per cent, of the average value. These differences disappear
again with increasing fields. The maximum value of <rp measured
was 208.3 for a field of 4090 Gauss. If we compare these ciuves
with the corresponding curves of magnetite, we find that the axes play
the opposite role in the two cases; maxima are found in the direc-

> Karl Beck. Das magnetiBche Verbalten von Eimmkristallen bei gew5hn!i ^her Tern-
p«ratur. Diasertation Zurich, 1918.

THEORIES OP MAGNETIC CRYSTALS AND MAGNETON: KUNZ 181

'^^fM

iSO

IOC

MO

\/\/

IL

ISO

/«0

MO

or 4r io' fss' lir

it

^ ^ ^ ^

I

I-

ISO

/0»

so

I I

^^_^— ^— $

^

o*

Act isor Its'

r ss'soras* lecr

^jmMt^CKsar 6/^,2111, Fires or

Cuts Ml^omdo^sif Ihifsetsffs^/Toi^ Oefa/?seOvf9

Fig. 14 Fio. 15 Fio. 16

tion of the diagonals of magnetite. In the case of iron, too, between
1500 and 1000 Gauss, small secondary maxima appear in the direction
of the diagonals*

182 THE0RIS8 OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

36

'M

f\

ZO

\\

~n

fO

i

B

'to

V

45 SO /3S

1-H= 71Gau8s
2-H» 189 Gauss
3-H- 392 Gauss
4-H- 595 Gauss
5-H» 1110 Gauss
6-H»2050 Gauss
7-H -6000 Gauss
8-H»9420Gauss

Fio. 17

The symmetry required by the cubical system, ia not rigorously
satisfied by the magnetic properties of iron crystals, but the deviations

are very small, at all events much
smaller than in the case of magnetite.
The component v^ of magnetisation
normal on the field shows in a range
of 18(f (Fig. 17) for middle and large
fields as function of the azimuth ^ of
(H); a curve with four seros and two
positive and two n^ative nearly
equal maxima. The zeros correspond
to the direction of the diagonals.
While (Tp remains always positive, v^
becomes once positive and once nega-
tive in every quadrant. This happens
because in the neighborhood of the
diagonals <r rotates more quickly
than H. The amplitudes of the
curves t)f v^ are very small in small
fields; increase in middle fields rapid-
ly to a maximum* and decrease
again with increasing fields. The
maximum of a^^ is about 20 per cent,
of the maximum of crp. Fig. 18 con-
tains the resultant a from crp of Fig.
14 and Oj^ of Fig. 17. Between these
curves and the corresponding curves
of pyrhotite characteristic di£Ferencee
appear. In the first place, there is
no special direction of easy mag-
netization, the circle of saturation is reached here in all directions
only for an infinitely strong field. Moreover, the sudden swinging
round of the elementary magnets across the direction of difficult mag-
netization does not occur in iron, at all events this phenomenon takes
place in the direction of the diagonals to a small extent. By means of
the given curves the curves of magnetization in any given direction can
be drawn.

We proceed next to plates cut parallel to the face of the rhomboidal-
dodecahedron. The paraDel component of magnetization <rp of a
favorable plate is given in Fig. 15. The plate contains at Od and 180d
a quartemary axis, at 90^ one binary axis and at 55^ and 125^ a ternary
axis. In middle fields we see again as in the cubical face the di£Ferenoe
between the principal or quartemary axes and the binary axes or

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 183

diagonals. Moreover, the directions of the ternary axes are weaker
magnetically than the other two axes so that (7p offers a maximum in
the binary axes between the two ternary axes. As in Fig. 14, so in Fig.
15, the differences of (7p disappear in the various directions for very weak
and very strong fields. The normal components otq show again four
zeros within 180° in the direction of the axes at 0°, 55°, 90°, 125° and
180°. The two maxima, to the right and to the left of the binary axis

1-H«392 GausB, 2-H-695 Gauss, 3-H- oo

Fig. 18

at 90° are smaller than the other two maxima. The amplitudes of the
curves reach a maximum in middle fields and decrease with increasing
fields.

We consider finaUy the face of the octahedron. The corresponding
curves show partly the same features as in previous faces, partly simpler
properties. Here we find only three binary axes at an angle of 60°.
Fluctuations of the magnetization must therefore repeat themselves
three times within 180°. This becomes evident from the curves of Fig.
16. The fluctuations disappear again in weak and strong fields. For
H s 189 and 392 Gauss o-p has a maximum in the direction of the binary
axes, but an inversion of the fluctuation appears already for H = 595
where the maxima appear in the binary axes; but the variations are
very smaD, and are apparently only of second order of magnitude.
0-0 assumes only very small values in aU fields; it describes three sinusoidal
cmrves within 180°. The amplitude of the waves of tr^ reaches a small
maximum already in smaU fields and vanishes in stronger fields almost
completely. The resultant o- in a polar system of coordinates is given
by a curve very little deviating from a circle. The phenomena of
hysteresis remind us of those of the curves of magnetization. The
energy dissipated by hysteresis increases at first with increasing field,
reaches a maximum and decreases again.

Ig4 THB0RIB8 OP MAGNETIC CRYSTALS AND MAONBTON: KUNZ

The magnetic properties of iron crystala on the whole are quite
different from those of pyrhotite. There ia no magnetic plane and the
existence of an intrinsic molecular field is not established.

HEMATITE

The magnetic properties of this crystal have been investigated by
Westmann, Bavink, Abt, Kunz and Smith.^ Crystals have been used
of different origin, which may be very different magnetically. They
are found on the Vesuvius, in Dagnaska in Hungary, in Ouropreto in
Brazil, in Bchabri in the Ural mountains, in Caveradi in Graubiinden,
Switzerland; beautiful crystals come from Elba and from Siebenbtirgen.
Eunz found that the crystals can be divided into two groups, similar to
those of pyrhotite, normal and abnormal hematites. Conmion to both
is a magnetic plane, as in the case of pyrrhotite, which coincides with
the plane of base. Perpendicular to this plane hematite is paramag-
luetic without hysteresis.

The crystalline plates, parallel to the base plane, often arranged in
table-like groups, from the Vesuvius, show a regular magnetic behavior.
They seem to possess only one component of crystallization and have a
small hysteresis both for alternating and rotating fields. At a tempera-
ture of 650^ under the simultaneous action of a magnetic field, they
undergo no change.

The other group of crystals from Caveradi in Graubtinden, from Elba
and Siebenbtirgen, seem to be composed of several elementary magnets,
which are inclined by 60^ to each other, and appear in various ratios.
They have such large hysteresis that often the hysteresis predominates
over the intensity of magnetization. Under the simultaneous influence
•of high temperature and a magnetic field, the magnetic structure of
the crystal changes, as in the abnormal crystals of pyrhotite.

Group I; Abnormal Crystals.

At first the component of magnetization I^ normal to the magnetic
field was measured by means of the couples and a great variety of phe-
nomena was found. In the weakest fields of 138 Gauss the crystals of
Siebenbtirgen and Granbiinden showed for I^ a pure sine curve (Fig.

1 J. Weetmann. TJpBala. Math, och Naturw. 11, 1807.

B. Bavink. Dissertation G6ttiiigeii, Neuea JahHmdi /fir Mineralogie, Bd. XIX,
877. 1904.

A. Abt. 'Ober die mag. Eigensehaften des Hematites. Ann. der Phynkt N. F. 68,
p. 668, 1899.

J. Kuni. t^ber die magnetiaehen Eigenshaften des Hematites. NeueB Jakrimeh der
Mineralooie, I Bd. p. 62, 1907.

T. T. Smith. The magnetic properties of hematite. Phynoal Renew, Vol. VIII,
p. 721, 1916.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 186

19, 1). The magnetization is like that of a pennanent magnet. For
H»339 Gauss in the second curve of Fig. 10 a deviation from the sine
curve appears with hysteresis of rotation, which under H = 700 Gauss
reaches a considerable amount; while curve 3 of magnetization still

l-H«138GauB8, 2-H«339 Gauss, 3~H-700 Gauss

FlQ. 19

preserves the same uniform character as in weaker fields. But at
H = 1400 Gauss in the interval of 180^ three equal maxima and minima
appear, and the hysteresis'assumes
values so large, that it is the princi- 4h.
pal phenomenon, Uttle influenced
by the variations of magnetiza-
tion. The band between the
curves a and b of Fig. 20 corre-
sponds to the hysteresis of rota-
tion. With increasing fields the
fluctuations of the curves of mag-
netization increase (Fig. 21 for
H = 11300 Gauss) at the same
time the maxima of magnetiza-
tion are displaced towards each
other for the forward and back-
ward motion of the magnetic field.
In other crystals the phenomena
become still more complex, for in-
stance, the three equal waves of the
last figure may be very different.

Flo. 20

186 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

Fig. 21

Group II; Apparently Simple Crystals.

These crystals are toiind on the Vesuvius. They form thin black
leaves of high luster, often arranged in crystallin aggregates of higher

l-H-i48.5 Gauss, 2-H-i 161.6 Gauss
3-H-421 Gauss, 4-H-862 Gauss

Fig. 22

order. The normal components 1^ of magnetization for different fields
are given by the curves of Fig. 22 and Fig. 23. For fields varying from
48.5 to 3500 Gauss the phenomenon repeats itself from 180^ to 180^.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 187

Two maxima and two minima appear, most pronounced for Hs862
Gauss. Above 6000 Gauss one maximum and one minimum disappears,
and dissymmetries make themselves felt. Surprising are also the points
of intersection of the curves A and A' above the axis of abscissas.
The curves of Fig. 23 are
essentiaUy different from
those of pyrhotite. The
amplitudes of the corres-
ponding curves of pyrhotite
reach a maximum with in-
creasing field and these
curves show very distinctly
the directions of easy and
of difficult magnetization.
These features do not appear
in the curves of hematite.
The resultant intensity of
magnetization increases with
increasing field, reaches a
maximum and decreases
again in higher fields. The
maximum value of I is only
about 3.6 absolute units.

The influence of temper-
ature on normal and ab-
normal hematite crystals is the same as on the corresponding crystals of
pyrhotite

ON THE THEORY OF FERROMAGNETIC CRYSTALS

Besides the theory of P. Weiss* for pyrhotite and K. Beck for iron,
we owe theoretical studies to K. Honda and J. Okfibo' and to O. E.
Frivold.'

K. Honda calculates at first the forces X and Y exerted on a magnetic
pole P by eight adjacent magnets arranged in a plane according to the
cubical system, he neglects the action of the remaining magnets as
insignificant. Let 2^, Fig. 24, be the sides of the space lattice; 2r
and m the length and the pole strength of the elementary magnets,
respectively. One side of the space lattice we take as the x axis, the
other side as the axis of y. Denoting by X and Y the sum of the com-

> P. Weiss. Le travail d'aimantation des cristaux. /. d. Phy. Ill, p. 194, 1904.

* K. Honda. Science Reports 5, p. 153, 1916.

' O. E Frivold. Ann. d. Phyeik, Bd. 65, p. 1, 1921.

1-H- 862 Gauss, 2 -H- 1775 Gauss
3-H» 3850 Gauss, 4-H» 7172 Gauss
5 -H» 10750 Gauss, 6-H- 13340 Gauss

Flo. 28

188 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

ponents of the magnetic forces exerted by the neighbors on the
central magnet in the directions x
and y, we find as the condition of
equilibrium :

m Hsin (a-^)«— Ysin^+Xcoe^

Without the magnetic field, the
magnets point all in the positive y
direction; if a magnetic field under
an angle a with respect to the y axis
is applied, the magnets will turn
through an angle $ and assume a
position of equilibrium given by the
last equation.

a 2k 2k

Putting k.-,p-—, and q-j-p^

Fio. 24

we have a-'kr; 0< p <1;

0<q<-. By simple trigonometry we calculate X and Y and expand
the expressions in powers of p and q and we get:

^ . , ^^ mk3 6 7 . I p» ,1 9p« v

H«n(.-.).----«n4^^^(-+-|+...)

(1+2 k»)| ^a''^2 5V 2 2 7r ^

2 2 L4 (l+k*)| V.'^2 2 9r''^

+Ormr(i?+?l+ )]ain«2.}or:

m

H sin (a-tf) -F W = - f (fi),

r*

where f {d) contains k only as a parameter. F (0) is a periodic function
of 6, having - as its period. F {$) may generally be written in the form

it

F(tf)-Asin49,

where

A ^-j— ^(k)-/(k)8in«2fl+...|,

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 18^

and ^, iff are functions of k only. For small values of k, iff (k) becomes
very small in comparison with ip (k), and the form of F (0) approaches a
sine ciurve.

By an elementary complex Honda means a very minute crystal, in
which all elementary magnets are arranged in a definite space lattice.
A piece of iron consists then of a large group of such elementary com*
plexeSy whose axes point in all directions. We shaU at first study the
magnetization of such an elementary complex of the cubical system.
The component of magnetization I in the direction of the applied field is
obviously:

I = 2mr ncoe (a— ^) = IoCOS (a— ^),

where n is the number of elementary magnets and lo the saturation
value of the intensity of magnetization. Denoting =- by i, we have

(1) i = cos(a— ^)

moreover:

H sm (a— ^) « A sin 4 ^

or putting -j- —h, we get:
A

(2) hsin(a-^)=sin4^

A depends on the properties of a particular substance so also Iq. But
if we use the reduced i and h instead of the actual intensity of
magnetization and of the field, the relations (1) and (2) apply for
all ferromagnetic substances belonging to the regular system. If
h and a are given, equation (2) gives B and equation (1) the value of i.
Honda considers therefore equations (1) and (2) as the laws of magneti*
zation. These two equations can be solved by a graphical method.
In Fig. 25 foiu: curves representing the relation between i and h are
given, in which the angle a was taken at 30'', 70'', 120^* and 170^*, respec-
tively. These curves give the intensity of magnetization in the direction
of the respective field, when the magnitude of the latter is so varied that
it is always in equilibrium with the internal resisting force sin 4 9. In
the curve for a— SO'', the point a corresponds to the value of cos 30^;
as h increases, % becomes greater, but always less than a, and therefore
i = co6 (30—^) steadily increases, tending assymptotically to the value
of i« 1 with h= 00 . In the curve for 0 = 70", the point b corresponds ta
the value of cos 70^; as h increases from O, B and therefore sin 4 0 also

190 THBORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

increases. Since, however, the latter quantity reaches a frutYifniim at
^■■'y h must be diminished from a certain value of 6 upward, if the

magnetization is to be effected reversibly . With 0=- the resisting force

4

sin 4 0 vanishes and therefore h must be diminished to zero, with a

further increase of 6, sign 4 6 changes sign and therefore h must be applied

in an opposite direction, if the magnetization is to be made reversibly.

If $ approaches to 7(f , h becomes oo in the limit and the magnetization

tends assymptotically to unity. The curve for a^l2(f, which begins

at the point c on the negative side of i, passes through a maximum and

a minimum of h and coincides with the curve for a»3(f as the value

of i increases. The curve for a = 17(f beginning at a point d on the

negative side of i, passes through two maxima and one minimum, of

h with the increase of i and approaches assymptotically to the line i" 1.

From this figure we draw the following inferences:

Case I. 0<a<2. With the increase of the field, the magnets are

turned more and more toward the direction of the field, and coincide in
direction with the field, when it becomes infinite. If the field is gradually
diminished, the magnets return to their original position.

Case II. '<«<-. By gradually increasing the field, the magnets
4 2

are turned toward the direction of the field, till the internal resisting

force attains a maximum. From this position, a further reversible

rotation of the magnets can only be effected by diminishing the field

beyond zero. With a negative infinite value of the field, the magnets

are brought in the direction of the field. If the field be reduced before

the resistng force reaches a maximum, the magnets return to their

original position; if, however, it exceeds the critical value, the magnets

do not return to their original position with the withdrawal of the

field, but take another position of stable equiUbrium, different from the

IT

original by -. Hence in this case the magnetization is irreversible.

Similar oonclusions can be drawn from curves c and d of Fig. 25.

So far we have only considered the magnetization of a single complex,
but we are now able to study the magnetization of a mass of ferromag-
netic substance, such as iron, which consists of a great number of such
elementary complexes with their magnetic axes uniformly distributed in
all directions. Honda considers a complex in the form of a sphere,

4

and puts the action of the adjacent elementary crystals equal to -rl,

3

where I is assumed uniform Equation (2) has to be replaced by (3).

(3)

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 191

(H+- T I) sin (a— ^) = A sin 40.
«5

With these a umptions Honda gets a curve of magnetization and a
hysteresis loop which resemble closely the experimental curves. He
proceeds then to the theory of magnetization of magnetite. It has
been shown by W. Voigt [Gott. Nachrichten, (1900) 331] and Wallerant
(Comptes rendus 1001, p. 630) that a crystal of the cubical system is
not isotropic with respect to the magnetic properties, if the permeability
is a function of the magnetic force.

We shall consider a crystal of magnetite acted upon by a field parallel
to one of the faces of the cube. In the case of magnetite stable orienta-

tions of the molecular magnets are the three positive and three negative
directions parallel to the sides of the cube, but the directions of the
diagonals of the faces of the cube are the orientations of unstable
equilibrium. Hence, when the space lattice is quite regular, and the

192 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

thermal agitation is zero, the mag;nets in each group will take one of the
six directions of stable equilibrium in the absence of a magnetizing field.
For the calculation of the magnetization we take the two sides of
the plane lattice as x and y axes, and consider two pairs of all magnets
in the direction of positive and negative x as well as y axes and the re-
maining pair in the direction perpendicular to the xy-plane, the total
effect being thus zero. If the magnetizing field be applied in a direction
making an angle a with the x axis, the intensity of magnetization is
given by:

1/ T

i=-S cos (a— ^i)+cos (t— a— ^,)+cos (-— a— O
6 V 2

+COS (2+«-^4)+2cos(J-0| ;

or

i = - <coa (a-^i) — cos (af^,)+sin (a+^,)+sin (^4— a)+2 8in^4>,

where a and 6 are related by the equations:

^ sm 4:6 1 ^ sin 4 (?3 ^ sin 4 ^, sin 4^4 sin 4 6^

sin (a— ^i) sin (a+d^) sin (a+^,) sin (^4— a) cos 6^

In the calculation we must take into account the abrupt change of
rotation of the magnets in passing through the critical values of 0.
If h is very large or very small, it is readily seen that i is independent of
€K, that is, the crystal behaves as an isotropic substance. Honda proceeds
to calculate the magnetization in the direction of the x axis and of the
diagonal of the plane-lattice and finds a qualitative agreement between
the experimental curves of magnetization by Quittner and his theoretical
curves. For moderate fields of H = 100 up to H = 500 Gauss, the mag-
netization along the diagonal is much stronger than along the direction
of the crystallographic axis.
Let us next keep the field constant and change its direction from 0 to

T T 3

r, starting from the direction of the x axis. At a=0, -, -, - t, t, the

direction of the magnetization coincides with that of the field, so that
there is no component perpendicular to the field. The results are
graphically given in Fig. 26, the dotted lines represent the change of the
curves as affected by some want of regularity in the space lattice and by

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 193

thermal agitation. If we compare these curves of magnetization with
the empirical curves of Quittner we find not more than a qualitative
agreement, while several phenomena, for instance the reversals in the
experimental curves, are not explained. Honda calculates also the
two components of magnetization parallel and perpendicular to the
field in plates parallel to the face of the dodecahedron and of the octa-
hedron. The perpendicular components agree better with the experi-
mental curves than the parallel components, but there are marked dif-
ferences between the empirical and the theoretical curves.

i

Para//e/ Co/r?poneA?/^

04

Perpe/7c//C(//ar Co/?7poj7e/7t

Fig. 26

Honda gives a theory of the magnetic properties of pyrhotite, assuming
the hexagonal system. But its magnetic properties correspond at most
to the rhombic system, in spite of its hexagonal appearance. Honda
considers also the case treated by Weiss, in which each of the magnets
is acted on only by its neighbors in the same row. The condition of
equilibrium is expressed by the equation:

where

Hsin(a-^)=F(^),
F W=A sin2^[^ (k)-^i (k) sin^^l,

while Weiss obtained the relation:

H sin (a— ^) = A sin 2 ^.

