Irving Langmuir and the "Octet" Theory of Valence

Author(s): Robert E. Kohler, Jr.

Source: Historical Studies in the Physical Sciences, Vol. 4 (1974), pp. 39-87
Published by: University of California Press
Stable URL: http://www.jstor.org/stable/27757327 .
Accessed: 27/10/2014 17:46

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

University of California Press is collaborating with JSTOR to digitize, preserve and extend access to Historical
Studies in the Physical Sciences.

http://www.jstor.org

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
Irving Langmuir and the "Octet"

Theory of Valence

BY ROBERT E. KOHLER, JR.*

1. INTRODUCTION

In a brief paper published inMarch 1916 Gilbert N. Lewis (1875-1946)

a
proposed that the chemical bond consisted of pair of electrons shared by
two atoms.1,2 Within a decade the shared pair bond was widely recognized
as the first
principle of molecular structure; but for over three years itwas
almost Its and were due
completely ignored. recognition acceptance

chiefly to Irving Langmuir (1881-1957) of the General Electric Research

Laboratory in Schenectady. Shortly after the end of the First World War
a
Langmuir realized the full significance of Lewis's ideas, and through se
ries of dramatic lectures and long papers in 1919-1920 he brought the new
ne
"Lewis-Langmuir" theory rapid and widespread popularity. The early
glect of Lewis's theory is readily understandable in terms of the intellec
tual context of 1916. More difficult to understand is the tremendous pop
of Langmuir's version of Lewis's views three years later. In
ularity only
this paper I argue that the contrasting popularity of Lewis's and Lang
muir's versions is due mainly to the different personal styles of Lewis and
a
Langmuir and to their different professional situations. Langmuir had
a
gift and a taste for popularization and popular reputation that Lewis did
not, especially among the sizeable group of industrial chemists. The new
theorywas well received in 1919 because itwas Langmuir's theory.
at first at the belated recognition of his
Although Lewis was pleased
he soon came to resent the acclaim alone en
views, popular Langmuir

joyed. A series of lettersbetween them in 1919-1920 illuminates the com

intellectual, and social issues involved in the brief
plex personal, dispute
over From these letters and other sources, the origin
priority. unpublished
of Langmuir's of Lewis's can be traced in some detail.
rediscovery theory

of History and of Science, University of Pennsylvania,

*Department Sociology
Philadelphia, Penna. 19104.
!G.N. Lewis, "The Atom and the Molecule," J. Am. Chem. Soc, 38 (1916),
762-785.
2The sources of Lewis's theory are discussed in R.E. Kohler, "The
probable
Origin of G.N. Lewis's Theory of the Shared Pair Bond," Hist. Stud. Phys. Sci., 3

(1971), 343-376.

39

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
40 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

The circumstances of Langmuir's rediscovery explain the real differences

between Lewis's and versions of the "Lewis-Langmuir"
Langmuir's theory
as well as theirvery different attitudes toward the question of priority.
For both Lewis and Langmuir valence theorywas a brief episode in their
careers. In 1921 Langmuir abruptly abandoned work on the subject, as did
Lewis after the publication of his book Valence and the Structure of
Atoms and Molecules in 1923.3 By that time the intense interest inLang
muir's "octet had died down, and the shared bond as pre
theory" pair
sented in Lewis's book was already being used by progressive chemists in

England and America. Whereas Langmuir was responsible for popularizing

the new valence theory, Lewis's version of it to be more influential
proved
in the long run. The general acceptance of the new ideas by the chemistry

profession in 1921-1926 will be the subject of a subsequent paper.

2. THE NEGLECT OF LEWIS'S THEORY, 1916-1919

Why was Lewis's 1916 theory so completely neglected at first?The war

hindered its in but America was not in
undoubtedly recognition Europe,
volved until one year after Lewis's paper appeared. In all likelihood, the
main reason for the neglect is simply that the
problem Lewis's theory so
was not as a
elegantly solved recognized problem by most chemists. For
nearly twenty years the prevailing belief had been that all chemical bonds
were polar bonds, formed
by the transferof a single electron from an elec
atom to an one. All bonds were believed to be
tropositive electronegative
the same as the polar bond of Na+Cl~, even the bonds of nonpolar organic
molecules. By 1916 the electrochemical theory as set forth by Richard
Abegg in 1904, J. J. Thomson in 1907, and others was a firmlyentrenched
orthodoxy.