194 THBORIBS OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

Honda points out that the assumption of Weiss, that the angle which
the internal force makes with the direction of the stable equilibrium,
is equal to the angle of deflection of these magnets, is not generally true.
The theoretical ciures, however, obtained by Honda do not closely
resemble the empirical curves. The directions of easy and of difficult
magnetizations are perpendicular to each other in the magnetic plane,
a fact which does not agree with the hexagonal system.

/

tg^

!^^5^^^^^^^^

S^S^BBhSIS

«

Af

^

^^

Njn

(/

HXIO^
Fio.27

/5 20

It should be mentioned that W. H. Bragg and M. Nishikawa investi-
gated the diffraction of Roentgen rays through magnetite (Phil. Mag. 30,
p. 305, 1915) ; and that A. W. HuU (Physical Review, 9, p. 84, and 10,
p. 661, 1917), investigated the iron crystal, which is characterized by a
centered cubic lattice, whose unit is a cube with an atom in each comer,
and one at the center of the cube.

O. E. Frivold considers at first elementary magnets, arranged along a
straight line, which are in statistical equilibrium under the influence of
the thermal agitation, their mutual action, and the external field. This
theory is therefore a straight forward extension of Langevin's theory.
The result is indicated in Fig. 27. Curve (1) is that of Langevin, curve
(2) contains the correction of (1) due to the mutual action of the mag-
nets, strongly magnified, because the ordinates of the curve of Langevin
should be increased only by 1/1000 part. The mutual action is of
coiu'se most marked at the lowest temperatures, where saturation might
be reached even in weak magnetic fields.

The problem for three dimensions is carried out in a way analogous to
that of one dimension. The elementary magnets are taken in the
comers of a cubic space lattice, and for a space centered cubical lattice.
The result of the calculation is given by the following two equations for
the two space lattices respectively:

M, ImHT/.
Nm 3kf(l

6.27

(

a<k

t)'+ }

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 195

+

/mHY/2.56. m^ « 1)1
VkT/ I 4 VkTr ""is/J '

M, ImHR 8.88, M« X* \

+

/mHV /3.22/ M» \ 2\1
VkT/ 1 4 \a»kT/ 15/J'

Langevin's formula is simply:

M ^ ImH
Nm "SkT'

a is the distance between two successive magnets. Even here the
mutual action plays a very insignificant role, especially in higher tem-
peratures. The mutual influence appears larger, when the electrical
forces of dipoles are taken into account. The field of ferromagnetic
crystals is fuU of experimental and theoretical problems.

MAGNETONS

The conceptions of magnetons are due to Amp)ere and Wilhelm Weber,
but magnetons as special physical realities were introduced into the
optical sciences by Walter Ritz for the explanation of the Balmer series
of the hydrogen spectrum. If an electron describes a circular orbit
perpendicular to a magnetic field H, with velocity v, it is acted upon by a

2

/"

Fig. 28

force F=»e V H; this force is balanced by the centrifugal force m v*/r,
where r is the radius of the circle. v=« r=2 t n r, n being the fre-
quency. We have therefore evH=mvVri or eH/m = v/r=2Tn.

196 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

The frequency is proportional to the field. Ritz assumed the existence
of elementary magnets like thin, short iron rods in all hydrogen atoms,
of length 1 and pole strength /i- In a distance a from m the magnetic

force Hi=-, while H2 = , or the resultant magnetic force H=Hi

a* (a+)*l

— H2=/Li ( "i^T^Tni)' *^^ ^^^ frequency of the electron:

eH

ni

m2

^ e/i/l 1_\

m2Aa« (a + l)V"

If we assume two magnetons joined together in the same line the
resultant magnetic force will be equal to:

a« (a+21)«'
and the frequency n2:

CM

ni =

CM /I l_\

2TmVa* (a + 21)V'

For 3 magnetons we would obtain:

^• = 2!^W~(ir^^

(a + 31)^

To every line of the spectrum corresponds a system of magnetons.
This explanation of the Balmer series has been replaced by the theory
of Bohr.

More important than this magneton of Ritz, which has only historical
interest, is the magneton of P. Weiss, which is based on a large
number of ph3rsical measurements on paramagnetic substances. Aa
references we mention: P. Weiss: Physikalische Zeitschrift 12, p. 935,
1911; P. Weiss: L'etat actuel de la question du magneton, Bibliothique
Universelle 35, p. 406, 1913. A comprehensive review has recently been
given by B. Cabrera: Magneto-Chimie, in Journal de Chimie phyBique,
16, p. 442, 1918, with a complete bibUography.

DEFINITIONS

LetM— the magnetic moment per gram atom at the temperature T.
Mo "^ the same at the absolute zero.
k« susceptibility per unit volume.
ds= density, m^the molecular weight of the substance.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ jil97

Then for the molecular susceptibility, Xm-

X.=^m=| where H=magnetic field.
For paramagnetic gases we have according to Langevin:

M , 1 ^ MoH

— =ctgha — where a =

M

a

RT

For gases we can write with sufficient approximation:

therefore:

1 a
coth a— " = ",
a 3

M=Mo-=r^:rr,or — r =r^; R= 8.315 10^

^ 3 3 RT H 3 R

XniT =

3R

But Curie found :

Xm T = Ca = constant. (Curie's constant) .

M«

C«=3-g,orMo=V3RC;.

The only paramagnetic gases are Os and NO for which Weiss, and^ Weiss
and Piccard, give the following values:

Xa.lO*

Cm

Mo

n

0,

NO

3381 8
1400.3

0.9937
0.4132

15746
10156

14.014
9.039

0,

NO

3449.6
1461

1.0107
0.4281

15920
10330

14.12
9.20

According to Weiss the magnetic moment per gram atom or the mag-
neton is 1123.5, deduced from magnetic measurements of Fe, Ni,
Co and FeiO* at low temperatures. If we divide 15745 by 1123.5 we
obtain the number of magnetons n = 14.014 per molecule of oxygen.
According to Weiss all the paramagnetic moments are integral multiples
of the magneton 1123.5. Unfortunately Weiss, Bauer and Piccard
(Comptes rendus T 167, p. 484, 1918), found for the number of magnetons
in oxygen 14.12 and in NO 9.20, numbers which deviate considerably
from integers.

198 THEORIES OF MAONETIC CRYSTALS AND MAGNETON: KUNZ

SOLUTIONS

For dilute solutions the same paramagnetic law has been assumed as
in gases. But the solvent is not without influence, in general. There
are cases where the nile of Wiedemann holds and cases where it does not
hold. The rule of Wiedemann states that in the solutions the suscepti-
bility X obeys the law of mixtures. Let the concentration of a salt be
C«, its susceptibility x*; the susceptibiUty of the solvent Xmt ^ts concentra-
tion 1— C«. Then the rule of Wiedemann will be expressed by the
equation:

X = C.Xa+(l-CJx..

Moreover in a compound like FeS04 the negative group SO4 will
have an influence on the magnetic moment of the salt. In the inorganic
compounds we know very Uttle of this influence, but in the organic
compounds of related constitution Pascal found that in many cases the
molecular diamagnetism of a compound is equal to the sum of the
diamagnetism of the atoms plus an additive constant X or in symbols
Xm^2n*X*+X;X is characteristic for the molecular constitution. X^o
only in the saturated carbon-hydrogen compounds C^ H^4.2, so that
the molecular susceptibility is purely additive. For all other compounds
of^the aliphatic series X is positive, for the aromatic compounds X is
negative. Much more compUcated are the relations in the oxygen
compounds, for instance Os and N O are paramagnetic, while C O and
Hs O are diamagnetic. Still more compUcated are the properties of
the inorganic compounds, where we know as yet no additive or analogous
law. Nevertheless Weiss assumes that the diamagnetic atoms maintain
their diamagnetism in the compounds with paramagnetic atoms. In
order to obtam the pure paramagnetism in the paramagnetic salt«, he
uses^the following atomic and molecular coefScients:

-x.X10«

-X.X10«

-XaXlO*

H

C

0

s

Se

Te

P

3.05
6.25
4.8

15.6

24.

39.

27.4

Fl 12
CCl 21
Br 32
I 46.5

Na 4

K 11

Hg 35

S0« 38 5
NO, 19.0
NH, 15.0

CN 11.25
H,0 13.5

Using these corrections Weiss deduced from the measurements of Pascal
the foUowing molecular moments and number of magnetons. Neglect-
ing the fourth and seventh example we obtain almost as good integers
as|, Weiss, by replacing the decimal point one cipher to the left. Un-
fortunately, Weiss had used for the susceptibility of water— 7.6* 10~*

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 199

Substance

Mo

Mo

n' -"integer

1123.5

K and NH4 Ferricyanide

Fe and NU4 Pyrophosphate. .
Fe and NH4 citrate

11700
24600
24680
27100
31500
31390
33800
29200
30480
31370
31120

10.41
21.69
21.96
24.04
28.03
27.93
30.09
25.99
27.11
27.91
27.69

10
22
22
24
28
28
30
26
27
28
28

Na f erripyrophosphate

Na ferrimetaphoq)hate

Ferrichloride

FerrisulDhate

K-f errometaphosphate

Ni^ferrooxalate

Napferropvrophosphate

Ferrosulonate

instead of —7.2* 10"^ This means a considerable error so that these
first numbers of Weiss have only historical interest. But they gave
rise to a careful study of the solutions of paramagnetic substances,
which revealed a number of interesting phenomena, and led in special
cases to integral numbers of magnetons. Not all solutions follow the
rule of Wiedemann, the susceptibility of the substance dissolved is often
a function of the concentration. The salts of Ni follow the law of
Wiedemann and give rise to integral numbers of n. Weiss et Bruins,
and Cabrera, Moles and Guzman give the following results :

Om

n

Nia,

Ni (NO,),
Ni (S0«)

1.300
1.299
1.306

16.03
16.02
16.07

More recently Theodorides (Archives de Geneve 3, 1921) finds for NiClj,
in the temperature mterval 0*'-125**, n = 16.03; and between 150"-300^,
n = 16.92.

Similar are the results of the salts of chromium, Or (N Os)! n ■> 18.99,
and Crs (S O^s, n» 18.99. On the contrary the salts of iron (valence
2) do not seem to give a whole number of magnetons but there exists a
lack of agreement among the measurements, as can be seen from the
following table:

\^m

n

FefSOi 7 H,0

3.400

25.9

3.589

26.6

F6fS0«

3.385

25.9

3.551

26.51

Fe a, 4 H,0

3 478

26 2

FeCl,

3.349

25.7

200 THSORJSS OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

The measurements of solid salts and their solutions do not always agree.
Approximately the number of magnetons in ferro-salts is 26. In the
three valent iron salts in solution Wiedemann's rule does not hold.
Here the influence of hydrolysis appears. Cabrera and Moles give
the following curves for the variation of C^ as function of the concentra-
tion.

These measurements seem to indicate horisontal asymptotes corres-
ponding to 29 magnetons for Fe CI| and 27 magnetons for Fet (SOOt.

Fio. 29

If these curves dependjon hydrolysis, they must undergo a change by
the addition of the cathion H + , which will oppose hydrolysis and increase

o-FcCl,+HCl •-FcCl,+HNOi

Fio. SO

the magnetic susceptibility. These ciu-ves (Fig 30) show the influence of
increasing quantities of HCl and HNOi in a solution of Fe C1|, whose
concentration was 0.00838 gr. per cm*. The curves consist of two

THSOBIBS OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 201

branches. In the first part 0^ increases rapidly by the addition of
small quantities of acid. In the following branch the curve rises much
more slowly. The first branch of the curve corresponds to decreasing
hydrolysis. If we continue the second branch until it cuts the axis of
ordinates, we obtain a value of 0^ for zero hydrolysis. This value
corresponds to 29 magnetons, but imfortunately the limiting values of
Cm for Fe Cls are not quite the same when the hydrolysis is reduced by
HCl and by HNO,.
Moreover, we obtain for C/fr
infinite dilution for Fe Cl|
and Fe (NO|)i 27 mag-
netons; for Fe2 (804)3 26
magnetons. If in complex
salts the paramagnetic
atoms form a part of the
anions, they lose some-
times their paramagnetic
character ;'f or instance, K4
Fe Cy4 is diamagnetic.

Interesting phenomena
appear if paramagnetic iron salts are dissolved in organic solvents, which
have been dried carefully.

They are represented by 3 curves (Figs. 31, 31a, 31b) taken from
Cabrera's review. Sharp maxima appear especially in the solution of

FeCli in GH,OH
Fig. 31

FeCl, in (CiHi),0
Fig. 31a

W^

FeQ, in HCX)OH
Fig. 31b

Fe Cls in (CsH6)20 and in CsHsOH. If we continue the curves until
they cut the axis of prdinates such values of Cm will sometimes appear
as will lead to integral numbers n.

The salts of manganese, cobalt and copper have in general not given
integers n, though in some limiting cases, where the hydrolysis is total
or zero, approximately whole numbers of magnetons appear. That
salts of the diamagnetic copper are paramagnetic is in itself interesting.

a02 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

PARAMAGNETIC SUBSTANCES IN THE SOUD STATE

P. Weifis extended the theory of paramagnetic gases to solid sub-
stances and found from the molecular susceptibilities of a series of com-
pounds the following approximately integral numbers of magnetons.

Fe ClIiNH/cinio* V//^ ','.'.'.'.'. '.','//.

FeF,3NH4F

FeF,2NH4FH,0

Ferriacetylacetonate

l/3Mni04

CrCl,

Cobaltacetylacetonate

1/2 Cr (NH,). Cr (C, OJi 3H, O

1/2 Cr (NH,)4 CO4 (Cr NH,), (C, OOi

28.83
26.99
28.94
29.19
21.23
17.97
20.04
21.12
20.16
20.16

ni

2.88

2.70

2.89

2.919

2.12

1.797

2.004

2.11

2.0126

2.0126

If we again replace the decimal point one C3rpher to the left, we obtain
even better integral numbers n, than Weiss (n). In these salts the
magneton would be 10 times larger than Weiss' magneton. The last
table contains older measurements. Newer measurements of K. Onnes
have shown that often paramagnetic salts with crystal water follow
Curie's law in a wide range of temperature. In most other cases, how-
ever, this law must be replaced by the following form:

x(T-hTi)- Constant,

where Ti is a constant (positive or negative). When liquid oxygen and
nitrogen are mixed together, then Ti depends on the concentration of
oxygen: it approaches zero as the dilution becomes infinite, i.e., if the
molecules of liquid oxygen are separated sufficiently. Curie's law holds
even for dilute solutions. Cans, Keesom and Lenz have given inter-
pretations of the modified Law of Curie. Even in these cases Cabrera
has deduced an integral number of magnetons, for instance, from the
measurements by E. H. Williams of the oxides of the rare earths. For
Didymium he found as many as 52 magnetons.

FERROMAGNETIC METALS AT LOW TEMPERATURES

The magnetism of ferromagnetic bodies increases even below the
temperature of liquid air with decreasing temperature, and reaches the
absolute saturation at the absolute zero:

I. = M.N.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 203

P. Weiss and K. Onnes have measured at 14^ abs. the magnetic moments
per gram atom M. given in the following table.

Mo

n

Nickel

Cobalt

Iron

3370
10042
12360

7417

3.00

8.94

11.002

Magnetite

The magneton per gram atom of Weiss 1123.5 is the common divisor
of iron and nickel. Magnetite does not agree, nor did cobalt because of
the extraordinary hardness of this element, which makes the measure-
ment of the intensity of magnetization I impossible at these low tempera-
tures. O. Bloch however has subjected the Ni-Co alloys to a sys-
tematic study and has measured I of alloys containing up to 70 per cent.
Co, down to liquid air temperatiu'e. The absolute saturation is propor-
tional to the content of cobalt. By extrapolation Bloch found for
Mo 10042 corresponding to 8.94 or 9 magnetons. The alloys of Ni-Fe
and of Co-Fe show a different behavior. If we represent I as a
function of the concentration, we obtain 2 straight lines which intersect
each other in a point corresponding to the compoimds Fe, Ni and
FcaCo. Continuing these straight lines until they cut the axes, we
find the following values of M© and n:

Ni
Fe

FetNi

M<

3450
12450

34390

3.07
11.09

30.6

Co

Fe /
Fe»Co

M<

10080
12355
11232
40544

8.973
10.997

9.998
36.087

In the iron-cobalt alloys the niunber of magnetons changes from 11 to
10. Interesting in itself is the fact that the iron cobalt compound
Fe^Co is the strongest magnetic substance known imtil now: stronger
than pure iron and pure cobalt! How the 36 magnetons distribute
themselves among the iron and the cobalt is unknown. In the neigh-
borhood of the absolute zero Ni contains 3, Fe 11 or 10, and Co.9
magnetons.

MAGNETIC PROPERTIES AT THE CURIE POINT

Finally, the magnetic properties of the Curie point offer a possibility
of determining the magnetic moment per gram atom, and the number
of magnetons. At that temperature, the ferromagnetism disappears,

<^<j.

204 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

but there remains afterwards a small interval of temperaturei in which
considerable magnetism may appear through smiultaneous action of an
external and an internal field. Now for ferromagnetism an internal
magnetic field is characteristic:

H«=NI,orH=NiM,

and Weiss assumed that a ferromagnetic body behaves like a gas, whose
molecules are acted on by an internal and an external field. Then we
have:

M*
M-=3^(H+N,M);

M Mp« N,M ,
H*3RT^ "^ H ^ '

or

(1) M Mq'Nk Mq*

H ^ 3RT^ "SRT'

We had:

^3' RT'

At the Curie point we have T^G, and H"N| M, hence:

3R0 ' 3R

Substituting this expression in (1) we obtain:
'=-('-|)-3^''"^(^-4

"••-c

3R

This last equation was confirmed by P. Weiss and his coworkers for
many ferromagnetic substances. The temperature curve of magnetiza-
tion, however, above 6 is not uniform, but shows different discontinuities,
which separate probably various modifications of iron. In iron the
modifications p and y are well known, but according to the magnetic
measurements the range p must be divided in 2 parts, at 828^, into
fii and /3s. The same holds for Ni. Some of the results obtained are
recorded in the following table due to Cabrera.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 205

Ca

n

Iron fit. . . .

Fe,

6.639
6.536
6.526

36.24
35.94
35.91

Iron/3t

Fe,

4.587
4.560
4.599
4.646

30.12
30.02
30.15
30.30

Iron^i

Fe,

18.306
18.580

60.15
60.60

Iron y

6.605

36.13

Ni^i

Ni

0.3258
0.3234
0.3261
0.3264

8.03
7.99
8.04
8.05

Ni/J,

Ni

0.4033
0.4032
0.4109
0.4033

8.96
8.93
9.03
8.96

n

36.03

30.15

60.37

8.03

8.97

The measurements of Honda and Takagi, and of Terry, do not agree
with those of Weiss, and the previous authors do not find the magneton.

Above the Curie point 6, magnetite behaves strangely: its curve [ -,T )

consisting of 5 straight lines, to which correspond magnetons in the

Method

Ferromagnetism

at low temperature

Curie point: Fe

Fe

paramagnetic solutions

Solid paramagnetic substances.

Fen

Nin

Valence

11

3

101

r 12.0

8.03

10

8.97

20

18

26

16.03

2

27 \
29 /

3

29

27

21

* In cobalt alloys.

ratios : 4 : 5 : 6 : 8 : 10. It looks as if at certain temperatures the magnetic
moment of the molecules changes in definite amounts so that the number
of magnetons increases with increasing temperature. Concluding the
report on the magneton of Weiss we shall collect in a table the number
of magnetons per atom, which have been found by the different methods
indicated.

206 THE0RIB8 OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

The statement holds in general that the number of magnetons per
atom in paramagnetic compounds of an element is larger than in the
metallic ferromagnetic compounds. No relations between Weiss'
magneton and other properties of the atom are known. All these
numbers of the last table raise new questions. While there is a large
amount of experimental evidence in favor of this magneton, its existence
cannot yet be considered as established; deviations from integral num-
bers are not infrequent: for instance, Ph. Theodorides investigated in
the solid state the following salts: M.SO4, COSO4, Fes (SOOi,
CoCU, NiCli, M„Cli, M„0, from 25* to 26* in an interval of WXf.
The sulphates have a negative molecular field (Ti« positive); the
chlorides a positive, (Ti— negative). The sulphates and the chlorides
of Co and Ni give integral numbers of magnetons, while MbCIs and
MqO give fractional numbers n, i.e. no magnetons.

The magnetic susceptibilities of nickel and cobalt chloride solutions
have been measured recently by Miss Laiura Brant.^ The suscepti-
bilities of the salts have been computed from the susceptibilities of the
solutions by application of the Wiedemann law, and the susceptibilities
of the metals obtained by extending the Wiedemann relation to the
salts. The molecular susceptibility of nickel was 0.004423, and of
cobalt 0.01036. These values give 16 magnetons for the nickel atoms
and 24.5 magnetons for the cobalt atoms.

ELEMENTARY MAGNETIC MOMENTS BY J. KUNZ

Before I became acquainted with Weiss' magneton, I applied in

M * N
the year 1910 the equation 9= ° ' or the equivalent equation

3 It

3 r G
m = ---— , to the ferromagnetic substances Fe, Ni, Co, Ft04 and

IN 1,

Heusler alloys. r = 1.36- 10""; 0 = Curie point; N = constant of the
internal magnetic field in the equation Hm^NI; I, = absolute intensity
of saturation; m = magnetic moment of the molecular magnet or the
magneton. If for instance the magnetic moment of iron is m, and if
there are Z magnetons per unit volume at temperature 0**, then:

Z.m=I,

This equation yields Z. If the mass connected with one magneton is
equal to /ipe^ &nd 5 the density at the absolute zero, then Zfi^=6; if
we assume /upe to be the molecule of iron and mh the mass of an atom of
hydrogen, then :

112"^"-

1 Phyuical Renew. Vol. 17, p 678, 1921.

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 207

(The absolute values of the moments of the elementary magnets of iron,
nickel and magnetite, Physical Review, Vol. XXX, p. 359, 1910).
I have assumed that the temperature law of corresponding states holds
down to the absolute zero; an assumption which has been shown to be
not correct. From mh we find the elementary charge e of the electron

by the relation — =9654. The magnetic moments obtained and the

Mh

values of e are collected in the following table:

mlO"

elO»

Fe

5.15
2.02
3.65
66.21
3.55
4.23

1.60
0.90
1.54
1.56
1.54
2.04

FciO*

Ni

Co

Heuder alloy No. 1

Heusler alloy No. 2

The average value of e is 1.53 • 10^*® instead of 1.59 • lO"*®. It had to be

assumed, however, that the elementary magnet of iron contains 2, of

Co, 4, and of Ni, 6, atoms. We have to consider that the ratio of the

densities of nickel and iron is nearly the same as the ratio of the atomic

weights, the number of atoms per unit volume is therefore nearly the

same in both metals, but the magnetic moment of nickel is only about

20 per cent, smaller than that of iron. We should therefore expect

the intensity of magnetization of nickel to be only about 20 per cent.

smaller than that of iron, while in reality the magnetization of iron

is about 3.5 times stronger than that of nickel. The numbers given

involve wide extrapolations and cannot claim high accuracy. Moreover

the modem theory of cr3r8tal lattice does not agree with these older

assumptions.

THE MAGNETON OF BOHR

The atom model of Bohr has been so successful in the explanation of
the line spectrum of hydrogen, that one might expect it would lead also
to an explanation of the magnetic properties. In the original theory of
quanta by Planck, an oscillator could only absorb and emit whole mul-
tiples of the quantum of energy E = hv, while according to Bohr the quan-
tum relation appears in two different ways in the atom. In the first
place the electron falls, while emitting light, from an outer into an inner
stationary orbit, so that the equation, E^— Ee=hi', holds. In the
second place, for the stationary orbits, the moment of momentum,

p=— , holds (here n is an integer). If the electron moves on a circle

with radius a, then:

p»may,

208 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

where m is the mass of the electron, and v the velocity /I'^ft 2 r r,
or:

p = ma*2ir r.

On the other hand the magnetic moment, Mi, of such a revolving electron
is equal to

. . • Pe nhe

1 A—e vTa*=- — =

2m 4Tm

he

whenn = l, Mi = T . This is the magneton of Bohr. Ifwemultiply

Ml with Loschmidt's number, L « 6.06 - 10", then we obtain the magnetic
moment, M, per gram atom :

M = M,L=7-^ = 5584; M,=6.2110-".
4 IT m

5 X 1123.5 «: 5617.5; i.e., Bohr's magneton is about 5 times larger than
Weiss' magneton. The magnetic moments per gram atom ought to
be integral multiples of 5584. We can not yet test this formula, as
measurements of monatomic paramagnetic gases are not yet available.
Perhaps the vapors of alkali metals are paramagnetic. Measurements
of these vapors would be very valuable. Bohr's magneton is only too
large in the case of nickel at lowest temperatures; in all other cases so
far known the paramagnetic elements and their compounds contain
more than 5 Weiss magnetons. If we proceed to the diatomic gases,
then Bohr's older theory gives us no satisfactory idea on the structure
of molecules. According to Bohr's model the molecule of hydrogen
ought to be paramagnetic while this gas is diamagnetic. We can not,
therefore, directly extend the above result to molecules. If we apply it
nevertheless to the only known paramagnetic gases, oxygen, and NO,
then we obtain according to Weiss, Bauer and Piccard (Comptes rendus
T 167, p. 484, 1918) the following nmnbers for the susceptibilities at
20^0:

02:xto= 1.077- 10-^;

NO:xio=0.48710-^.

The magnetic moments per molecule are:

Mo,= 1.587- 10*,
and

Mno= 1.033 10*.

THBORJBS OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 20fr

Hence, according to Weiss, oxygen would contain 14.12 magnetons,
and NO 9.2 magnetons: not integral nmnbers. According to Bohr
we obtain, by division with 5584, for oxygen 2.86 and for NO 1.86
magnetons: numbers which are not satisfactory. Now, W. Pauli Jr.
(Physikalische Zeitschrift 21, p. 615-617, 1920) has suggested that the
formula of Langevin requires a correction, if the paths have to be
quantified, i.e., if the an^e 6 between the axis of the magnetic moment
and the direction of the magnetic field can not assume continuously all
values but has to be restricted so that cos 6 can only assume the rational
values

k

cos^=it->k = l, 2, 3 n,

n

where n is the number of quanta of the moment of momentum, or the
number of magnetons. Then he obtains instead of M© = \/3RC" the
corrected formula:

Mo=)/.

3RC

m

l/2(n+l)(2n+l)
Forn—1,

M,=V3RC„.^^.

For NO we obtain by means of this formula :

Mo -10330: 1,732=6960,

and

5960

5584
For n=2 we obtain:

= 1.066.

Mo=\/3RC„\/l6.

2

This gives for oxygen :

M«= 15870: 2- 739 =5800

and,

5800
5584 ^•"*'

or 4 per cent, deviation from the assumption that oxygen contains two
Bohr magnetons. For NO the deviation amounts to 7 per cent, from

210 THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ

the assumption that this gas contains one Bohr magneton. Paul! has
assumed that the axis of the magnetic moment is perpendicular to the
axis of figure and that for the spatial quantification only the moment of
momentum of the electrons and not the total moment of momentum has
to be used. The magneton thus determined, in spite of the deviationsi
is very probable, as the niunber of magnetons is restricted to one and
two.

THE MAGNETON OF PARSON

No theory explains so far why the electron, moving on a circular or
elliptic orbit, about a nucleus, does not lose the energy by radiation, or
why the orbit remains stationary or free from radiation. In order to
eliminate this question, Parson assumed that the electron is at the same
time a magneton, i.e., the electron consists of a linear circular ring in
which the electricity moves with the velocity of light. It is a closed
circuit without resistance and without radiation. It is almost the
conception by Ampere of stationary currents in the atoms, which give
rise to the magnetic properties. Such a circuit, at rest or moving with
constant velocity, loses no energy through radiation. Parson assiunes
the radius r» 1.5* 10~* cm. The moment of this magneton is equal to

A i^M^xr*— ;

T'

2Tr.
c '

M = ^^i^ « EJ15 = 3.5 . io-i» e.m.u.,
2irr 2

while that of Bohr is equal to 9.21 • IQ-^^ and that of Weiss equal to
1.85 ' 10~~^^ As Parsons' magneton is 38 times greater than the magneton
of Bohr, which almost corresponds to the magnetic moment of NO,
Parsons' magneton appears almost impossible for magnetic reasons.
In order to eliminate this difficulty Parson assumes that this is the
greatest magnetic moment which an atom may assume, and that the
moment of most atoms will be smaller, because the different magnetons
of an atom neutralize each other. Moreover, in the molecules with
several atoms in the solid and liquid state the magnetons of the various
atoms will oppose each other.

This magneton of Parson can not be discovered in the devation of
cathode rays through its magnetic moment. We might expect to
increase the concentration of electrons in a conductor connected to

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 211

earthy by the application of a ipagnetic field. The magnetic work re*
quired to bring up the magneton, would be equal to H M, and the elec^
trie work gained equal to

Ve
2

HM;

V=?^^=4.5- 10-^ volts,

for H»1000 Gauss. A theory of spectral Unes, Roentgen spectra,
fine structure, Stark effect, is not based upon this conception of the
electron. Parson discusses especially chemical questions. The chemi-
cal forces are intimately con-
nected with the magnetic and
electrostatic effects of the
magneton. According to
Parson the magnetons or elec-
trons are imbedded in a
sphere of positive electricity
which is uniformly distrib-
uted, and in which special
elastic forces are in equili-
brium with the repelling
forces of the positive parti-
cles. The positive electricity
is compressible and may con-
dense roimd about the mag-
netons and weaken their elec-
trostatic effects. With these

conceptions Parsons tries to build up a qualitative theory of the chemical
compounds and of the natural system of the elements. An idea, which
has been emphasized later in the considerations of Kossel, Lewis, and
Langmuir, appears here at first, namely, the idea that the most perfect
form of symmetry i^ the distribution of 8 magnetons in the comers of a
cube; which gives a high degree of stability with a minimum of magnetic
energy. Parson has also built a model with 8 coils, in which are made
visible the mutual positions of 8 magnetons. Figure 32 shows the upper
4 coils, which are just equal to the lower 4 coils, so that the model is
diamagnetic. These models correspond to the inert gases, which indeed
are diamagnetic. Parson, in contradiction with modem evidence — as-
sumed already in in helium 8 magnetons, then the niunber of magnetons
increases almost regularly by one imit. Li has therefore 9 magnetons of

Fig. 32

212 THEORIES OP MAGNETIC CRYSTALS AND MAGNETON: KUNZ

which only one has a free magnetic momenti the element is therefore para-
magnetic. The hydrogen atom contains one magneton and is therefore
paramagnetic in the free state, while the molecule is diamagnetic so
that the two magnetons neutralize each other. For iron, nickel, and
cobalt, Parson assumes 32 magnetons while the atomic numbers are

4ViSr

Fio. 33

26, 27, 28 respectively. Parson does not explain why these elements,
which fall in the group 8 with the inert gases, are so strongly magnetic,
nor does he give a sufficient explanation of the periodic change of dia-
magnetic and paramagnetic properties of the chemical elements. In
connection with the theory of cubes Parson makes the interesting

f

THEORIES OF MAGNETIC CRYSTALS AND MAGNETON: KUNZ 213

observation: the compounds HF, HsO, HiN, H4C, are perfect cubes
and are diamagnetic as well as the salts of Li, Na, K, Rb, Cs with F,
Cl| Br, I, and the analogous compounds: CaO, SrO, BaO, NaNOs,
Na^04, KNO,, K,S04.

In connection with Parson's magneton a few remarks may be made
on the relation between magnetism and chemistry. The experimental
facts are l^on, the theory is hardly started. It is well known that a
certain periodicity exists between atomic weights or atomic numbers
and the magnetic susceptibilities of the elements. The curve represent-
ing this relation is of an irr^ular character representing seven distinct
maximal among which that of the iron group is by f ftr the predominating
one. If the sign only of the magnetic properties is taken into account,
one gets the best representation perhaps by the method of the helix due
to B. E. Eknerson, which is given in Fig. 33.

The strongly magnetic groups appear on a diameter, where we find
Fe, Ni, Co, then Pd, Ru, Rh, then Gd, Eu, Sm, then Pt, Ir, Os.
To the right of the diameter D all elements are paramagnetic or ferromag-
netic; to the left of D all elements are diamagnetic with the exception of
tin and oxygen. Tin in some temperature interval is paramagnetic,
[ in another interval diamagnetic. Oxygen is surprisingly paramagnetic.

Its regular diamagnetic properties appear only in some of the organic
I and inorganic compounds.

Moving on the spiral from iron to the right, we meet Mn and Cr,
elements which are paramagnetic, but whose strongly magnetic proper-
ties appear only in some of their alloys such as the Heusler alloys,
manganese-antimony, manganese-tin, CrsO^. On the right hand side
from the ferromagnetic elements, there are paramagnetic elements; on
the left hand side the diamagnetic elements.

The Uterature on the magnetic properties of chemical compounds is

very rich in facts, which however are not yet correlated by a theory.

It is very surprising that some iron compounds are diamagnetic. The

rules discovered by Pascal for organic compounds have already been

f mentioned.

214 MAGNBTOSTBICTJON: WJLUAM8

MAGNETOSTRICTION AND ITS BEARING ON MAGNETIC

THEORIES

Bt 8. R. WlLUAMB

FtofeaBor of Fhsrnoa, Oberiin College

Under the title of magnetostriction may be claasified those mutual
relations which exist between magnetic and mechanical deformations
of ferromagnetic substances. In general, a magnetic field causes a
change in dimensions of such bodies while reciprocally mechanical
deformations produce changes in the magnetic properties. Among
such phenomena may be listed:

1. Joule Effect. (Change in length due to a magnetic field.)

2. Wiedemann Effect. (Twist due to the superposition of longitudinal
and circular magnetic fields.)

3. Longitudinal Currents due to a twist superimposed on a longitudi-
nal field.

4. Longitudinal Magnetization due to a twist superimposed on a
circular field.

5. Volume Change due to a magnetic field.

6. Villari Effect. (Change in magnetic properties due to a mechanical
stress.)

7. Production of Sound due to a magnetic field.

8. Change of Resistance due to a magnetic field.

9. Effect of Magnetic field on Thermo-electric phenomena.

It should be emphasized that all of these phenomena entail a definite
relation between mechanical and magnetic characteristics. To investi-
gate these inter-relations is one of the most important studies of magnetic
phenomena which may be undertaken, for not only is there wrapped up
in such a study the possibility of judging mechanical properties of
substances by their magnetic behavior, which is important for the
industries, but for the particular subject in hand, magnetostriction
has a very sijcnifi'^ant contribution to make to a comprehensive magnetic
theory. Thus far magnetic theories have avoided trying to explain the
magnetostrictive phenomena and as a consequence the newer investi-
gations along this line must have a very definite outlook on magnetic
theories in general. A comprehensive theory of magnetism must explain
the entire range of magnetic phenomena. A glance at the Ust of effects
given above will demonstrate what a real theory of magnetism must
explain. In the interest of clarity a short description of each of the
above experimental facts wiU be given. To a large extent references
will be given to the more recent researches in which a bibliography of
earlier papers may usually be found.

MAONBTOSTRICTION: WILUAMS 215

1. If a rod of iron or steel is subjected to a magnetic field which may
be varied continuously from zero upwards it will be found that the rod
first increases in length and after a certain field strength is attained it
begins to shorten and becomes shorter than in its virgin state. At large
field strengths there appears to be no change in length and the curve
showing the changes in length becomes asymtotic to the field axis.
This is known as the Joule' eGFect. It varies in different ferromagnetic
substances and the characteristic changes in length for different materials
are shown in Fig. 1. Temperature, tension, extranrnus magnetic

.1

Fig. 1

fields,' etc., all have an influence on the changes in length which occur
due to a magnetic field.