Although both Abegg and Thomson (at first) explicitly denied the exis
tence of a bond, there was a small current of dissent from their
nonpolar

opinion. In 1913 W. C. Bray and Lewis proposed that a nonpolar bond

must exist, and more detailed theories of a
nonpolar bond were proposed
by J- J- Thomson himself in 1914, W. Arsem in 1914, Alfred Parson in
1915, and others.4 But in 1916 the theories were isolated; more
dissenting
over, they were highly individual, even eccentric, and led to strange and
dubious To see how Lewis's looked in historical con
consequences. theory
text,we must keep inmind that the shared pair bond, represented by dou

3G.N. Lewis, Valence and the Structure of Atoms and Molecules (New York,
1923; reprinted, New York, 1967).
4See R.E. Kohler, "Origin," op. cit. (note 2).

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 41

ble dots, was part of a small avant garde of theoretical speculation, for
which the vast majority of chemists had little interest or sympathy. For
most chemists the simple, more familiar electrochemical bond was per

fectly adequate.
The few references to Lewis's paper before 1919 reveal just how little
Lewis's understood it. The reviewer
contemporaries, including Langmuir,
forChemical Abstracts discussed only the familiar part of Lewis's theory,
how the cubic atom electron transfer bonds.5 He men
namely, explained
tioned without comment the tendency of atoms to hold even numbers of
electrons and also Lewis's idea that "atoms were i.e.,
interpenetrable,"
that could share electrons; one he was an in
they suspects passing along
itmeant.
triguing bit of mystification without really understanding what
In a lecture on valence to the American Association for the Advancement

of Science in 1918 William Noyes (1857-1941) briefly referred to Lewis's

some exceptional organic compounds might
theory.6 He admitted that
have nonpolar bonds, but he saw the exceptions as a minor appendage to
the polar orthodoxy. Noyes had been one of the earliest supporters of the
most important advance in valence theory
polar theory, and for him the
was still the electron transferbond and "the study of positive and negative
atoms in inorganic
compounds."7
in a long re
In 1917 Lewis's theory was more sympathetically discussed
view of recent theories of atomic structure by Saul Dushmann (1883
a of at General Electric.8 Dushmann too
1954), colleague Langmuir's
failed to understand Lewis's most novel points. He provided methane with
electron transfer bonds, and used Abegg's old terminology of polar va
lences and contravalences to Lewis's cubic atom. He did not men
explain
tion Lewis's view that electrons could be shared, or that they could form
stable pairs.9 His attempt to explain Lewis's formula for the ammonium
ion with its shared electron pairs reveals how little he understood what
Lewis was doing; he tried to force Lewis's elegant formula into the
Yet Dushmann did find
cramped framework of the familiar polar theory.10
Lewis's ideas and in view of later events at General
"extremely suggestive,"
Electric that fact has the greatest historical importance.

Langmuir too
was familiar with Lewis's theory and had referred to its
5E.B. Milland's review in Chem. Abst, 10 (1916), 112.
6W.A. Noyes, "Valence," Science, 49 (1919), 175-182.

1lbid.f pp. 180-181.

8S. Dushmann, "Structure of the Atom," Gen. Elec. Rev., 20 (1917), 186-196,
397-410.
*Ibid? pp. 403-406.
lOrfcid., p. 405.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 42 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

as as 1916. But was interested in

importance early Langmuir primarily
Lewis's theory of the cubic atom; his understanding of the shared pair
bond was no more than Dushmann's, and his interest
profound apparently
less. The limitations of Langmuir's viewpoint in 1916 are closely related to
his activities at General Electric, to which we must now turn.

3. LANGMUIR AND GENERAL ELECTRIC, 1909-1916

an
IrvingLangmuir had come toGeneral Electric in 1909 from unhappy
teaching position at Stevens Institute of Technology inHoboken.11 His in
terest in chemistry had been an early enthusiasm acquired from his older
brother Arthur, a successful industrial chemist. had studied metal
Irving

lurgy at Columbia (because it required much physics and chemistry), and

had made a brilliant record. After graduating in 1899 he studied physical
with Walther Nernst at where with characteristic en
chemistry G?ttingen,

ergy he also performed fifty-milehikes and sent fifty-page letters to his

family. His Ph.D. dissertation in 1906 was a study of the dissociation of
a hot filament. was a studious, and intense
gases by Langmuir shy, young
man, fascinated by ideas. His ambition was to spend his life at the fron
tier of research. In 1906, however, few American uni
physical-chemical
versities had facilities or resources for research, and with Arthur's encour

agement he the at Stevens.