2. Wiedemann' foimd that if a rod of ferromagnetic substance is
niagnetized longitudinally and then simultaneously a current is sent
along the rod producing a circular magnetic field, the superposition of
these two magnetic fields causes the two ends of the rod to rotate in
opposite directions. This is often spoken of as a special case of the
Joule effect in that changes in length occurring along the heUcal direction
of the resultant m^netic field produces a twist. This experiment*

> Joule, Pha. Mag. 30, 1847, 76 and 22fi; BidweU, Proe. Roil. Soe., 55, ISM. 228; 56,
1804, 94; Wilti&mB, Phv*. Sat., 34. 1B12. 268.
■WilLiuDB. Ph],t. Rn., 10. 1917. 133.

■ WwdamMUi, BUktridUt. i. 689.

■ WilluiDl, Phv- Ra-. 33, 1911. 281.

216 MAGNETOSTRICTION: WILLIAMS

as usually carried out maintams either the longitudinal or the circular
magnetic field constant while one or the other is varied in some con-
tinuous manner from zero upwards. Inasmuch as this procedure
causes the direction of the resultant field to vary continuously there is
a wide departure from a true Joule effect in which the direction of the
field is always constant. A comparison with the Joule effect should
be made only when both the longitudinal and the circular fields are
varied together so as to keep the resultant field direction constant.

3. When a ferromagnetic rod is magnetized longitudinally in a sole-
noid, a twist imparted to the rod establishes a circular magnetic field in
the rod which may be detected by the electric current which is transiently
set up along the length of the rod while the rod is being twisted.

4. Magnetize a rod circularly by running a current along the same
and if a twist is imparted to the rod in this state a longitudinal magneti-
zation will be produced in the rod which may be detected by the current
induced in the surrounding solenoid. Both of the special phenomena
enumerated in (3) and (4) are reciprocal relations accompanying the
Wiedemann^ effect. In (2), (3) and (4) there are three factors which go
to make up (he effects; a longitudinal magnetization; a circular mag-
netization; and a twist. By establishing any two the third will be
produced.

5. Joule,' the first to observe changes in length magnetically, also
observed that while changes in length were occurring, variations in
dimensions transversely were also taking place. That is, if the rod
increased in length, the dimensions at right angles to the length decreased.
This led to the question as to whether the transverse and longitudinal
changes just compensated each other and gave no change in volume.
Joule was not able to confirm the idea that there was a change in volume
ftccompan3ring magnetization because of the smallness of the effect, but
later on an effect was confirmed by Cantone* in the case of nickel.
Nagaoka and Honda^ are the ones who have really established beyond a
doubt the presence of this effect in all ferromagnetic bodies. In all of
the magnetoj^trictive phenomena each substance has its own peculiar
effect. In this instance cobalt is just the opposite of iron, viz., iron has
its volume increased by magnetization while cobalt is decreased. There
is also a very definite relation between the effect of magnetization on the
change of volume and the change in intensity of magnetization due to
hydrostatic pressure. They appear to be reciprocal relations.

6. Thomson* has pointed out that there are certain reciprocal relations
in magnetism where if the changes in length were known it might be

> Wiedemann, Elek,, 3, 692.
t Joule, loc. cit.

* Cantone, Rendieonii d. R. Acoad. d. Lineei, 6, 1890, 252.

« Nagaoka and Honda, Phil. Mag., 46, 1898, 261; 4, 1902, 46.

• Thomson, AppUcationa of Dynamics to Phsrs. and Chem., p. 47 et aeq., 1888.

MAGNETOSTRICTION: WILLIAMS 217

predicted with certainty the effect which a longitudinal pull or compres-
sion would produce in that same specimen in the way of changing its
magnetic properties. This latter phenomena is known as the Villari^
effect. If a ferromagnetic rod shows an increase in length due to a
magnetic field, that same rod will show an increase in magnetization
when stretched or a decrease in magnetization when compressed longi-
tudinally. If the rod shortens in a magnetic field a corresponding Villari
effect ensues. For substances which show an increase in magnetization
for weak fields and a decrease for strong fields, there is a certain critical
field strength where the intensity is the same whether the rod is stretched
or not. This is know n as the Villari reversal point. Substances showing
a Villari reversal also show a Joule reversal such as Bidwell demonstrated
in iron. The question as to whether the Villari reversal occurs in nickeP
seems now to be pretty well settled in favor of the negative.

7. The magnetostrictive effects have a relationship to the tones
which are emitted by a rod when placed in an alternating magnetic field.
It is a complicated phenomena because if a rod is placed in a periodically
varying magnetic field, the alternating changes in length will not only
give rise to vibrations in the rod, but there is danger of not getting the
rod placed symmetrically in the field and this will give alternating puUs
and thrusts on the rod and so set up vibrations. Bachmetjew's' work
on strongly stretched rods, which gave no tones, would indicate that
magnetostriction was the cause of many of the tones produced. Honda
and Shimizu^ have carried out some interesting experiments on this
subject which indicate very definitely that the tones emitted by a rod in
an alternating magnetic field are largely produced by the change in
length. Maurain^ observed the frequency of tuning forks in a magnetic
field and in various azimuths and found a change in the period. This is
a field which needs further investigation, but changes in dimensions of
the fork must effect some change in the period. Warburg* and St.
Meyer' have also shown that mechanical vibrations affect the magnetic
properties of ferromagnetic substances. This is really a special case of
the Villari effect.

8. A change in resistance due to a magnetic field is a complicated
phenomenon. It seems to be pretty well accepted that it is primarily
related to the Hall effect. However, with change in dimensions which
occur, there should be some change in resistance and if in the process of

> ViUari, Pooa. Ann,, 126, 1868, 87.

•Ewing ftnd Cowan, PhU, Trans., 179, 1888, A. 325: H^dweUer, Wied. Ann,, 52,
1894, 462; 15, 1904, 416: Honda and Shimisu, Ann. d. Phy%„ 14, 1904, 791; 15. 1904. 866:
Williams, Phy; Rec„ 10, 1917. 129.

• Bachmetjew, Rep. d. Phya., 26, 1890, 137.

« Honda and Shimisu, Phil, Mag., 4, 1902, 646.

• Maurain, C. R. 121, 1896. 248.

• Warburg, Pogg, Ann., 139, 1870, 499.

' M^er, Boltsmann Feetschrift, P., 68, 1904.

218 MAGNETOSTRICTION: WILLIAMS

magnetization an orientation of oblate spheroidal particles exisU>, a i

change in resistance similar to that found on rotating spheroidal particles

in an electrolyte^ may be expected, and would augment the change in

resistance more than the mere change in dimensions. For the most

part measurements of the changes in resistance which occur in magnetic

fields have been carried out with powerful electromagnets. It hardly

seems possible to carry out such experiments without getting some real

distortions in the specimens.

9. Thermocouples may be made out of substances which are the same
chemically, but not ph3rsically.' If magnetized and unmagnetized iron i

are used as the elements, a P. D., ca. 1/22 volts, as maximum exists. ^

Bachmetjew' in applying a tension to the iron wire thus magnetized
foimd that the direction of the P. D. could be reversed, thus placing a
partial cause of the phenomenon at the door of magnetostriction. The
results of those who have studied the behavior of thermocouples in a
magnetic field indicate a marked similarity between the character of the
curves showing the change in E.M.F. with magnetic field and the Joule
eflfect.

The above phenomena must be explained by a comprehensive mag-
netic theory. What must be the character of that theory? Preemi-
nently it must be a mechanical theory. On the one hand there are plain
straightforward mechanical effects due to a magnetic field; while on the '

other, by impressing on the same substance a distortion of any character
whatever, a very distinct change in the magnetic qualities obtains.
As Burrows^ expresses it, "Experimental evidence seems to point to
the conclusion that there is one and only one set of mechanical character-
istics corresponding to a given set of magnetic characteristics, and con-
versely there is one and only one set of magnetic characteristics corre-
sponding to a given set of mechanical characteristics."

Po3rnting and Thomson^ have called attention to the fact that the
magnet ostrictive effects are yet to be explained on the molecular hypoth-
esis. They state, " It would obviously require some further assump-
tion as to molecular grouping or as to molecular dimensions in different
directions." This point will be discussed in greater detail later on. ,

If a catalog of magnetic phenomena is made it will be seen how impor-
tant a role magnetostriction plays in studying magnetic theories, for
magnetostriction occupies a large section of the list of magnetic effects.
Not only will the following table show how the magnetostrictive effects
bulk up among other magnetic phenomena but it will be useful in orient-
ing on<)'s thinking in the field of magnetism.

> Williams, Phy%. Rev., 2. 1913, 241.
i Williams, Science, 40, 1914, 606.

* Bachmetjew, Wied. Ann., 43, 1891, 723.

« Burrows. Bid. Bur. Standi., 173, 1916, 13.

* Poynting and Thomson, Blee, and Mag,, p. 201, 1914. "^

MAGNETOSTRICTION: WILLIAMS 219

I. Induction Effects.

1. Relation between field strength and magnetic induction,
permeability, susceptibility, coercive force, retentivity,
hysteresis, etc.

2. Dia-, para- and ferromagneti&m.

3. Terrestrial magnetism.

4. Alternating currents.

5. Inductive effects as influenced by temperature, mechan-
ical strains, ageing, extraneous fields, etc.

6. Relation between susceptibility and chemical properties.

II. Mechanical Effects.

(a) Reaction effects between magnetic fields.

1. Attraction and repulsion of magnetic poles.

2. Motion of electric conductors, (solids, liquids and gases),
carrying currents when placed in a magnetic field.

3. Hall effect and its reciprocal relations.

(b) Magnetostrictive effects.

1. Joule effect. Its reciprocal relations.

2. Villari effect.

3. Wiedemann effect. Its reciprocal relations.

4. Volume change. Its reciprocal relations.

5. Change in resistance due to a magnetic field.

6. Production of sound.

7. Piezo- and Pjrromagnetism.

8. Magnecrystallic action.

9. Effect of magnetic field on thermo electric phenomena.

III Magneto-optical Effects.

1. Faraday effect.

2. Kerr effect.

3. Zeemann effect.

4. Magnetic double refraction.

Naturally one might question some points in this classification.
Certainly changes would be made if more were known about the subject.
Whatever the arrangement of subjects a complete magnetic theory must
explain all of the above phenomena. This is a real task.

Following the suggestions of Ewing in his theories it seems to be pretty
well conceded that induction phenomena are to be explained by the
orientation of something within the ferromagnetic substance which we
may call the elementary magnet. If the rotation of the elementary
mi^nets due to an external field explains ferromagnetism then one may

220 MAONBTOSTRICTION: WILLIAMS

properly ask if the rotation of the elementary magnets might not also
explain the magnetoetrictive efifects, since these efifects appear in ferro-
magnetic substances. Swing's theory and model have been quite
successful in the general field of magnetism. Why should they not be
effective also in magnetostriction? This leads to an emphasis of the
suggestion^ that if dimensions in different directions be ascribed to the
elementary magnets then orientation of such a group of magnets would
give rise to changes in dimensions such as are found in magnetostriction.
It would appear that some such picture must be made of what happens
in a ferromagnetic substance when it is magnetised or else a line of argu-
ment such as that suggested by Borelius* and others must be followed,
viz., that there are two effects present and their combination gives rise
to the results which we obtain in magnetostrictive phenomena. Swing's
work and that of his followers have been on the basis of a specific model
which could be set up and tested in the laboratory.

Another class of theorists is that in which mathematical formulae
have been set up for the mechanical stresses produced by magnetic
fields. Maxwell' deduced the first one which was followed by one of a
more general character by Helmholtz^ in which he emphasized the terms
arising from the change in density of the medium, a fluid being the
medium he had particularly in mind. In 1884 Kirchhoff * gave a formula
which was even more general than those preceding him which included
terms dependent upon the elongation as well as those dependent upon
changes of density. Kirchhoff's theory really confirms the idea from a
mathematical standpoint that when a substance, at least a ferromagnetic
substance, is subjected to a magnetic field, strains are set up in the body
thus placed and give rise to anisotropic susceptibility.

The equations which Kirchhoff set up show the relations which exist
between the intensity of magnetization and the field strength when this
relation is modified by the effects of the strains set up in the magnetized
medium. The equations which define these coefficients are the following :

I.= (k-k' (X.+X^+Xj-k"X,} H.:

I^- {k-k' (X,+Xy+X.)-k"Xy} H^;

I.« (k-k' (X.+Xy+X.)-k"X.} H..

The ordinary relation between I and H is that I » k H. In the above
equations it will be noticed that the terms within the brackets are of

1 Willuuns. Phy. Rett., Abetraot, Feb.. 1911; Phy$, Re9., 34, 1912, 4a Phy$. Ra.,
35, 1912, 282; Poynting uid Thomaon, Elee. and Mag., p. 201, 1914.

* Borelius, Ann, d. Phv§., 58, 1919, 489.

* Maxwell, SUc. and Mag,, p. 257, 2iid. Ed.

* HelmholU, Wied. Ann,, 13, 1881. 386.

» Kirchhoff. SiiAer, d, K, Akad, d. Wi$9, m BtHin, p. 47. 1884.

MAGNETOSTRICTION: WILLIAMS 221

the character of k in the simple equation. If there were no strains set
up then k' and k'' would be zero and the equations of Kirchhoff would
be the usual relation between I and H. k' is the coefficient of mag*
netization introduced because of the changes in volume altering the
magnetization of the specimen. This is the term which Helmlioltz
introduced in his equations and which he thought applied more par-
ticularly to liquid media, k'^ is the coefficient which changes the
intensity of magnetization because of the change in length which has
occurred in the material parallel to the direction of the magnetic field.
Kirchhoff assumed that in applying his formulae, solid elastic media,
free from hysteresis and time lag, were dealt with and that they were
initially isotropic. Sano^ has extended the theory of Kirchhoff some-
what by making all of the coefficients, k, k' and k'^, some function of the
field applied.

When Kirchhoff developed his mathematical theory of magnetostrio*
tion there was little or no data on the values of k' and k^' because these had
to be determined from the changes of voliune and of length which occur
in ferromagnetic substances when subjected to a magnetid field. Par*
ticularly the change in volume due to a magnetic field had not been
definitely detennined. It is to Nagaoka and Honda' and their pupils
that we are indebted for the careful, painstaking work which has put the
theory of Kirchhoff to such rigid tests. Nor should the elaborate
researches of Cantone' in calculating the coefficients of Kirchhoff,
k' and k'^ from the experimental values of the change of volume and of
length in nickel by magnetization be forgotten. Nagaoka and Honda
extended the work of Cantone to various other ferromagnetic substances.
One cannot read the various papers of these two indefatigable investi-
gators without feeling how thorough-going their work is. From the
theory of Kirchhoff, Cantone calculated and fotmd the change in length
and in volume of ovoids to be of the following values respectively:

/N «1 H« /4Tk«,, .,, .k-k^ k"/. .o.A

(b) — =

/ (k-kO k"l

— •{Tk*+3 -— ?;

«)r 4 4/'

V K(l+3

where E is Young's modulus, K the rigidity, and 9 a constant defined
by the relation ;

s Suio, Phy. R€9., 14, 1902, 16S.

• Nasaoka and Honda, PhQ, Mag., 46, 1808. 261; 49. 1900. 329; 4, 1902. 46.

• CantoDA, Mem. d. R. Aeead. dei Lineei, 6, 1890. 487.

MA0NBT08TRICTI0N: WILLIAMS

Id an analogous way Nagaoka and Honda calculated these same changes
for long wires or rods when placed in a uniform magnetizing field and
found them to be:

St
(d) -

H*

V 2K(l+3

2 2(l+2»)/'

«{"'"+l*-''''-f}'

The results of Cantone and Nagaoka and Honda show the volume
change for the ovoid and the extended rod to be the same while the length
change is less for the ovoid than for the rod or wire. Inasmuch as
magnetic fields beget mechanical strains and vice versa, mechanical
strains produce changes in the magnetic properties of ferromagnetic
substances and these effects are reciprocal, Nagaoka and Honda cal-
culated both k' and k'' from the changes in magnetization due to increase
in volume by hydrostatic pressure a, viz.,

(e)

and also from the change in susceptibility due to longitudinal stretching^
X,of alongrod:

(0

«k=|k'|-3(k'+^k'o}x.

From the ovoid Nagaoka and Honda got:

(A)

while from a long rod:

a<)

k'=

i

p(l+2tf)-q
2(l+3ff) '

k"-

3q-p .
2(l+3tf)'

k'=k+

E«

2KH«

(X-(l+2ff)a);

1

k"-4Tk*-

E*
2EH>

(3X-»);

MAGNETOSTRICTION: WILLIAMS 223

where p and q are defined in terms of known quantities and are used here
for brevity's sake. From changes in magnetization and susceptibility

(C)

Young's modulus and coefficent of rigidity were determined in the usual
way. Nagaoka and Honda "found wide quantitative divergences
between the results of experiment and calculations, though in nearly
all cases there was agreement as to quality."^ It is with ''imperfect
success" that mechanical deformations due to magnetic fields may thus
be explained by a theory of magnetic stress and vice versa. One has
only to turn to the careful work of Nagaoka and Honda^ to see how far
Kirchhoff's theory comes from explaining the facts as they exist. They
assert that Kirchhoff's theory is a ''rough approximation and will
perhaps only hold when the strain is infinitely small. We cannot,
therefore, expect that such a theory can explain the relations between
the strains caused by magnetization and the effects of stress on magneti-
zation in all their quaUtative and quantitative details." According to
Eirchhoff the change in magnetization is made proportional to the strain;
experimental data shows k, k' and k'' as fimctions of the strain.

In the case of the change of magnetization produced by the elongation
X of the wire, Nagaoka and Honda calculated the various values of 5 I
and obtained results which agree quaUtatively with those obtained
experimentally in iron and nickel and yet the values of the term (3 k'+
k'O used in the calculation varied 50 per cent, from the observed values
for nickel at low fields.

Computing the change of magnetization, 8 I, due to a decrease of
volume, — <r, the agreement with experiment was fairly good for nickel,,
but there was a big discrepancy for iron and steel.

They applied the theory of Kirchhoff to the Wiedemann effect and
found here a better concurrence between theory and experiment for
nickel than they did for iron. The curves showing the calculated and
observed values for the Wiedemann effect in iron would indicate a
difference for some field strengths greater than 50 per cent. While the
correspondence between calculated and observed values, quaUtatively,
are not to be minimized, yet it must be conceded by all that a theory
which gives variations of over 50 per cent, between calculated and ob*

> Bneyl, Brit., p. 340. 11th, Ed.

> Nagaoka and Honda, Phil. Mag., 46, 1898. 277; 49. 1900. 336; 4. 1902. 66.

:224 MAGNBTOSTRJCTION: WILUAMS

served data leaves much to be desired in the way of coordinating the
phenomena. As Nagaoka and Honda said, "The present state of the
theory of magnetostriction may perhaps be compared with that stage in
history of the theory of magnetism when the intensity of magnetization
was supposed to be simply proportional to the magnetizing force. In
fact the theory is still in its infancy, so that thei-e are ample grounds
for further development and research."