accepted teaching position

Langmuir certainly realized he would have little chance at first for re

search at Stevens, but was assured that within a few he would have
years
the opportunity.12 He threw himself into teaching, trying to upgrade the
low teaching standards and to build up the slender and neglected labora
tory facilities. Langmuir's athletic intellectual ideals and his eagerness to

provide rigorous laboratory trainingwere not well received by his students,

who recalcitrant and at the new or
grew unruly regimen, by his colleagues,
who resented his efforts to shake them
up.13 The burdens of single-handed
reform and a heavy teaching load leftno time for research.His letters from
this reveal him as an ambitious man, con
period intensely energetic, young
fident of his abilities, but unaware of the problems of communication and
naive in his expectations that his superior abilities would be
quickly ac
In a
knowledged by others.14 1909 Langmuir demanded large raise and a
11A.
Rosenfeld, The Quintessence of Irving Langmuir (Oxford, 1966). See pp. 98ff.
12Letter from I. Langmuir to FJ. Pond, 2 July 1909.
Langmuir Papers, Manu
script Division, Library of Congress.
13Letter from I. Langmuir to (?) Grant, 12 April 1909. Langmuir
Papers.
14Letters in the Langmuir I am preparing a more detailed
Papers. study of Lang
muir's early career.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 43

promotion; his colleagues were only too happy to call his bluff, and in the
summer of 1909 he found himself jobless and with no prospect for a uni

versity post.15
had another In 1908 a former classmate,
Fortunately Langmuir prospect.
Colin Fink, had described to him the stimulating intellectual atmosphere at
General Electric and the research he was there. In Janu
interesting doing

ary 1909 Langmuir had contacted Willis R. Whitney, director of theGen

eral Electric about a summer research In March
Laboratory, job.16 Lang
muir an informal talk at and returned "more eager than
gave Schenectady
before" for summer research.17 was also
Whitney apparently impressed,
for it seems that a more was discussed.18 was
permanent post Langmuir
confident when he left for Schenectady inJuly that he would be offered a
He wrote to his mother that he to accept, since a of re
job. planned year
search at General Electric would be excellent experience and would give
him time to look around for what he really wanted, a university post.19
His at General Electric such a success, however, that he never
year proved
left. In industry Langmuir unexpectedly realized his ideal of complete
freedom to pursue research wherever his interests led him.

The General Electric Laboratory was unique at the time. It had been
founded in 1900 by E. W. Rice, Jr.,General Electric's technical director,
who was disturbed by the paucity of new developments in electrical tech
new technical applications would only come
nology. Rice believed that
about through fundamental research in chemistry and physics.20 It was a
a was then in one
propitious time for such faith, since physical chemistry
of itsmost productive periods. The discovery of the electron in 1897 revo
lutionized electrochemistry and the study of the electrical properties of
of obvious to the electrical Rice per
matter?subjects importance industry.
suaded Willis R. Whitney (1868-1958), then a young professor of chemis
Institute of Technology, to become the first director,
try at Massachusetts
and his choice was a happy one.21 Whitney was convinced that theway to
results for was not to force creative men such as to
get industry Langmuir

15Letter from I. Langmuir to FJ. Pond, 2 July 1909; letter from FJ. Pond to

I. Langmuir, 14 July 1909. Langmuir Papers.

16Letter from I. Langmuir to W.R. Whitney, 30 January 1909. Langmuir Papers.
17Letter from I. Langmuir to W.R. Whitney, 15 March 1909. Langmuir Papers.
18Letter from I. Langmuir to Mrs. Charles Langmuir, 16 July 1909. Langmuir
Papers.
Ibid.
20A.
Rosenfeld, op. cit. (note 11), pp. 109-112.
21G. Suits, "Willis R. Whitney," Biog. Mem. Nat. Acad. Sci., 34 (1960), 350-367.
See also Laurence Hawkins, Adventures into the Unknown, the First Fifty Years of
the General Electric Research Laboratory (New York, 1950).