The correlation between magnetostriction and other magnetic and
physical properties must be extended and in all cases should be carried
out as far as possible on the same specimens, as there are no two samples
exactly alike either magnetically or mechanically. Comparisons between
the results of workers using specimens different from those of others are
of Umitcd value.

MAGNETOSTRICTION: QUIMBY 225

THEORIES OF MAGNETOSTRICTION.

Bt S. L. Quiubt
Instructor in Physics, (Columbia University

When a body is placed in a magnetic field the matter in every element
of volume experiences certain forces due to the action of the field upon
the magnetic particles of the body. In addition the forces between
neighboring molecules may undergo considerable change as a result of
the molecular re-orientation which accompanies magnetization. Both
of these effects contribute to produce a deformation of the body which
is known as " Magnetostriction."

The problem of deriving a theoretical relationship between the strength
of the magnetizing field and the consequent strain of the magnetic
medium evidently resolves itself into an evaluation of these forces.
This done, the strains may be calculated by the ordinary procedure of
the theory of elasticity.

We shall first review the theories of magnetostriction of Maxwell and
von Helmholtz. These investigators concerned themselves solely with
an endeavor to evaluate the f orcive per unit voliune acting on the medium
as a whole due to the external magnetic field. Their results, therefore,
will not include the intrinsic stresses arising from the mutual actions of
neighboring molecules.

Both Maxwell and von Helmholtz sought, at the outset, expressions
for the potential of the desired field forces. The point of departure
of the two theories exhibits itself at once in the quite different values
obtained for this quantity.

Maxwell's Theory of Stress in a Magnetized Mediiun.^

If a voliune element, dr, of a magnetic medium be assumed to contain
a large number of elementary magnetic bipoles, then by calculating the
work done in bringing these bipoles from a position where the field is
zero into a field of strength H the magnetic potential energy of the
medium in dr may readily be shown to be.*'

(1) dW--H-Idr.

We seek, now, an expression for the potential function of the mechani-
cal force on dr. Let the matter in dr be shifted from a place where

> Maxwell, Treatise, II, S 639 et seq.
• ibid., S 389.

226 MAQNBTOSTRICTION: QUIMBY

the field is H to one where it is H+5H. The consequent change in
potential energy will be:

(2) »W--H«I«r-I-5H«r.

The first term on the right hand side of equation 2 evidently represents
a decrease in potential energy arising from a change in the internal
configuration of the elementary magnets in dr. It measures the work
done against intermolecular forces of other than magnetic t3rpes opposing
the change in polarization and is stored up as internal energy of the
medium of a purely elastic or thermal character. This part of the
potential energy clearly has nothing to do with the mechanical forces
acting on the medium as a whole. ^

The second term of the right hand member of equation 2 is the change
in potential energy which would have occurred had the magnetization
of the element been held rigid, so that no work oould be done internally.
It therefore represents the work of the mechanical bodily forces on dr,
and:

-h'l:

(3) W'-- / dr / IdH

is the required potential function of these forces.

From equation (3) we may write at once for the external forcive per
unit volume on the magnetized medium:

(4) F=Vh (I-H),

where Vh indicates that V operates upon H alone.

If the medium carries an electric current whose density is J there will
be an additional electromagnetic force JXB on this current/ so that
the total force per unit volume is given by:

(5) F-Vh(IH)+JxB.

Using the relationships:

4tJ-VXH;

VB«=0;

* On the subject of energy in a magnetised medium see Liyens, **Tlie Theory of Eleo-
trieity" Ch. VI and Ch. XI; Lannor, '*The Electrodynamic and Thermal RelationB of
Energy of Magnetisation/' Proe, Roy, 8oc. 71 (1903), p. 229; Langevin, "Magnetiame el
Theorie des Eleotrons," Ann. de Chim. H de Phyt, 5-6 (1905), p. 106.

> Cf., however, p. 61 infra.

MAONBTOSTIUCTION: QUIMBY 227

we may obtain the X-component of F in the form:

(6)

""''Ui^^-l^^ ^i^'^^^i^'^'^} '

with similar expressions for Fy and Fs.
This f orcive Maxwell identified with a stress composed of :

1. A hydrostatic pressure equal to r— IP.

8 T

2. A tension along the line bisecting the angle, c, between the direc-
tions of the magnetic force and the magnetic induction equal to

7- B H cos* €.

3. A pressure along the line bisecting the exterior angle between these
directions equal to -— B H sin* c.

4 T

4. A couple tending to turn every element of the substance in the

plane of the two directions from the direction of magnetic induction to

1
the direction of magnetic force equal to 7— B H sin 2 c.

4t

Thus we may imagine every element of volmne of a nuignetized body
to be strained under the action of the magnetizing field as though it
were acted upon by Maxwell's distribution of stress. In addition to these
strains there will be others arising from the intrinsic magnetic stresses
before mentioned.

The Theory of Magnetostriction of von Helmholtz and Kirchoff.

Von Helmholtz first attacks the problem of investigating the stresses
which arise in a dielectric placed in an electric field when it is in electrical
equilibrimn. Unlike Maxwell, von Helmholtz based his development
upon the presupposition that the expression for the total potential
energy per unit volume in an electrostatic field would constitute the
potential function of the mechanical bodily force on unit volume of the
dielectric medium.^

The electric field is assumed to arise from a continuous volimie dis-
tribution of charge of density p. The work done in bringing this charge
up gradually in the presence of the dielectric is given by either of the
following expressions:

> von Helmholti, Pogg. Ann. 13 (1881), p. 386. YFtM. Ahk„ I, p. 708.

228 MAGNBTOSTEICTION: QUIMBY

w-Jira,,

or:

W- I'Ppdr,

where £« —V ^ and E is the dielectric constant of tiie

Neither of these expression satisfies the imposed condition of electrical

dW
equilibrium, i.e., -— 9^0. By combining the two forms we obtain:

o 1p

(7) W

/{"-.-.»•}

whose variation with respect to ^ is nulL

This expression, von Helmholtz uses as the potential function of the
mechanical forces acting on the medium as a whole.

By variation of W as given by equation (7) von Helmholtz obtains a
forcive which he identifies with Maxwell's electric stress.^ His method
of performing this variation has, however, been criticized by Larmoi*
and Livens,' who show that in its correct analytical form von Helm-
holtz's theory involves the existence of a bodUy forcive on the elements
of the free ether, which could not, therefore, be in equilibriimi.

The source of this discrepancy has already been pointed out. In
von Helmholtz's theory no distinction is drawn between the two fun-
damental constituents, ether and polarizable matter, of the field, which
is regarded as consisting of a single uniform medimn capable of transmit-
ting the electric and magnetic actions in the same manner as an ordinary
elastic solid transmits mechanical forces. His theory would be valid
if there were only one medium under consideration, of which W is the
energy function. But in fact we have to deal with the ether with its
stress and the polarized matter with its reacting mechanical forces, and
there is no means of disentangling from a single energy function such as
that of von Helmholtz the portions of the energy associated with these
different effects.

The formulae of von Helmholtz, which are equivalent to Maxwell's
electric stress and quite different from his magnetic stress, are now ap-

> Maxwell, op. dt., I, p. 159.

• Lannor, "A Dynamical Theory of the EUeotrio and Luminifetoua Medium," Mtl.
Tram,, A 190 (1897), p. 280.

• livena, PAtZ. Mqq., 32 (1916), p. 102.

MAGNETOSTRICTION: QUIMBY 229

plied by him and by Kirchoff^ to account for the mechanical f orcive in a
magnetized medium. Hertz/ using a method similar to that of von
Helmholtz, arrives at expressions identical with his. These writers
include in their formulae subsidiary terms involving the change in K
and M arising from the strains in the medium. More recently Sano' has
extended the method to include the variation of /i with the intensity of
magnetization and has given formulae applicable to crystalline media.^
A general form of the theory of magnetic stress based upon the method of
energy has been developed by Cohn* and further elaborated by Gans* and
Eolacek.' This theory is, however, subject to the same criticism as
that directed by Larmor at von Helmholtz's procedure. This criticism
seems to have been entirely overlooked by the majority of writers on
magnetostriction, with the result that a great part of the theoretical
work on the subject is fundamentally at fault.

Larmor's Application of the Energy Principle.

Larmor* has indicated the way in which the complete results sought
by Kirchoff may be obtained by a different application of the energy
principle. The forcive of von Helmholtz's and Kirchoff's theory is
now regarded as that which would result if the magnetostrictive
deformations arising from the action of the field on the body were
prevented by a constraint.

The procedure may be illustrated by an investigation of the chaagd
in intrinsic length of a bar of magnetic material, caused by its int.o-
duction into a magnetic field. Clamp the bar to its natural length when
at a great distance; then introduce it into the magnetic field sj as to
lie along the lines of force; then imclamp it in such a way that ic .nay ao
as much work as possible in pushing away resistances to its magnetic
elongation; finally remove the undamped bar to a great distance. If
this cycle is performed at uniform temperature, it follows from Camot's
principle that there can be no resultant work done in it. Now the
work done by the magnetic forces in introducing the bar is:

/idH;

that is:

> Kirchoff. Wied. Ann, 24 (1885), p. 52; 25 (1885), p. 601.

s Herts, Wied. Ann. 41 (1890): "Electric Waves," pp. 259-268.

• Sano, Phjft. Re9. 13 (1902), p. 158.
« Phy$. ZeU. 3 (1902), p. 401.

*Cohn, "Dae Electromagnetieche Feld," p. 510.

• Gane, Ann. d. Phya. 13 (1904), p. 634; Bncye. der Math. Wuwntth., 15.
' Kolacek, Ann. d. Phya. 13 (1904), p. 1.

• Lannor, loo. dt., p. 283.

230 MAONETOSTRICTION: QUIMBY

«+Q^ +I^^ H d H

d^
dQ

per unit volume, where c is the magnetic susceptibility which is pre-
sumably a function of the internal longitudinal pressure Q in the bar
and of its intensity of magnetization I. The work done in unclamping
it is I Qi li per unit volume, where U is the intrinsic magnetic elongation
and Qi is the pressure corresponding to the strength Hi of the part of the
field in which it is undamped. This is on the assumption that the bar
is so long that there are no free magnetic poles near together which would
Himiniith Q by their mutual attraction. The work done per unit
volume by the magnetic forces during the removal of the bar is:

— ^
dl/

- /l«+I— IHdH.

The resultant work in the cycle being null, we have:

2 ^ 2 M

where M is Young's elastic modulus. This can only be satisfied if Q is
of the form X H', where X is a constant, and it then gives:

d» _2X
dQ* M'

and the elongation:

1 d « ,^
2dQ '

while the corresponding stress:

^ 2dl"*

The efifect of the variation of the elastic coefficients with magnetisation
may be included by a similar analysis.^

^ Lannor, loo. cit., p. 299.

MAONBTOSTRICTION: QUIMBY 231

This method has been utilized by Heydweiller^ and Houstoun,' who
obtain relationships similar to those of Larmor between the elastic and
magnetic constants of a stretched wire. Their formulae are applied to
experimental data secured by Bensing* and Nagaoka and Honda^ but
the agreement between theory and experiment is far from satisfactory.

The Investigations of J. J. Thomson

J. J. Thomson* has made use of Hamilton's principle to obtain expres-
sions for the reciprocal relationships observed to exist between strain
and magnetization. The first step in the deduction is to set up the
Lagrangian function for a magnetized mediiun in terms of suitable
magnetic and strain coordinates. For the former we may take the
magnetic field H and the intensity of magnetization I. If the compo-
nents parallel to X, Y and Z of the displacement of a volume element of
the mediiun are a, /9 and y, the resulting strains will be:

da - dp d7

d X dy d z

d7 . d/9 , da . d7 dfi . da

dy dz dzdx dx dy

Considering the case of a cylindrical bar of homogeneous isotropic
material whose axis coincides with the axis of X, the Lagrangian function
per unit volume is now written:

(8) l=^AP+HI-i m (e«+P+g»)-in(e»+P+g»-2ef-2eg-2fg),

where n is the coefficient of rigidity and m— n/3 the bulk modulus.
"A" is a function defined by the equation:

;-n.<*''-

This gives

i"-/

Hdl;

> Ann. d. Phy, 12 (1903), p. 602.
s Pha. Mao. 21 (1011), p. 78.

> Ann, d, Phy. 14 (1904), p. 363.
« Pha, Mag. 46 (1898), p. 260.

* J. J. Thomson, "Applications of Dynamics to Physics and Chemistry,*' p. 47.

232 MAGNETOSTRICTION: QUIMBY

80 that the expression for the magnetic energy of equation (8) is seen
to be in accordance with that used by Maxwell.
Now by Hamilton's principle:

a /Ldt-0;

where

L-/ldr.

Equating to zero the variation caused by a smaU displacement in the
material we get for each coordinate an equation of the t3rpe:

¥i -

inaide the body, and:

d e

at the boundary. The first term of (9) represents the mechanical
forcive on the medium which is equivalent to Maxwell's magnetic
stress. We may assume that the strains arising from this term are
known. If we now let e, f , and g, be the strains due to the second term
alone we can derive expressions relating the intensity of magnetization
with the state of strain in the medium. These appear in the form:

^ ^ "^dP \ diAsm-nicIde 3m-n«Idf/'

d£
"^dP

ix

The sign of the effects now under review will evidently depend, for a
given specimen, upon the coefficients dl/de and dl/df. Also, as
these effects are superimposed upon the strains arising from Maxwell's
stress and in many cases are much larger than the latter, the resultant
deformation may be either an extension or a contraction.

A study of the experimental data bearing upon magnetostriction is
reserved for a later section, but it may be remarked that Thomson's
equations are adequate to account qualitatively for all the observed

p ♦

MAGNETOSTRICTION: QUIMBY 23S\

magnetostrictive efifects, including the reciprocal relationship between
torsion and magnetization. So far as the writer is aware no satisfactory
attempt has been made to apply them quantitatively.

This completes our review of the applications of the energy principle
to problems of magnetostriction. Properly used, it is adequate to
give us certain correct relations between the statistical elastic and mag-
netic coordinates of the molecular system. It cannot, however, furnish
us with any idea of the nature of the mechanism which produces the
phenomenon. A complete theory of magnetostriction must start with
the magnetic forces exerted on the moving electrons within the molecule,
and from these, in combination with known intermolecular forces of
other nature, calculate the resulting changes in the molecular configura-
tion. The orderly array presented by the structure of crystals probably
offers the most promising field for effort in this direction.

In the following investigation, due to Larmor, it will be shown that the
aggregate f orcive on the moving electrons within a body arising from the
presence of an impressed magnetic field may be represented by Maxwell's
stress.

Maxwell's Stress on the Electron Theory.^

Considering first the electrons which are in orbital motion within the
molecules, it is shown elsewhere in this report that the force of magnetic
origin acting on an electron is:

F«e(vXB),

where v is the velocity of the electron and e its charge.
The average value of this over the orbit of a single electron is given by :

vXBIdt

c^'

=i/d rxB.

We may replace the line integral by a surface integral over a cap bounded
by the orbit, whence we get:

F=iY[VB(B-n)-(V-B)n]dS

=i^VB(Bn)dS,

eince V-B=0 always. If we sum expressions of this type for all the
slectrons in an element of volume, assuming that V (B* n) is constant

> Cf. Lannor, Pha. Tnau., A 186 (1896), p. 736.

234 MAGNETOSTRICTION: QUIMBY

over the orbit of a ang^ electron^ we find for the force per unit volume
on the electron orfaita:

P-Vb(BI).

If, in addition, there is a conduction current of density J' traversing
the volume element there will be an additional force J'XB giving a
total fordve per unit volume on the medium of:

(11) P-V, (BD+J'XB.

It is to be observed that this f orcive is not wholly external to the
volume element under consideration. A portion of it equal to J'X4 w I
represents the actions of the electrons furnishing the magnetism upon
the free electrons in the element. Similarly the part Vb(4«'I-I)
represents the local reactions between the various electron orbits.
In a calculation of the mechanical force on the mediimi as a whole these
locally compensated forces should be omitted. This procedure is
generally adopted with r^ard to the latter of the two. The part
PX4«'I has, however, with curious inconsistency, been allowed to
remain.^

The X-component of the force given by equation (11) may be expressed
in the form:

Neglecting the ethereal displacement current' the stress system
equivalent to this forcive is seen to be identical with that of Maxwel
except for the addition of the term 2 t'I' to the hydrostatic pressure
constituent. This term represents the stress arising from the mutual
actions of the polarized molecules. The intrinsic stress due to the
interaction of the molecular and finite currents is not so easily separated
out, though it is implictly contained in the above as well as in the Max-
wellian representative stress system. For this reason neither of them
expresses accurately the actual mechanical bodily force of extraneous
origin on a volume element of a magnetized medium carrying finite
electric currents.

I Cf. Livens, "The Theory of Eleotrieity," p. 689: Larmor. Pha, Trant,, A 190 (1897),
p. 361.

• The retention of this term leads to second order effects only. Cf . Livens, op. eit^
p. 698.

ANOVLAR MOMENTUM IN MAGNETISM: BABNBTT 236

THE ANGULAR MOMENTUM OF THE ELEMENTARY

MAGNET

Bt S. J. BAKiraTT

Carnegie Institution of WaohingtoD

1. Ampere's afisumption that the elementary magnet, or magoetOD,
is a permanent whirl of electricity, and Weber's assumption that eleo-
tricity in general, and that of Amp^'s whirls in particular, has maoi
require together that the elementary magnet should possess angular
momentum, unless it is constituted of both positive and negative
electricities rotating in opposite directions. In this case a finite mag-
netic moment might be accompanied with no angular momentum. If
the m^neton has angular momentum, it must exhibit the djniamioftl
properties of a gyroscope. Furthermore, if all the magnetons in a
magnetized body or magnet have ftngiilar momentum in the same diretv
tion, or if the angular momentum in one direction is preponderant, t he
whole magnet must behave like a gyroscope. Similarly, a coil of wire
ti^versed by an electjic current consisting of a stream of one kind of
electrons only, or with the b'near momentum of one kind preponderant,
must have angular momentum.

»---' — I'

2. In Fig. 1 n shown a modification of a common type of gyroscope,
whose wheel, pivoted in a ring, can be rotated rapidly about ito axis A.
Except for the action of two springs, the ring and the axis A are free

236

ANGULAR MOMENTUM IN MAGNETISM: BARNBTT

to move in altitude about a horizontal axis B, the axifl A making an
angle $ with the vertical C; while the axis B, together with the wheel
and the framework supporting it and the springs, can be rotated about
the vertical axis C. If the wheel is spun about the axis A, and the
instrument then rotated about the vertical C slowly, so that the cen-
trifugal torque is negligible, the wheel tips up or down so as to make
the direction of its rotation coincide more nearly with the direction of
the impressed rotation about C. The greater the rotary speed about C
the greater is the tip of the wheel; it would tip until the axes A and C
became coincident if it were not for the springs (centrifugal torque
being still supposed negligible).