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
44 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

pursue immediate applications, but to allow them to follow theirown in

terests.Whitney protected his most able chemists from routine administra
tion and red tape, and encouraged them to work freely on anything that
interested them without regard to immediate payoff for General Elec
tric.22 According to it was custom to visit his men
legend Whitney's regu

larly and to ask them if theywere "having fun." Langmuir, who invariably
was having fun, remarked one day that he did not know what
good his
fun was for General Electric. "That is not your worry; it's
having doing
mine," Whitney replied.23
at General a that was
Whitney's style created Electric spirit of research in

contrast to that at most industrial and many academic institutions.

striking

Open discussion of recent chemical and physical theorywas encouraged;

and in 1910-1915 there was much to discuss, especially in electron and
atomic Seminars and review sessions on the latest were
theory. topics orga
nized by Saul Dushmann.24 Publication and discussion of the results of
work at General Electric were and there was commu
encouraged, vigorous
nication with the best academic scientists in America and Europe. In short,

the General Electric was one of the liveliest, most forward

Laboratory

looking and productive centers of chemical and physical research in

America.

The atmosphere at General Electric was strikingly similar to the atmo

sphere of G. N. Lewis's department at Berkeley. Lewis too believed in a
free and open of research, and he and a
spirit encouraged independence
broad interest in fundamental theories. and discussion were
Cooperation
fostered by research conferences.25 Lewis, like Lang
free-wheeling weekly
muir, was addicted to theoretical it is no coincidence
speculation; perhaps
that the new of molecular structure emanated from and
theory Berkeley
where innovation was and new ideas were sure to
Schenectady, encouraged
be enthusiastically debated.

Langmuir's first project at General Electric made good use of his early
research with Nernst. In 1909 William
Coolidge (b. 1873) had just devel
oped a ductile tungsten filament that could be heated tomuch higher tem
peratures than any before. Langmuir began to study the effect of very high
on various and within about five years this work had
temperatures gases,

yielded a series of major results: the dissociation of molecular hydrogen,

22Ibid, pp. 354ff.

23A. Rosenfeld, op. cit. (note 11), pp. 124-125.
24I. Langmuir, "Saul Dushmann, A Human Catalyst," in Langmuir. The Man and
the Scientist (Oxford, 1962), pp. 409-410.
2SR.E. Kohler, "Origin," op. cit. (note 2).

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. K.OHLER, JR. 45

the invention of the atmospheric lightbulb, and the construction of a new

of the constitution of matter. An anomalous heat loss from very
theory
hot filaments led to the discovery thatH2 molecules dissociated to free hy
drogen atoms,26 and a series of important papers in 1912-1913 on the
properties of atomic hydrogen brought Langmuir widespread recognition
among physicists and chemists, including Nernst, Rutherford, Lewis, and
Bohr.27 investigations of the evaporation of tungsten fila
Simultaneously,
ments (which caused the blackening of vacuum bulbs) led Langmuir to the
idea of an improved light bulb filled with an inertgas. The nitrogen (later
argon) bulb, developed in 1913 and patented in 1916, proved an enormous
commercial success and an fame indus
gave Langmuir unparalleled among
trial chemists. His success became the most often cited proof that basic
research did pay off.
Langmuir's new theory of matter also grew out of his studies on fila
ments. He observed that formed a monomolecular film on
oxygen tung
sten so phenomenally stable that it reacted only with atomic hydrogen.
He was deeply impressed; he began to suspect that surface adsorption,
which was generally ascribed toweak "physical" forces,was in fact due to
the formation of true chemical bonds. His opinion was strengthened about
that time, 1915, by William Bragg's work on x-ray crystallography. Bragg
showed that a crystal of NaCl was a regular lattice in which each sodium
atom was bound to six chlorine atoms and vice versa, both atoms
although
were as monovalent.28 concluded that the
traditionally regarded Langmuir
force between atoms in crystals, usually regarded as a "physical" force,
was identical with the "chemical" force between atoms in a molecule, and
that molecules existed as such in the vapor state. For Lang
separate only
muir a was one vast molecule, and a in state was a chemical
crystal change

change.29 In his
two long papers of 1916-1917 he applied his new theory
to a vast of chemical and a feat he enthusiasti
array physical phenomena,
described as an outline."
cally "merely
While dramatic and ambitious new of matter was
Langmuir's very theory
never widely accepted (it had a brief life), it did attract a good deal of at
tention. Indeed, he did was attention. In 1915 he was
everything attracting
awarded the Nichols Medal by the American Chemical Society, and in

26A. Rosenfeld, op. cit. (note 11). See Chapter 9 and references, pp. 340-352.
27 Letters in the
Langmuir Papers.
28I. Langmuir, "The Constitution and Fundamental Properties of Solids and

Liquids," J. Am. Chem. Soc, 38 (1916), 2221-2295. Part II, ibid., 39 (1917),
1848-1906. See especially pp. 2221-2223.
2*Ibid, pp. 2222-2225.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
46 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

1917 the Hughes Medal by the Royal Society. Already the author of over

fiftypapers and many patents, touted for his theoretical and practical dis
coveries and in demand as a lecturer, was acknowl
increasingly Langmuir

edged to be one of the brightest stars inAmerican physical chemistry.

In his 1916 on the constitution of matter, tried to rec
paper Langmuir
oncile his new theory with the various existing theories of the chemical
bond.30 Although his taste in theorizingwas eclectic, it is not clear that he
took of these theories or even that he understood
any entirely seriously,
all of them. His treatment of Lewis's theory is a case in point. Langmuir
out Lewis's as an it
singled theory "extremely important" one; yet he gave
the least attention of all those he discussed. He emphasized Lewis's distinc
tion between and but did not even mention Lewis's
polar nonpolar types,
ideas of shared electrons and electron Yet in his discussion of Stark,
pairs.
Thomson, and Bohr he out those features that were clos
picked precisely
est, at least in appearance, to Lewis's shared bond; e.g., Thomson's
pair
two electron bond. Langmuir clearly was intrigued by Lewis's shared pair
bond but did not understand it, and thatmay be why he avoided closer
discussion. A remark Thomson's two electron bond reveals that
concerning

Langmuir had completely missed the point: "This theorywould seem to

fit in badly with the facts of organic chemistry, forwe find no evidence of

organic compounds inwhich carbon acts as if it had a valency of eight or

hydrogen a valency of two."31 Langmuir had not escaped the traditional
identification of one bond with the transferof one electron.
The key to his confusion is simply that Langmuir was not concerned
with the chemical bond or the structure ofmolecules, but with the forces
between molecules in and surface films. His conviction that chemi
crystals
cal affinitywas identical with physical cohesive forces would hardly have
led him to appreciate Lewis's theory. The shared pair bond was designed
to explain the rigid, stable geometry of organic molecules;
Langmuir went
so far in the direction as to that solid methane
opposite suggest might be
regarded as solid hydrogen held together by carbon atoms.32 There was
very little inLangmuir's own interests in 1916 thatwould have forced him
to understand Lewis's of the chemical bond.
theory

Langmuir's notebook for 1916 reveals what it was in Lewis's paper that
he did find chief concern was
"extremely important."33 Langmuir's

30Ibid., pp. 2225-2230.

^Ibid, p. 2229.
^Ibid., pp. 2232-2233.
33I. Langmuir, Notebook 803. Container 46, Langmuir Papers.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 47

atomic structure, and in his first entry, dated 20 October 1916, he went
directly to the crucial point: Bohr's orbital electrons and theirviolation of
classical He that there were better theories than
dynamics.34 thought
Bohr's, such as Alfred Parson's magneton theorywhich he had just been
Parson had that electrons, or con
studying.35 proposed "magnetons,"
sisted of circular bands of electricity rotating around fixed points at the
corners of cubic octets. He had also J. J. Thomson's
adopted positive
and this Langmuir felt was "a greater violation of classical dynam
sphere,
ics than ... a failure of Coulomb's law at short distances (Lewis)." Lang
muir was referring to Lewis's 1916 paper, which proposed that the atom
consisted of concentric cubic shells of electrons electron at each cor
(one
around the nucleus. Lewis was to Bohr's
ner) adamantly opposed dynamic
orbital atom, which he felt ignored the facts of stereochemistry.He stated
that Bohr's atom was to "save Cou
quantized planetary designed simply
lomb's Law" at short distances. Lewis abandoned Coulomb's law and

hinted that the stability of his static atom was due to a reversal of the at
tractive electrostatic force between electrons and nucleus at short dis

tances. went with this view and said so in

Langmuir along print.36
In a second entry of 26 October Langmuir went further to say that the
electrons in atoms were situated in concentric, nonrotating Rings
rings.37
of two electrons were stable, and of were very stable "as in
rings eight
Parson's theory" (with no mention of Lewis). On these principles Lang
muir constructed an Aufbau table of the elements with shells of 4, 6, 8,
12, and 16 electrons, these numbers determined quite
apparently