3. If in apparatus of the kind just described we replace the wheel,
or the wheel and its supporting framework complete, by a magnet,
coil of wire traversed by an electric current, or

electromagnet, with its axis along the axis AA
and its center of mass on the axis B, the body, if
it possesses angular momentum, should behave
Uke the gyroscope of section 2. The centrifugal
torque will not, however, in general be negligi-
ble, especially as great speeds about the vertical
will be necessary to make the gyroscopic effects
appreciable when the angular momentimi due
to the magnetons or electric current is small.
(Mechanical disturbances due to the two springs
may be made negligible). The centrifugal torque
may be altered at will, or made to vanish by
adjusting suitable weights attached to the body
along an axis CC intersecting A and B nomudly.
See Fig. 2.

4. The first attempt to detect by direct experiment the angular
momentum of electricity flowing in a coil of wireror the angular momen-
tum associated with the Amp^ian currents in magnetic matter, was
made by Maxwell^ in 1861, with apparatus somewhat similar to that of
Fig. 2, but with the springs removed. Maxwell did not use a per-
manent magnet, but a symmetrical coil of wire traversed by an electric
current and provided at will with a core of iron.

Let A, B, C denote the moments of inertia of the magnet (or cofl)
about its own axis, the horizontal axis B, and the central axis normal
to the two, respectively.

Let $ denote the angle between the axis A of the magnet and the
vertical C, 0 the impressed angular velocity about the vertical, J the

Fig. 2

^ Maxwell's Electricity and Maghetasm, § 676.

ANGULAR MOMENTUM IN MAGNETISM: BARNETT

237

total angular momentum under investigation, and fi the angle between
J and the axis A. See Fig. 3.

Let us suppose that under the action of the springs, producing a
torque T in the direction of increase of 6, the angular velocity Q and the
angle B are maintained constant. J can be resolved into two rectangular
components: one parallel to the axis of the impressed rotation, J cos
(0—0), which is constant; and one perpendicidar to tliis axis, J sin
(9-/9), which has the constant rate of
change 0 J sin (9-/9). By the second .

law of motion this is equal to the JstNfO-ff

torque T; thus:

(1)

Now

(2)
and

(3)
Hence:

T=0Jsin(^-j9).

J cos fi—A Q cos 0+Mf

J sin /SaCfisin 0.

FiQ. 3

(4)

T= (A-C) tf sin ^ cos ^+M 0 sin ^.

If C is somewhat greater than A the applied torque T necessary to
maintain the motion constant will vanish and equilibrium (under the
action of the centrifugal torque) will be stable with the springs removed
when

(5)

cos^=

M

(C-A) 0

By means of two nuts acting on screws the axis CC was adjusted to
be a principal axis with C just exceeding A, so as to make the instrument
very sensitive. On account of disturbances due to the earth's magnetic
field the results were very rough, but no change in 9 with reversal of
M or Q could be detected even when an iron core was inserted in the coil.

Maxwell concludes that if a magnet contains matter in motion the
angular momentmn of the rotation must be very small in comparison
with any quantity which we can measure.

By calculating M as the product of the constant R, determined below,
and the magnetic moment of the magnet, and by taking account of the

238 ANGULAR MOMENTUM IN MA0NBTI8M: BARNBTT

fact that equation (5) holds only if the horizontal axis about which
rotation is possible passes exactly through the center of gravity of the
magnet, W. J. de Haas and G. L. de Haas-Lorentz^ have shown that a
change in $ would hardly be perceptible even in very favorable con-
ditions.

By making A— C very small and $ very nearly 90^, and measuring
T and fi, M could also be determined from equation (4).

5. Let M now denote the angular momentum of a magneton, |i its
magnetic moment, and

(6) R.^

the ratio of the first to the second.

Let us assume that in the experiments of section 4 we have a perma-
nent magnet synmietricaUy magnetized about the geometric bjos, and
that the magnetons are all alike. If 0 denotes the angle between the
axis of a magneton and the intensity of magnetization I in its neigh-
orhood, we have

(7) I»2mcos9,

the summation extending over the unit of volume. The internal
angular momentum j per unit volunae will then be:

(8) j >=2; M cos e^R 2 M cos 9»R L

Thus if I denotes the mean intensity of magnetization along the axis of
the magnet, V its volume, and I V its magnetic moment, its total internal
angular momentum will be R I V. Thus R can be determined from
(4) and (5).

6. Maxwell's experiment was an experiment on the angular momentum
of a gross magnet, and it does not seem to have occurred to him to make
an experiment in which every one of the countless multitude of magnetons
in a magnetic body should simultaneously replace his magnet, and to
measure the total change in the orientation of all the magnetons by a
magnetic method.

The first experiment based upon this idea appears to be one made
nearly a third of a century ago by John Perry*, who tried, but without

> W. J. de Haas and G. L. de HasB-Lorents, K. Ahad. AmMUrdam, Proe. 19, 1916 •
p. 248.
s John Ferry, Spinning Tope, October, 1890, footnote, p. 112.

ANGULAR MOMENTUM IN MAGNETISM: BARNETT 239

Buccess, to detect a change in the magnetization of an iron rod produced
by rotating it about its axis. In 1912 and 1915, respectively, Schuster^
and Einstein and de Haas' had the same idea. In the meantime, in
1909, the same idea occiured to the author', who, with the help of L. J.
H. Bametty then began experiments which were first successful in 1914.
These proved to be the first successful experiments in this whole field.^
The quahtative theory of the phenomenon with which these experi-
ments are concerned is as follows. When the body of which it is a part
is set into rotation about any axis, the magneton, if it has angular momen-
tum, will behave much Uke the wheel of section 2, and will thus change
its orientation so as to make its direction of rotation coincide more
nearly with the direction of impressed rotation; the coincidence will
finally become exact if not prevented by the action of the rest of the body.
In an ordinary ferromagnetic body in the usual state with which we
are familiar only a slight change of orientation can occur on account of
the forces due to adjacent molecules, which perform the function of the
springs in the case of our gyroscope. The rotation causes each molecule
to contribute a minute angular momentum, and thus also a minute
magnetic moment, parallel to the axis of impressed rotation; and thus
the body, whose magnetons originally pointed in all directions equally,
becomes magnetized along the axis of impressed rotation.

If the revolving electrons, or rotating magnetons, are aU positive,
the body will thus become magnetized in the direction in which it would
be magnetized by an electric current flowing around it in the direction
of the angular velocity imparted to it. If they are all negative, or if the
action on the negative magnetons is preponderant, it will be magnetized

in the opposite direction. This is what actually
happens.

7. To develop a quantitative theory', assume the
magneton to consist of a symmetrical electrical
system rotating with angular velocity <a, magnetic
moment Mi ai^d angular momentimi M»R /i about
the axis of symmetry, the electrical charge in rota-
tion being all of one sign. Fig. 4 illustrates a special
case.

The vectors representing M and m are in the same

direction or opposite directions according as the

moving charge is positive or negative.

Let A denote the moment of inertia of the magne-

FiG. 4 ton about its axis of rotation, and suppose B = C the

^ A. Schuster, Phya. Soc. Lond. Proc, 24, 1911-12, p. 121.

> A. Einstein and W. J. de Haas, Verh, d, D. Phya. Gea. 17, 1915, p. 152.

* S. J. Bamett, Science, 30, 1909, p. 413
^ 8. J. Bamett, Phya. Reo. 6, 1915, p. 239.

* See S. J. Bamett, 1. c. ante and Einstein and de Haas, 1. c. ante.

I

240 ANGULAR MOMENTUM IN MAGNETISM: BARNBTT

i

4

(mean) moment of inertia about any central axis normal to the asda S

of symmetry. |.

If now the body of which this magneton is a part is set into rotation ||

with angular velocity 0 about jin axis C, the magnetoui behaving like ^

the wheel of a gyroscope, will strivei as it were, to take up a position
with its axis of revolution coincident with that of the impressed rotation;
but it will be prevented from turning so far by a torque T due to the
action of the rest of the body and brought into existence by the dis-
placement. In a minute time kinetic equilibrium will be reached, and
the axis of the magneton will then continuously trace out a cone fnitlnng
a constant angle 6 with a line through its center parallel to the axis C
of the impressed rotation. When this state has been reached T will
be given by equation (4) above, which may be written:

(9) T» JM 0+(A-C) ff cos ^j sin ^.

Now imagine the body, instead of being rotated, to be placed in a
uniform magnetic field whose intensity H is directed along the previour
axis of rotation, and consider a magneton whose magnetic axis, aftes
displacement by the field, makes the angle d with H. The magneton
would keep on turning under the action of the field untU its axis coin-
cided with H, but IB prevented from doing so by the torque T' upon it
due to the action of the rest of the body and brought into exist^ice by
the displacement. This torque is weU known to be:

(10) T'=MHsinft

To find, therefore, the magnetic intensity which would produce the
same effect on the orientation of the magneton as would be produced by
rotating the body at the angular velocity Q, all we have to do is to equate
TandT'. This gives:

(11) M H sin 9» JM 0+(A- C) 0> cos 9 j sin 9;
or:

/tON TT MO/, , (A-C)O A

(12) H= (1+^^ — . ^ cos g 1.

M \ A« /

The values of 0 experimentally attainable are so small in comparison
with any possible values of fi in the case of any magneton of probable

ANGULAR MOMENTUM IN MAQNBTISM: BARNBTT 241

type that the last term is negligible. Henoe we have for any magneton
in the body, whatever its orientation, with all sufficient exactness :

(13) H = — = R0=2tRN,

where N is the impressed angular velocity in revolutions per second.

From what precedes it follows that if aU the magnetons in a body are
alikCf rotating it at an angular velocity of Nr.p.s. will produce the
same intensity of magnetization in it as placing it in a field of strength
2 V R N Gausses.

// the magnetons in a body are of two kinds, positive and negative,,
with constants Ri and Rs, rotating the body will have the same efifect as
if a magnetic intensity Hi » Ri Q were applied to the positive magnetons
and an intensity Hs=Rsfi were applied to the negative magnetons.
If the effect on the negative magnetons is preponderant, the rotation
will thus produce an intensity of magnetization in the direction of Ht
but of magnitude less than that which would be produced by the inten-
sity Ri fi if all the magnetons were negative.

8. The ratio R will now be determined for three types of magnetons.
(1) Suppose the magneton (Fig. 4), to consist of n similar electrons, all
positive or all negative, with total charge n e revolving in a circular orbit
of radius r, with constant angular velocity co (and areal velocity a^
1/2 0) r*) about a much more massive, and fixed, nucleus with charge
— n e. If the radius of the orbit is great in comparison with that of an
electron, no appreciable error in calculating the angular momentum will
be made if we assume the (electromagnetic) mass and the charge of
each electron both to be concentrated. In this case, if m denotes the
mass of one electron, we have

(14) /i=nea; A=nmr*,

(15) M=A€i)»nmr'€i)=2nma,
and

(16) R=^=.2??.

M e

If the distance between successive electrons in the orbit is great in
comparison with the diameter of an electron, m will be practically
identical with the mass of the free electron. Otherwise it will be
greater. If the diameter of the orbit is reduced, the electromagnetio

^42 ANGULAR MOMENTUM IN MAGNETISM: BABNETT

field being restricted more nearly to the space between each dectrcm
and the nucleus, the angular momentum will become smaDer, as sug-
gested to me by Dr. Tobnan.

(2) Voigt^ has examined the behavior in a magnetic field of mag-
netic elements, or magnetons, consisting of hcHnogeneous uniformly
charged solids in rotation. No account is taken <rf the electromagnetic
origin <rf the mass, but the mass density is taken as everywhere propor-
tional to the electric density. For this type of magneton it is easy to
show that, as in the case of an electron ring,

(17) R»2??.

6

(3) Abraham* has considered the behavior in a magnetic field of a
spherical electron in rotation and uniformly charged either over the
surface or throughout the volume, and has calculated the angular
momentum on the assumption that the mass and momentum are purely
electromagnetic.

The electron, or magneton, nuisses, m^ and m^ for surface and volume
charge, respectively, are

/to\ 2e* , 4e«

oa oa

if e denotes the charge of the electron and a its radius.

For the angular velocity (a radians per second the ocNrresponding
angular momenta are

(19) Ms= ' m« a' w and M^^- m^ a' w;
and the corresponding magnetic moments are

(20) Mi=o e a* « and Mt=* c © ^ «•

Thus the corresponding values of the ratio R » —are

(21) R,-— 'andR,=J^.

e 7 e

^ W. Voigt. Ann, der Phya. 9. 1902, p. 130.

* M. Abraham. Ann. der Phys. 10, 1903, pp. 151, 169, 171.

1

L

ANGULAR MOMENTUM IN MAGNETISM: BARNETT 245

The first of these is just one-half the value for an electron orbit, and
the second somewhat smaller.

9. In very weak fields ferromagnetic bodies all receive magnetic
moments proportional to the intensities of the fields applied. Similarly,
since R 0 is equivalent to a very small value of H for even the greatest
speeds practicable, these bodies must be magnetized by rotation propor-
tionaUy to the speed.

If, however, in either case we start with a ferromagnetic substance
not in or near the neutral state but at the steep portion of the magneti-
zation ciuire, the application of either a small increment of magnetic
intensity or of a small speed may be sufiScient to produce a considerable
and irreversible change in magnetization. Proportionality exists only
for elastic displacements.

10. An interesting case of the theory of magnetization by rotation
is its application to a single magneton or a swarm of magnetons, whose
behavior in an ordinary magnetic field has been considered by Voigt^
and Abraham.*

Consider first the case of Voigt's magneton, a solid of revolution with
electric and mass densities proportional, placed in a uniform magnetic
field with intensity H, and permanently endowed with an angular
velocity o) about its unique axis greater than H/R (=H-^ 2m/e, on
Voigt's assumption).

In this case, in accordance with classical electromagnetic theory,
the rotation proceeds imdamped about the unique axis, while it is
damped about the other (equal) axes, and the action of the field on the
magneton is as follows: When the field is applied, precession of the
magneton's axis about the direction of the field, accompanied by nuta-
tion, begins, with angular velocity H/R. The nutation is damped out
by dissipation or radiation, and the precession is retarded for the same
reason. Hence the direction of the axis of the magneton gradually
approaches coincidence with the direction of the field, when it is in
stable equilibrimn. During this process the velocity of rotation dimin-
ishes sli^tly, the motion being affected as in the case of the electricity
in Weber's molecular grooves.

If there are N such magnetons in the unit of volume, and if the
demagnetizing and molecular fields and the upsetting effect of collisions
are negligible, all the magnetons will ultimately become oriented with
their axes in the direction of the magnetic field. In this case the moment
of unit voliune will be

« -¥("-g)^

^ W. Voigt, 1. c. ante.
' M. Abraham, I. c. ante.

1244 ANGULAR MOMENTUM IN MAGNETISM: BAENETT

or

(23) i-NAe(«-;_H\
^ ' 2in \ 2 m/*

where A is the moment of inertia of the magneton about the unique axis
of p^manent rotation.

The first and principal term is entirely independent of H. The
change of orientation is produced by the magnetic field, but only the
time taken to arrive at the steady state is affected by its magnitude.
The second term is a diamagnetic term, and arises from the fact that
owing to the change of extraneous flux through the magneton during
the process of its orientation its velocity is decreased, just as in the
ease of the Weber-Langevin theory.

In this case we have, except for the small diamagnetic term, which
vanishes with the intensity, saturation for even the weakest fields; and
we have lera nearly complete saturation for stronger fields.

When collisions are not absent, a magneton's axis will be repeatedly
deflected in its approach toward coincidence with the direction of the
field, and the intensity of magnetisation will not reach saturation; but
it will increase with the field strength, being greater for a given field
strength the greater the mean time between collisions and the weaker
the molecular and demagnetizing fields. Increase of temperature,
shortening this time between collisions, and increasing their violence,
will, if the magnetons remain unchanged, thus diminish the magnetisa-
tion for a given field strength.

The precessional process described above is doubtless similar in a
general ^ay to the process by which in every case in paramagnetic and
ferromagnetic substances the magnetons are aligned more or less com-
pletely with the magnetic field.

The behavior of Abraham's spherical magneton is very similar to
that of Voigt's magneton. Although the moments of inertia about
three rectangular axes are identical, motion about any axis normal to
the unique axis of permanent rotation changes the magnetic field and
thus radiates or dissipates energy, while rotation about the unique
axis leaves the field fixed and unidtered.

The angular velocity of the precession of the magneton's axis about
the direction of the field is given for both magnetons by Q~H/R, which
becomes in the case of Abraham's magneton

(24) p,=H^5?

e

lor surface charge, and

ANGULAR MOMENTUM IN MAGNETISM: BARNETT 246

(25) a,=H-^|^'

for volume charge. Abraham has referred to this difference as a possi-
ble mode of descriminating between the validities of the two h3rpothe8es
as to the distribution of the charge.

If the magnetons are subjected to an angular velocity 12 instead of to
a magnetic field with intensity H, we have, when the effects of collisions
and the molecular and demagnetizing fields are negligible,

(26) I«NA/R(«+fi).

The first and only important term is independent of 12. Here the orien-
tation is produced by the velocity impressed, but only the time taken to
reach the steady state is affected by its magnitude.

The second term, here added to the first-, corresponds to Voigt's
diamagnetic term. We have, except for the small second term, satura-
tion for even very small values of 12.

If collisions are not absent, or if the magnetic fields of adjacent mole-
cules and the demagnetizing field become appreciable, the intensity of
magnetization will not reach saturation; but it will increase with 12,
being greater for a given value of 12 the greater the mean interval between
collisions, the less their violence, and the weaker the field.^

11. In studying magnetization by rotation experimentally and thereby
determining the ratio R in both sign and magnitud3 two general types
of methods may be used, viz., methods of electromagnetic induction
and magnetometer methods.

In all methods the substance under investigation is in the form of an
approximately cylindrical rod, which is mounted with its axis horizontal
and in the magnetic prime vertical in a region in which the earth's
magnetic field has been neutralized.

In the methods of electromagnetic induction the intrinsic magnetic
intensity of rotation is determined by comparing the change of flux
through the rod under investigation, produced by rotation about its
axis at measured speed, with the change of flux produced through the
same rod by the application, parallel to the axis of rotation, of a uniform
magnetic field of known intensity. The changes of flux are proportional
to the intensities, if small.

In the only method of this kind which has hitherto been used the
changes of flux are measured ballistically, with a galvanometer of the
t3rpe which has come to be known as a fiuxmeter, a coil of wire surroimd-
ing the rod being in the fiuxmeter circuit.

In another method,' which has not yet been used, the rod may be

^ See S. J. Barnett, Nature vol. 107, p. 8, 1921.

s Suggested by an inveetigation of Tolman's in a different field.

246 ANGULAR MOMENTUM IN MAGNETISM: BABNBTT

ofldDated about its axis instead of being aet into continuous
tional rotation; and the alternating electromotive force thus developed
in the surrounding coQ of wire may be compared with the electromotive
force produced by an alternating field of known intensity, an amplifier
being used to increase the sensibility if necessary.