arbitrarily.
On 28 November set down some ideas on structure sug
Langmuir crystal
a reading of W. Kossel's 1916 theory of electron transfer
gested by
bonds.38 The polar theory, he noted, entailed that both Na+ and Cl~ atoms
in NaCl have complete cubic groups of eight electrons; Langmuir had
not mention Lewis's
clearly accepted Lewis's cubic octet, though he did
name. Albert W. Hull an
However, Langmuir's colleague (1880-1966),
fit
x-ray crystallographer, had found that the lattice data of NaCl crystals
well with octahedral ones.
poorly with outer cubic shells, but Langmuir
to devise an once based on an idea
proceeded appropriate theory, again

**Ibid., pp. 34-39.

35A. of the Structure of the Atom," Smithsonian
Parson, "A Magneton Theory
Misc. Coll., 65
(1915), 1-80.
36In "Constitution . . . Part II," op. cit. (note 28), pp. 1850-1856.
37I. Langmuir, Notebook 803, op. cit. (note 33), pp. 82-84.
Mlbid., pp. 107-116.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
48 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

from Lewis's paper: "Lewis has found an exceptionally strong tendency

for the total number of electrons in the shells in any molecule to be an
even number. Let us assume that this is an invariable rule."39 ar
Langmuir
octets caused
gued that inNaCl crystals the strong repulsions of the cubic
the electron shells to rearrange to Na (2,2,6) and Cl(2,10,6), retaining a
stable, even number of electrons in each shell and an octahedral sextet in
the outer shell.40

In his last entry of 1 January 1917 Langmuir set down a revisedAufbau

table and a statement on atomic structure .in terms of stationary electrons,
the stable pair and octet, and the reversal of Coulomb's law within the
atom.41 Lewis's name was mentioned in to the first
Although only regard
all three were essential in Lewis's of atomic structure.
point, points theory
But it is also clear how littleLangmuir was concerned with Lewis's theory
of molecular structure and the shared bond. inter
pair Langmuir's primary
est was the physics of matter, not chemical structure. It was not Lewis's

lengthy discussion of the sharing of electron pairs that interested Lang

muir, but his brief hint of non-Coulombic forces. So too it was Parson's

elaborate, but obsolete, that he noticed, not Parson's more

physics sugges
tive ideas on the chemical bond. A list Langmuir compiled of sixteen
points any atomic theorymust explain included fifteen from physics, in
and electrons, and one from
cluding gravitation positive only chemistry,
"the nature of chemical
namely, affinity."42
did touch on one of molecular structure:
Langmuir important problem
he proposed that the very stable N? molecule consisted of a single cubic
octet surrounding both kernels with their inner shells.43 But instead of ar

ranging the six inner electrons in one shared pair and two free pairs, as he
did after his rediscovery of Lewis's shared pair in 1919, Langmuir depicted
them as two groups of three electrons. His paradigm was theAufbau of
stable electron groups, not the of electron bonds.
spatial arrangement pair
To understand later attitude on his indebtedness to Lewis, it
Langmuir's
is essential to see how in 1916 he adopted isolated parts of Lewis's theory
without it as a whole on Lewis's own terms. The most ex
seeing striking

ample is his theory of crystal structure, in which he used Lewis's observa

tion that most molecules contained even numbers of electrons. For Lewis

39Ibid., p. 109.
40Ibid., pp. 113-116. Entry dated 29 November.
41Ibid? pp. 151-157.
^Ibid
*3lbid., p. 115.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 49

this fact was evidence of a more general theory; but as usual Langmuir was
was
totally occupied with his own thoughts and saw in this fact only what
relevant to his current enthusiasm. For to see more would re
Langmuir
a occasion.
quire special

4. LEWIS'S WORK ON VALENCE THEORY, 1916-1919

Lewis made curiously little effort to publicize his theory.He even failed
to discuss the shared pair when opportunity presented itself. In September
1916 he published a short paper on the color of triphenylmethyl com
to the presence of an "odd" or unpaired elec
pounds, which he ascribed
tron.44 Such "odd molecules" had been one of Lewis's key pieces of evi
dence for the electron pair, yet he only referred in passing to his "recent
even mention the electron pair bond. He did draw an
paper" and did not
elegant picture of the dissociation of several compounds to free radicals,
must simply
using his double dot bonds;45 but without any explanation it
have baffled his readers in 1916.
Lewis was more to defend his cubic atom Bohr's orbital
eager against
one. In December 1916 he revealed his plan for a counter-revolution at the

Societies in New
joint meeting of the American Chemical and Physical
York.46 He had been slated to speak on the relation between atomic struc
ture and valence bonds?a fine opportunity to give the shared pair bond
some publicity. But in view of the "wide acceptance" of Bohr's model
among physicists, he chose instead to defend the position that electrons in
atoms were at rest. He even tried to meet the on their own
physicists
a crude mathematical some years previ
ground with theory, developed
Lewis
ously
on the basis of "electron theory
alone" (i.e., without quanta).
that the force between electrons at short was some
proposed range
attraction and repulsion.
damped periodic function alternating between
The dips determined the equilibrium positions of electrons and explained
the series of atomic the need for orbital or quan
regular spectra, obviating
tum hypotheses. It seems unlikely that Lewis's sketchy, ad hoc physics
At the end of his lecture
made much of an impression on the physicists.
Lewis mentioned his idea that atoms could share electron pairs, but im
to his idea of non-Coulombic forces. Few chemists
mediately linked it

44G.N. Lewis, "Steric Hindrance and the Existence of Odd Molecules," Proc, Nat.
Aca?L Sel, 2 (1916), 586-592.
*Hbia\, p. 591.
46G.N. Lewis, "The Static Atom," Science, 46 (1917), 297-302.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 50 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

would have been stimulated to look up Lewis's earlier paper on the

subject.

however, was interested in Lewis's He was in

Langmuir, very theory.
New York for themeeting, giving a paper on "The Structure of Solids and

Liquids and the Nature of Interatomic Forces." He and Lewis had lunch
together on 28 Decmeber, "but therewas so much left over to talk about
that they arranged tomeet again the following day to continue the discus
sion."47 the discussion centered on the static atom not
Presumably theory,
the chemical bond.
In April 1917 the entry of the United States into thewar put an end to
atomic discussion, as laboratories mobilized for war research. In November

1917 Lewis was commissioned in the chemical warfare service, and in Jan

uary 1918 he went to France where he was soon in of de

charge anti-gas
fense training.48 He returned to Washington in September 1918 and to
Berkeley some months later.During his firstmonths at Berkeley Lewis suf
fered from a brief case of postwar letdown, and was occupied with a severe
administrative crisis in the university. In June 1919 he wrote Millikan:

When I reached Berkeley I found I had the greatest aversion to

chemistry
and the chemical laboratory. For several weeks I
hardly entered the
building; but this feeling has entirely disappeared and I have never en
joyed my work more than I have for the last few months. ... As you
perhaps know, outside of the department the conditions in the univer
are unsettled, and have been at times discour
sity extremely extremely
But we are that our worst will not come true.49
aging. hoping predictions

Soon Lewis was again hard at work on his

compilation of thermodynamic
data, a project of fourteen years standing. He apparently had no inclina
tion to pursue his neglected ideas on the cubic atom and the shared
pair,
and had no inkling that theywould soon be a nine
days wonder.

5. THE ORIGIN OF LANGMUIR'S "OCTET" THEORY, 1919

Even before April 1917

Langmuir had begun war research, and through
out 1917 and 1918 he was
largely occupied with submarine detection de
vices and other projects.50 He had the
opportunity, however, whenRuther
47A. Rosenfeld, op. cit. {note 11), p. 168.
48J.H. Hildebrand, "Gilbert N. Lewis," Nilt. Acad Set. Biog. Mem., 31 (1958),
209-235.
49Letter from G.N. Lewis to R. Millikan, 7 June 1919. G.N. Lewis Archive, College
of Chemistry, University of California,
Berkeley.
50A. Rosenfeld, op. cit (note 11), pp. 156-163.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 51