In the magnetometer method, an astatic magnetometer is mounted
with the center of its lower magnet system in the polar axis, or, preferably,
in the equatorial plane, of the rod, or rotor, under investigation. The
magnetometer deflection produced by reversing the direction of rotation
of tiie rotor, driven at a measured speed, is compared with the deflection
produced by reversing a known magnetic field in the rotor parallel to
its axis. The deflections are proportional to the changes in the mag-
netic moments in the two cases, and these are proportional to the
intrinsic intensity of rotation and the standardizing field intensity.

12. In section 5 it has been shown that if but one type of magneton is
present the angular momentiun of the magnetons per unit voliune in a
magnetised substance is

(8) bis i»RI;

an equation first developed by O. W. Richardson^ for the special case
of electron-orbit magnetons, for which R = 2 — .

If the frame holding the g3rro8Cope of Fig. 1 is forcibly turned about
the axis B, thus altering the component of the wheel's angular momen-
tum about the vertical, there will be a reaction in the frame producing
an equal change of momentimi in the opposite direction.

In the same way the change in the angular momentimi of the mag-
netons produced by a change in the magnetization will be accompanied by
an equal and opposite change of momentimi. Probably the reaction
measured by the rate of increase of this momentum will be upon the
remaining part of the molecules of which the magnetons are constituents;
that is, the total reaction will be exerted upon the rod. But this is not
certain; a part of the reaction may be upon the magnetizing coil. The
probability of this, however, is small; for the rotation of a symmetrical
coil about its axis in no way affects its magnetic field.^ If the rod is
magnetized from a neutral state the total angular momentum acquired
by rod and coil together wiU be:

-jV=-RIV
where V is the volume of the rod.

^ O. W. Richardson, Phya. Rev. 26, 1908, p. 248. For discuflBioiis of the ease in which
more than one kind of magneton is present. Richardson's paper and also that by J. Q.
Stewart Q. c. infra) may be consulted.

' A second order effect on the field, which in general occurs, is here neglected.

ANGULAR MOMENTUM IN MAGNETISM: BARNETT 247

If the magnetons are aU positive, the vector representing the angular
momentum of the reaction will be opposite to that representing the
intensity of magnetization; if they are all negative, the two vectors will
have the same direction.

If two t3rpe8 of magneton are present, we shall have :

(27) j =ix+j,=Ri Z Ml cos ^i+R, Z Mt cos fl,=Ri h+R^ I,.

13. Two general methods of investigating this effect experimentally
have been tried. In the ballistic method, proposed and tried by O. W.
Richardson^ to whom the idea of rotation by magnetization is due, a
cylindrical rod of the substance under investigation is suspended with
its axis vertical by an elastic fibre, and the total angular momentum
imparted to the rod is determined from the throw occurring when a
known change in the rod's magnetization is produced by altering
a vertical magnetic field.

In the method of resonance, first applied to this problem by Einstein
and de Haas,' a system similar to that just mentioned, which has a
definite frequency Uo of vibration about its vertical axis, is magnetized
by an alternating field whose frequency n is variable over a range
including the frequency of the vibrating system. The angle a through
which the cylinder vibrates is determined as a function of the frequency
of the impressed field. From the maximum value a^ of the angle at
resonance, the maximum magnetic moment A of the rod, which is kept
saturated during nearly the whole of each half cycle, or from a^, a, A,
n— Uo, and the moment of inertia of the rod about its axis, the magnitude
of R can be determined. And from the phase relation between the
deflection of the rod and the magnetizing current, the sign of R can be
determined.

A modification of this method, in which the magnetizing solenoid is
wound directly on the rod itself and in which the reversals of the cur-
rent are produced by the motion of the electromagnet thus formed,
has been used by de Haas.' It has the advantage of doing away with
two of the experimental difi&culties, though it introduces others; it has
also the great theoretical advantage that the total torque upon both
rod and solenoid is measured. Only a few rough experiments have
hitherto been made by this method.

Another modification has been used for a lecture experiment by Ein-
stein,^ but only rough experiments have been made. In this method

^ O. W. Riehardflon, 1. e. ante.
' A. Einstein and W. J. de Haas, 1. c. ante.

* W. J. de Haaa, K. Akad. Amsterdam Proe. 18, 1016, p 1280; verh. d. D., Phys. Om,
18. 1016, p. 423.

« A. Einstein, Verh. d, D, Phys, Oea 18, 1016, p. 173.

248 ANGULAR MOMENTUM IN MAGNETISM: BABNBTT

•n ingemoufl device is used for applying the magnetizing field for only
a small fraction of the period, and thus greatly reducing certain extra-
neous disturbances.

14. As stated in section 6, the first successful experiments in the
field to which this article is devoted were made by L. J. H. Bamett and
myself in 1914/ in November and December of which year they were
described before the Ohio Academy of Sciences and the American
Physical Society. These were experiments on the magnetisation of
cold rolled steel by rotation and were made by the fluxmeter method of
electromagnetic induction. They confirmed equation (14) as to pro-
portionality of the intrinsic intensity of rotation with speed, showed
that the role of negative electricity in magnetism was at least pre-
ponderant, and gave for 2 t R, on the assumption of only one type of

2m
magneton, the value — 3.6X10~', or, for R, about one-half of — for

e

the negative electron in slow motion, viz., — 1.13X10~^=— 7.1X
10-V2 T.

In the early part of 1915 we completed another and more thorough
vivestigation of steel by the same method, with considerable improve-
ments. This gave for 2 t R the value — S.lftX 10"^

In 1917^ we completed an investigation of steel, cobalt, and nickd
by a magnetometer method, and obtained values of R which were, as
before, all negative, and whose means were intermediate between the
values previously obtained for steel and twice those values.

We have since' made by improved magnetometer methods very
extensive and conclusive experiments on steel, soft iron, cobalt, nickel,
and Heusler alloy, and have obtained for all these substances values
of R similar to those of 1914 and 1915 for steel, viz., the value of

— for the negative electron in slow motion, within the limits of the exper-
e

imental error. In the course of this work we have discovered a sys-
tematic error in the work of 1917, which made the values of R too high.
These experiments must be taken to prove in a direct and conclusive
way, on the basis of classical dynamics alone, (1) that Ampireian cur-
rents, or molecular or electronic currents of >dectricity in orbital revolu-
tion or in rotation, exist in the ferroma^etic substances; (2) that
the role of negative electricity in ferromagnetism is at least preponderant;
(3) that the magnetons have inertia and angular momentiun, so that

^ fl. J. Bamett, 1. c. ante.

> S. J. Bamett, Phys, Rev. 10, 1917, p. 7.

' For brief deeeriptions of the earlier part of the more recent work see S. J. Bametti
Froe. Phil, Soe. Wa^inaUm for Oct. 9. 1920 {Jour. Wath, Acad. 8ei. 11, 1931, p. 102);
and 8. J. Bamett and L. J. H. Bamett, Proc. American Phye. Soe. for Deo. 1920 (i^v*
Rev, 17. 1921. p. 404). For the latest work see /Voc. Ameriean Phye. Soe. for April, 1923.

ANGULAR MOMENTUM IN MAGNETISM: BARNETT 249

each behaves like a minute gyroscope. Furthermore; if we admit the
classical theory of radiation, these experiments, together with the
exitence of permanent magnetism, prove (4) that the arrangement of
th^lectricity in the magnetons is continuous or Satumian, rather than
planetary. Finally (5) the value obtained for R shows either that
negative magnetons, such as that of Abraham, with a value of R much
less than that for an orbital ring are responsible for magnetism, or else
that positive electrons or magnetons, whose rotation produces an oppo-
site eSectf are also involved. It is suggestive that the value of R for
the superficially charged electron of Abraham and Lorentz (section 8,
(3)), when rotating as a magneton, is equal to those found by ex-
periment on the assumption that all the magnetons in a substance
are alike. The investigation has established, within the limits of the
experimental error, a new and apparently universal constant in magnet-
ism.

15. The first experiments to yield any results on rotation by magneti-
zation, the effect converse to that discovered and described in 1914, were
made early in 1915 by A. Einstein and W. J. de Haas,^ by the first method
of resonance described in section 13. On the assumption of a single type
of magneton, and the assumption that all the reaction to the magneton
momentum goes to the rod, they obtained from a few experiments on

soft iron the magnitude 2 - , or 7.1X10-V2 x, for R; but, as was soon

e

pointed out by Lorentz,' the sign of the effect was not determined with

any certainty.

In September, 1915, de Haas,' by another resonance method referred

to in section 13, obtained from a few experiments about the same mag-

m
nitude of R, viz., 2 —f and determined the sign of the effect, making R

e

negative, as Einstein and de Haas had originally claimed. These are

the only experiments on rotation by magnetization yet made in which

the magnetizing coil was wound on the rod.

In 1916 Einstein's resonance method yielded about the same magni-
tude of R.^

The first thorough experimental investigation in the field of rotation
by magnetization was completed in 1917 by J. Q. Stewart', who used
an ingenious modification of the ballistic method first attempted by
Richardson, and studied the angular momentiun produced in small
rods by annulling their residual magnetism. Stewart investigated

^ A. Einstein and W. J. de Haas, 1. c. ante.

' See A. Einstein, Verh, d. D. Phya, Oea, 17, 1015, p. 203; also W. J. de Haas, 1. c. ante.

' W. J. de Haas, I.e. ante, K. Ahad. ArMierdam, 18, 1916.

* A. Einstein, 1. o. ante, Verh, d. D, Phys. Oes, 18, 1016, p. 173.

* J. Q. Stewart, Phys, Rev, 11, 1018, p. 100.

250 ANGULAR MOMENTUM IN MAGNETISM: BARNBTT

many small rods of iron and nickel, and, on the assumptions already r

referred to, fomid R negative and equal to about one4ialf of 2 — for I

both iron and nickel. |

In 1919 TSmSi Beck^ described a thorough investigation in which he ,

repeated with great improvements the work of Elinstein and de Haas

and extended it to nickel. For R he found 53 per cent of 2 - in the case

e

of iron, and 67 per cent in the case of nickel.
In 1919 G. Arvidsson' made a similar investigation on iron, and for

two specimens obtained a value of R about equal to one-half of 2 — .

If we admit the validity of the assumption that the reaction to the
angular impulse given to the magnetons on magnetizing a rod is applied
to the rod itself, the results of Stewart, Beck and Arvidsson confirm
in a striking way those obtained by the method of magnetization by rota-
tion, into the theory of which no assumptions appear to enter except
such as can be justified completely. Otherwise they must be taken
simply as a proof that the assumed relation between the momenta
is at least approximately true, and not as affording evidence with re-
spect to R, independent of the experiments on magnetization by ro-
tation.

1 Emil Beck, Ann. der, Phys, 60. 1919, p. 100.
s G. ArvidHon, Phyt, Zmi., 21, 1920. p. 88.

I

MAGNETO-OPTICS: INQERSOLL 251

MAGNETO-OPTICS

Bt L. R. Ingbrsoll

Associate Professor of Physics, University of Wisconsin

Introduction. The connection between magneto-optic theory,^ as it
stands today, and current theories of magnetism is not as intimate as
might be expected. The explanation of this lies, perhaps, in the fact
that — speaking in general terms — ^the most acceptable magneto-optic
theory concerns itself chiefly with the changes produced by the magnetic
field in the period of vibrating electrons (which are either emitting or
absorbing Ught), the matter of orientation of the orbits being of secondary
importance, while in theories of magnetism — at least f erro-m^gnetism —
the reverse is the case. In spite of this difference in point of view,
however, magneto-optic phenomena fmnish a very interesting sidelight
on the general subject of magnetism and one whose importance cannot
be overlooked.

The discovery by Faraday* in 1845 of the rotation of the plane of
polarization of Ught produced by transmission through a block of
heavy glass in a magnetic field was the first connecting link found
between the fields of magnetism and light. This was followed some
thirty years later by Kerr's discovery' of the phenomena accompanying
reflection at the polished poleface of a magnet. A still more striking
phenomenon, however, was that of the magnetic resolution of spectral
lines brought to Ught by Zeeman^ in 1896.

These three classes, known respectively, as the Faraday, Kerr and
Zeeman effects, constitute, together with certain magnetic double-
refraction effects, the phenomena of magneto-optics. While more or
less closely connected in theory, we shall, for purposes of clearness,
consider them separately.

THE FARADAY EFFECT

Chief Charcuieristice. Magnetic rotation is a property common to
(probably) all isotropic media. While analogous to natural rotation

> General references. W. Voigt. Art. " Magneiooptik," pp. 303-710 in Graetii
**Handb. d. Elekt. u. Magn." Bd. IV. Leipzig 1915. The moat comprehensive work
on the subject. Very extensive list of references up to 1913. Largely supercedes the
author's earlier *'Magneto-u. Elektrooptik" (Leipzig, 1908).

H. A. Lorents, *'Theorie der Magneto-optischen Ph&nomena/' Encydopadie der
Math. Wiss. 5, 3, Heft 2 (Leipzig, 1909).

P. Zeeman, "Researches in Magneto-optics" London, 1913. Contains complete
bibliography of Zeeman Effect up to 1913. Also (p. XI) an excellent list of general
references on magneto-optics.

* M. Faraday, "Exper. Researches 19, No. 2146 ff.

■ J. Kerr, Phil. Mag. (5) 3. 321, 1871.

« P. Zeeman, PhU. Mag. (5) 43, 236, 1897.

262 MAONETO^PTICS: INQERSOLL

it differs from it in that the absolute direction in which the plane is
turned is independent of whether the light is passing from the north pole
to the south, or vice versa; hence reflecting the light back on itself
produces a double effect instead of annulling the rotation, as would be
the case for natural activity. The rotation may also be doubled by
reversal of the magnetic field — a fact which is generally taken advantage
of in its measurement. The majority of substances produce a positive
rotation, i.e., in the direction of the ciurent producing the field, and
this rotation is proportional (Verdet's Constant) to the field strength,
save in the case of strongly magnetic substances where it varies (Eundt's
Constant) as the intensity of magnetization. Any time lag of the
effect, if it exists at all, is exceedingly minute.^ Magnetic rotation
measurements are made with polarimetric apparatus of various kinds
dependent on the spectral r^on to be investigated. In the visible a
polarimeter with tri-field lippich' system is probably the best. In the
ultra-violet, photographic methods such as used by Macaluso and Cor-
bino' or by Landau^ are id general required, although some work has
also been done with fluorescent oculars. In the infra-red the methods
are radiometric, with selenium mirrors for polarizers as used by Meyer,*
or a system involving double-image prisms as developed by Ingersoll.*
Magnetic RoUUory Dispersion, With the exception of iron, nickel
and cobalt, in which the rotation shows an increase for longer wave--
lengths, the effect in most substances is roughly proportional to the
inverse square of the wavelength, that is for a spectral r^on which does
not contain an absorption band in which absorption and rotation are
dependent on the same set of electrons. Near such bands we must
distinguish two separate cases, in the first of which the rotation increases
(and has the same sign) on each side of the absorption band, while in
the other it shows a decrease on one side (short wavelength) and an
increase on the other, of the usual "anomalous dispersion" type.
Sodium vapor, investigated by Macaluso and Corbino' and especiaOy
by Wood,^ is a good example of the first case, while the results of Elias^
on solutions of a salt of praseodymium may probably be taken as con-
clusive evidence for the existence of the second case.

1 0. Lodge, Chem. News. 59. 191; Eng. 47. 250. 1889.

> F. Lippich. ZeU. f, Instrum. 14. 326. 1894.

< D. Macaluso and O. M. Ck>rbino, C, R, 127. 548. 1898.

* St. Landau. Phyt. Zeit. 9. 417, 1908.

■ U. Meyer. Ann. d. Phya. 30. 607, 1909.

•L. R. IngeraoU, Phyt. Rev. 23. 489. 1906; Phi!. Mag. (6) 11. 41, 1906; 18, 74, 1909.

' R. W. Wood. PhU. Mao- (6) 10, 408, 1905; 12, 329, 499, 1906; 14, 145. 1907; R. W.
Wood and H. W. Springsteen. Phyt. Rev. 21. 41. 1905; R. W. Wood and F. E. Hackett,.
Aetrophys. Jour. 10. 339, 1909.

« G. J. Elias, Ann. d. Phye. 35, 298, 1911.

MAGNETO-OPTICS: INGERSOLL 253

Of the many formulas which have been suggested for expressing mag-
netic rotatory dispersion, perhaps the most satisfactory is of the type:

-Hr.-.

for an absorption band at wavelength Xi. This, or formulas of a similar
type have been found to represent the rotatory dispersion of carbon
bisulphide and of certain optically isotropic crystals over considerable
spectral ranges. It will be noted that near an absorption band this
formula gives a dispersion of the first type mentioned above.

Wiedemann's Law. The proportionality of magnetic and natural
rotation for different wavelengths was enunciated as a law by G. Wied&*
mann,^ as a result of experiments on turpentine. It is a law to which,
however, there are very many exceptions.

Theories of Magnetic Rotation. That any theory of magnetic
rotation may start from the Fresnel point of view, i.e., the resolution of a
plane vibration into two circular ones which travel with different
velocities in the medium, has been amply established by the experiments
of Righi,* Becquerel' and Brace.^

This does not, however, lead very far towards the ultimate explanation
of a phenomenon which is essentially electronic in character. Drude^
has worked out at some length two theories, founded respectively on
the hypotheses of molecular currents and of the Hall effect. In the
first, the rotating electrons whose motion is either induced or modified
by the magnetic field are supposed to be acted on by the Ught waves
so that their centers of rotation are set in vibration. Treating them
then, more or less as rigid bodies, he arrives at an expression for magnetic
rotation. While useful in explaining magnetism and diamagnetism
and therefore a natural h3rpothesis on which to found a theory of mag-
netic rotation, this assumption of molecular currents leads to equations
which call for rotations of opposite sign on the two sides of an absorption
band and this has scant experimental verification. Voigt* criticises
the standpoint of this theory as well as that of the somewhat similar
Langevin molecule in that they involve some disagreements in the
matter of the Zeeman effect.

In the second hypothesis the electrons set in motion by the light
waves are acted on by the magnetic field with forces proportional to

1 G. Wiedemann, Pogg, Ann. 82, 216, 1851.
s A. Righi, N. Cim. (3) 3, 212, 1878.

• H. Becquerel. C. R. 88, 334, 1870.

« D. B. Brace, Wied. Ann, 26, 576, 1885; PhU, Mag. (6) 1, 464, 1901.

• "Optik." p. 407 ff. Leipiig, 1912.

• Loe.6a. p. 661.

254 MAGNETO-OPTICS: INQBRSOLL

their velocitiee. This is shown to result in an expression for magnetic
rotation of the type given above, which gives a rotation of the same sign
on each side of an absorption band as is the case for sodium vapor.

With certain simplifying assumptions this expression can be shown
to be in agreement with one involving the refractive index Mi of the
type:

e d M

m dX

which is similar to that of H. Becquerel.^ Siertsema' has applied this
formula to a series of transparent substances and the calculated values
of e/m are in fair agreement with the accepted one. Voigt' deduces
a similar formula; he also by an ingenious application of the inverse
Zeeman effect (see later) arrives at an explanation of the rotation
phenomena which is particularly satisfactory for the immediate nei|^bor-
hood of an absorption band. Magnetic rotation theories (for gases)
on the basis of the Bohr molecule have also been worked out by Som-
merfeld^ and by Pauer,* while certain electronic considerations have
been exhaustively treated in a theory by S. S. Richardson.*

THE ZEEMAN EFFECT

Chief Characteristics, Using the word in its widest sense we may
include under the Zeeman Effect any changes in light emission or absorp-
tion produced by placing the emitting or absorbing body in a magnetic
field. In a narrower sense, however, we understand by this term the
magnetic resolution of sharp spectral emi&sion or absorption lines into
two, three or more components. If a line source — flame, spark or
vacumn tube — is placed in a strong magnetic field the hnes are split into
components which, while only slightly separated, may be easily observed
and studied with modem high-resolving-power spectroscopic apparatus,
such as the Michelson echelon grating, Lummer-Gehrcke plate, inter-
ferometer or speculimi grating of good size.

This resolution is accompanied by characteristic polarization phe-
nomena. In the simplest case a single line, when the vision is at ri^t
angles to the lines of force (transversal effect), is separated by the mag-
netic field into three, of which the central one retains its original fre-
quency and is plane polarized with electric vector parallel to the field,
while the components on either side are polarized in a plane at right
angles to this. On looking parallel to the field, however, (longitudinal

> H. Becquerd, C. R. 125. 097, 1897.

• L. H. Siertsema, Amtt. Proe. 6, 413; Letden Com. No. 82. 1902.
I Lac. cU. p. 561.

• A. Sommerfeld, Ann. d. Phyt. 53, 497, 1917.

• E. Pauer, Ann. d. Phyt. 56, 261, 1918.

• B. S. Riohardflon, PhU. Mag. (6), 31, 232, 464, 1916.

MAGNETO-OPTICS: INQERSOLL 26&

effect) the central component is lacking while the two side ones are,,
respectively, right and left-handed circularly polarized.

As might be anticipated from Eirchhoff's law, an exactly similar
effect holds in the case of absorbing gases and vapors. A dark line such
as may be observed in the continuous spectrum when white light traverses
an absorbing vapor (e. g. flame containing a metallic salt, between the
poles of an electromagnet) is resolved into three (or more) when the
field is excited. On examination it is found that the absorption i»
limited entirely to the states of polarization described above. This,,
which is known as the Inverse Zeeman Effect, has an intimate bearing
on the theory of the Faraday Effect and on magnetic double-refraction.
It is also the basis of the fundamentally important researches of Hale^
and his co-workers in Solar magneto-optics.

While the simple triplet, as described, is the commonest type of reso^
lution it is by no means the only one. The sextette is only second to it
in frequency of occurrence and there are other cases of separation inta
from four to fifteen or more components, while a very few lines remain
unaffected by the field. Preston has stated as a rule that for any given
element all lines belonging to the same spectral series exhibit resolutions
of the same t3rpe and, if drawn to the scale of frequencies, of the same
magnitude. Also, corresponding lines of the different elements behave
in the same way. Runge and Ritz have also developed rules applying
particularly to the more complicated types of resolution.

Theory. The Lorentz theory of the Zeeman Effect, reduced to its
simplest terms is this: Any electronic vibration may be resolved into
three axial components, of which the Z vibration, parallel to the lines*
of force, will not be affected by the field. The other two are each resolv*
able into a right and left circular orbital motion in the X Y plane, which
may be combined into one right and one left circular. The effect will
be exactly the same as if the light emitted by the source came from
three groups of electrons vibrating in these three simple ways.

Now looking in the direction of the field a negative electron moving^
in a right-handed circular path will experience an electromagnetic pull
towards the center of its orbit which will shorten its period of revolution,
while the reverse effect will take place for the other direction of rotation.
As seen in the Z direction, then, the longitudinal) effect will be confined
to these two opposite circular polarizations. Viewed transversely,,
however, the circular motions appear "edge on" as linear vibrations
perpendicular to the field, while the central or unshifted component is
polarized at right angles to these.

Putting the matter now in mathematical form, we may write down
the equations of motion of an electron of mass m whose coordinates

> G. E. Hale, Mount Wilson Sol. Obs. Contr. No. 30 and many aubaequent papara.

260 MAONBTO-OPTICS: INOBSaOLL

"6 If 9> f relative to Ha mean porition, as:

"d^""''*' '"d^--'^' "dlp""''^'

where k is the coefficient of the "quaaiehistic" forces. That the restoring
force must be proportional to the displacement foDows from the fact
that the frequency of vibration is independent of the amplitude. Fol-
lowing the usual methods of integration these equations give as a fre-
quency (expressed in terms of an angular velocity) :

«o« Vk/m.

Now in the presence of a field H these equations must be modified to:

d«( , ..eHd^ d*n , eHdf d*f , ^

m — ^«— kfH -: m — •«— kit ^ m-r^—— kf;

dV ^^ c dt' dt« ^ c dt dt« *'

the last being unchanged by the field.
Combining the first two we eventually get as soluticxis:

e ^^ H , e 60. H

cm cm

or, to a very close approximation:

where

Now since

and

we have:

«i->

«o±q;

q-

eH
2 cm'

<■>•■

2tc
X '

a

« (-q).

, 2*

c

ax
x«'

ax

e H

■ •

X» m 4»c«'

MAGNETO-OPTICS: IN0ER80LL 257

or, putting e in electromagnetic units:

ax e H
X* m 4 T c'

The separation dX is, therefore, proportional to the square of the
wavelength and its measurement allows a calculation of the ratio

— , which was first determined by Zeeman in this way.
m

The preceding theory ia adequate in explaining only the simplest case

of the Zeeman phenomena, i.e., the normal triplet, or separation which

gives a value of — in agreement with the accepted one as determined

m

by other means. Ebctensions of the theory require in general specific
assimiptions as to the nature of the atoms or molecules. By assuming
molecular magnets much like Weiss' magnetons Bitz has been able to
represent satisfactorily many of the complicated resolutions, but his
theory has been severely critized by Voigt.^ J. J. Thomson's model
atom was thought at first sight to possess possibilities for the inter-
pretation of quartets, quintets, etc., but now appears to yield only
the simple doublet and triplet of the elementary theory.

The Bohr atom does not, at first sight, lend itself very readily to
the explanation of the Zeeman effect. However, Debye,' Bohr,' Lorentz,
and Sommerfeld^ have put forth a quantum explanation of this phe-
nomenon, which, while necessarily more complicated than the simple
Lorentz theory, has an advantage in that it accoimts not only for the
simple but also for some of the more complicated effects as well. It
is of interest to note that while the use of the Bohr model in explain-
ing the Stark electro-optic effect gives resultant equations in which the
quantimi constant A, explicitly occurs, this constant is eliminated in the
final equations for the corresponding magneto-optic (Zeeman) effect,
and the results are consequently in agreement with those of the simple
Lorentz theory.

Lastly, Lorentz* has built up a general system of equations founded
on the supposition of magnetically isotropic atoms, which Voigt* has
modified, abandoning the assumption and making use of coupled elec-
trons. Voigt's equations, while necessarily complicated and lacking in

1 W. Voigt, Ann. d. Phyt, (4) 36, 873, 1011.
< P. Del^e. Phy$. ZeU, 17, 507, 1916.

* N. Bohr, PhU. Mag. 27, 506, 1914.

« A. Sommerfeld " Atombau u. Spektrallinien*', p. 416 £F. BraunBchweig, 1921.
«H. A. Loronts, Wied. Ann. 63, 278, 1897.

• Lac. eU. p. 592.

258

MA0NBT0-0PTJC8: INGBBSOLL

deameflB, Mie, with a suitable choice of conBtaiita, capable of explain-
ing aU the leeolutions obeerved.

Vaigf$ Theory of Magneiie Roiation. On the basis of the inverse
Zeeman Effect — ^longitudinal case — a single absorption band is resolved
into two, sli^tly separated in wavelength, of right and left circular
polarization. Now resonance absorption, such as occurs in a flame with
metallic vapor, has a characteristic effect on the index of refraction,
Vfbich shows, in the neighborhood of the absorption band, "anomalous
dispersion/' Such curves are illustrated, on a frequency scale, in Fig.l.
Remembering that on the Fresnel basis magnetic rotation is due to

Fig. 1

the difference in velocity of the right and left circular components of a

plane polarised beam (rotation R»t n ( ) radians per unit

length) we may conclude that such rotation for any particular wave-
length (or frequency, n) will be proportional to the differeneo of the
ordinaUa of the two curves, as shown in the lower figure. The rotation
should have the same sign on each side of the absorption band and the
opposite sign within the band. Also, as is seen by drawing the figure
for a larger magnetic resolution, the rotation within the band, after a
certain field intensity has been reached should decrease with further

MAGNBT0-OPTJC8: INQBBSOLL

259

increase of field. These predictions have all been amply verified in
experiments with sodium flames by Zeeman,^ Hansen,' and others.

Magnetic DatMe-R^raetum. If the preceding reasoning is applied
to the transverse case, we may expect a condition of affairs given by
Fig. 2. Here the absorption band is resolved by the field into a triplet
with plane-polarized components. At any particular frequency, then,
we shall find waves of the two planes of polarization having different
refractive indices and this means double-refraction. The magnitude
of the effect is indicated by the dotted line. Experiments by Zeeman and
Geest' on sodium vapor, and particularly those byjVoigt and^ Hansen^

Fia. 2

on that of lithium, verify in a most satisfactory manner the existence
of this new magneto-optic phenomenon.

There are, however, some cases of magnetic double-refraction, such
as that discovered in pure liquids by Cotton and Mouton,* which cannot
be explained by Voigt's theory. According to this theory magnetic
double-refraction and rotation are directly related and one ou^t to be
calculable from the other, whereas, the observed double-refraction for
liquids exceeds the calculated a thousand times. In a statistical-
mechanics application of his hypothesis of molecular orientation,
Langevin* deduces formulae which fit this latter case. According to

> P. Zeeman, Anut Proc, 5, 41, 1902.

* H. M. Hansen, Ann, d. PhyaA3, 169, 1914.

* P. Zeeman and J. Geest, Anut. Proe, 7, 436, 1904.

« W. Voust and H. M. Hansen, Phy%, Zett. 13, 217, 1912.

« A. Cotton and H. Mouton, Ann. Chim. Phy. (8), 11, 146. 289, 1907.

* P. Lancevin, C. R, 151, 476, 1910.

aOO MAGNBT0-0PTIC8: INOBBSOLL

his theory magnetic rotation and doubleH:«fraction must be aascribed
to essentially di£ferent causes. The first being allied to magnetic
resolution and diamagnetism is to be explained through modification
of electronic paths by the magnetic field, the other by re-orientation
of the molecular axes. While this theory will apply to liquids, Voigt's
is eminently the most satisfactory for metallic vapors. It is indeed
quite possible that in the general case both of these factors are operative.

THE KERR EFFECT

General Charaderistica. There may be distinguished three different
types of phenomena — collectively referred to as the ''Kerr Effect/'
although Kerr himself was able to observe only the first two — ^which
arise when a surface of iron, nickel or cobalt on which is incident plane
polarised light, is subjected to a strong magnetic field. The three
types correspond to the three possible cases of magnetization, vis., with
the lines of force normal to the reflecting surface (polar case), or lying
in the surface and respectively parallel (meridional case), or perpendicular
(equatorial case) to the plane of incidence of the light.

The phenomena in general are somewhat compficated, inasmuch as
they are superposed on the optical effects which accompany the reflection
of light from the unmagnetized metallic surface. In the simplest and
best known case — polar magnetization and nonnal incidence — ^the effect
is practically nothing but a very small rotation of the plane of polari-
zation. For large angles of incidence, however, or for the other types
of magnetization an accompanying ellipticity, which is very slight in
the nonnal polar case, becomes prominent. The experimental side
has been worked out by Kerr, Righi,^ DuBois* and others. (For
additional references see Voigt, loc. cit., p. 671 ff.) An interesting point
in this connection is the fact that Heusler's magnetic alloy shows no
appreciable Kerr effect.

Theory. Inasmuch as reflection of light involves a certain amoimt of
penetration into the material, any theory of the Kerr phenomena must
involve the (Faraday) effect resulting from the penetration, as well as
the general boundary conditions for the passage of a vibration from one
body to another. That the first consideration alone is insufficient will
be gained from the fact that in the polar case — nonnal incidence — ^the
Kerr effect is just opposite in sign to the Faraday rotation due to pene-
tration.

In a continuation of the same type of reasoning as he uses in explaining
magnetic rotation on the basis of the inverse Zeeman effect, Voigt

> A. Righi. Ann. Chim. Phyt. (6). 4. 435, 1885; 9, 65» 117, 1886.
I H. E. J. G. Du Bois, FTted. Ann, 39, 25, 1800.

MAONETO^PTICS: INQERSOLL 261

has worked out a theory of the Kerr phenomena from the modem elec-
tronic view point. Applying the general boundary conditions of elec-
tromagnetism which express the continuity of the electric and magnetic
field components parallel to the surface and taking care to satisfy the
energy principle he arrives at a certain complex parameter of the metal,
Q, on which the effect uniquely depends. This constant, which involves
the vibration periods of free and bound electrons and is proportional to
the inner field or magnetization, has been computed, using as a basis
Ingersoll's results^ on the equatorial case, by Snow^ and by Voigt.'
The latter concludes that for steel and cobalt the magneto-optic be-
havior in the wave length region around 2/i is conditioned principally by
conducting electrons. Calculation of the internal fields leads to a value
of the order of 10^ and this seems to warrant the conclusion that in the
ferromagnetic metals, there exist molecules or molecular complexes
with very strong magnetic polarity, which through the external field
maintain a tendency to parallelism. The closed internal field computed
on this basis by Weiss agrees in order of magnitude with the above
value deduced from magneto-optic considerations.

> L. R. IncenoU, Phys. Ret. 35, 312, 1912*

> C. Snow, Phv9. R0V, (2). 2, 20, 1913.
• W. Voigt. Phy9. Zeit. 16, 299, 1916.

Reprint and Circular Series of the National

Research Council

Hiunber 1. Report of the Patent Committee of the National Research Council.
Presented for the Committee by L. H. Baekeland, Acting Chairman. February,
1910. Pages 24. Price 30 cents.

Niunber 2. Report of the Psychology Committee of the National Research CouneiL
Presented for the Committee by Robert M. Yerkes, Chairman. March, 1919.
Pages 5L Price 60 cents.

Hiunber 3. Refractory materials as a fidd for research. By Edward W. Wash-
bum. January, 1919. Pages 24. Price 30 cents.

Number 4. Industrial research. By F. B. Jewett. 1918. Pages 16. Price 25 cents.

Number 5. Some problems of sidereal astronomy. By Henry N. Russell. Oc-
tober, 1919. Pages 26. Price 30 cents.

Number 6. The development of research in the United States. By James Rowland
Angell. November, 1919. Pa^ 13. Price 26 cents.

Number 7. The lari^r opportunities for research on the relations of solar and
terrestrial radiation. By C. G. Abbot February, 1920. Pages 14. Price
20 cents.

Numbers. Science and the industries. By John J. Carty. February, 1920.
Pages 16. Price 25 cents.

Number 9. A reading list on scientific and industrial research and the service of
the chemist to industry. By Chuence Jay West. April, 1920. Pages 45.
Price 50 cents. '

Number 10. Report on org^sation of the International Astionomieal Union.
Presented for the American Section, International Astronomical Union, by
W. W. Campbell, Chairman, and Joel Stebbins, Secretary. June, 1920. Pages
48. Price 50 cents.

Number 11. A survey of research problems in geophysics. Prepared by Chair-
men of Sections of the American Ueophysieal Union. Octoba, 1020. Pages 57.
Price 60 cents.

Number 12. Doctorates conferred in the sciences bv American universities in 1920.
Compiled by Callie Hull. November^ 1920. Pages 9. Price 20 cents.

Number 13. Research problems in colloid chemistry. By Wilder D. Bancroft.
January-April, 1921. Pases 54. Price 50 cents.

Number 14. The relation of pure science to industrial research. By John J.
Carty. October, 1916. Pages 16. Price 20 cents.

Number 15. Researches on modem brisant nitro orolosives. By C. F. van Duin
and B. C. Roeters van Lennep. Translated by Cfaaries E. Munroe. Felouary,

1920. Pa«es 35. Price 50 cents.
Number 16. The reserves of the Chemical Warfare Service. By Charles H.

Herty. February, 1921. Pages 17. Price 25 cents.

Number 17. Geology and geography in the United States. By Edward B. Mathews
and Homer P. Gttle. April, 1921. Pages 22. Price 20 cents.

Number 18. Industrial benefits of research. By CSiaries L. Reese and A. J. Wad-
hams. February, 1921. Pages 14. Price 2^ cents.

Number 19. The university and research. By Vernon Kellogg. June, 1921.
Pages 10. Price 15 cents.

Number 20. Libraries in the District of Columbia^ Compiled by W. I. Swanton.
June, 1921. Pages 19. Price 25 cents.

Number 21. Scientific abstracting. By Gordon S. Fuldiw. September, 1921.
Pages 15. Price 20 cents.

Number 22. The National Research Counml. Its services for mining and metal-
lurgy. By Alfred D. Flinn. October. 1921. Pages 7. Price 20 cent*.

Number 23. American research chemicals. By Clarence J. West. Septembo*,

1921. Pages 28. Price 50 cents.
Number 24. Urganomagnesium compounds in svnthetic chemistry: a bibliography

of the Grignard reaction 1900-1921. By Clarence J. West and Henry CSlman
January, 1922. Pages 103. Price $1.50.

Number 25. A partial list of the publications of the National Research Council to
January 1, 1922. February, 1922. Pages 15. Price 25 cents.

Number 26. Doctorates conferred in the sciences by American universities in 1021
Compiled by Callie Hull and Clarence J. West. (In press.)

Number 27. List of manuscript bibliographies in j;eology and geopai^y. Com-
piled by Homei" P. Little. February, 1922. Pages 17. Price 25 cents.

s;.

THE BORROWER WILL BE CHARGED
AN OVERDUE FEE IFTHIS BOOK IS NOT
RETURNED TO THE LIBRARY ON OR
BEFORE THE LAST DATE STAMPED
BELOW. NON-RECEIPT OF OVERDUE
NOTICES DOES NOT EXEMPT THE
BORROWER FROM OVERDUE FEES.

I