ford visited Schenectady in June 1918 to air his speculations on atomic

structure. According to Rosenfeld, Rutherford was impressed and urged

Langmuir to publish.51 But itwas not until early 1919 thatLangmuir's in

terestwas aroused by the question of molecular structure and the shared

pair bond.
In his enthusiastic letter to Lewis of 22 April 1919 Langmuir revealed
the general course of his rediscovery of Lewis's theory (Letter 1, Appen
The occasion for his renewed interest in Lewis's was a collo
dix). paper
at which Langmuir pre
quium organized by Dushmann in January 1919,
sented Lewis's theory and his own extensions of it.The entry inLangmuir's
notebook52 for 12 January 1919 begins: "Yesterday I gave a colloquium
at the lab on 'Relation between Atom Structure and Properties of Ele
ments' based on Lewis' on the Atom and Molecule. I gave
mostly paper
the ideas that I have developed during the last 2 or 3 years."53 Langmuir's
notebook reveals that his postulates had undergone slight but significant
modification since 1916. The three postulates, all concerning the stability
of of electrons in the atom, now read: the most stable was
groups group
that of two electrons with one or two nuclei; the next most stable was the
some of which could be common to two groups;
group of eight electrons,
the third most stable was even number of electrons.54 These postu
any
lates show a new concern with the bonds between atoms; the first two
and the thirdwas apparently
clearly refer to the shared electron pair bond,
meant to cover atoms with two or three electron pair
bonds. Nevertheless,

was clearly still the structure of the static atom.

Langmuir's first interest
He noted that he had often discussed the static atom with Albert Hull, and
had even suggested to Hull that they write a joint paper in which Hull
would the experimental evidence for stationary electrons (presum
provide
and Langmuir the chemical evidence.55
ably from x-ray crystallography)
was also still concerned with the problem of reconciling the
Langmuir
static atom with the Bohr model. He made some attempts to derive spec
tra from a static atom and discussed at length the failings of Bohr's mecha
nism for quantum transitions.He returned to his 1916 idea that transitions
involved changes in the structure of the electron itself,56which he now
conceived might be a continuous band of electricity around the nucleus.

SlJbfd, p. 173.
52I. Langmuir, Notebook 966. Container 46, Langmuir Papers.
ttlbid., pp. 99-107.
s*lbid., p. 99.
ssjbiU, p. 100.
stlbid, pp. 101-107.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
52 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

His excitement an of 22 1919 on "Structure of Elec

grew: entry January
tron" records speculations, stimulated by discussions with Whitney and
Hull, that the band electron might be wound up, the number of turns cor
to Bohr's number r.57
responding quantum
In his next entry of 24 January 1919 Langmuir returned to theAufbau
of the higher elements.58 He also added to his postulates?which covered
the first three rows of 2, 8, and 8 elements?another the next two
covering
rows of 18 elements each,59 and an for 25 elaborates the
entry January
of these His atom model became more elaborate: to ex
Aufbau groups.60

plain why there were two rows of 8 and 18 electrons, he postulated a

planar symmetry in the atom, every electron above the plane paired with
one below.61 In the
following pages of his notebook Langmuir's theory of
the atom emerges, differing in only some refinements from the form in
which itwas published.
Thus far Langmuir was occupied wholly with his new ideas of atomic
structure. His into Lewis's of molecular structure first ap
insight theory
pears in an entry of 6 February 1919. referred to this crucial in
Langmuir

sight in his long letter to Lewis of 3 April 1920 (Letter 6, Appendix,

pp. 76-87). Although he had accepted Lewis's idea that a single bond
consisted of an electron pair (postulate one) and presumably held the com
mon belief that a molecule was most stable when each atom had a full oc
tet (postulate two), he had not seen that the electron
pairs in bonds and
the free unshared pairs together comprised a full octet. The four
edges
(electron pairs) of Lewis's cubic atom, whether free or shared with other
cubic atoms, still constituted a complete cube. Langmuir had also failed to
see that the number of shared pair bonds of any atom was
strictlylimited
to four. With the exception of his model of N2, he had never been con
cerned with the number and of electrons in a molecule.
disposition present
Until the end of January 1919 Langmuir still assumed that the number of
shared pair bonds in a molecule was no less
arbitrary than the number of
simple line bonds in ordinary structural formulas?hence his third postu
late (Letter 6, Appendix, pp. 76-87). new
Langmuir's insightwas that
the number of shared pair bonds p was
strictlydetermined by the total
number of valence electrons e and the number of octets n. This he
insight
-
called the "octet
theory" and it the equation: e = Sn
expressed by 2p.

51Ibid., pp. 118-119.

pp. 126-133.
59Ibid., p. 130.
60Ibid., pp. 134-142.
61Ibid., p. 134.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 53

notebook some hints as to how he rediscovered this

Langmuir's provides
crucial point of Lewis's theory. In an entry of 24 January 1919 he re
turned to the structure of N2 he had suggested in 1916, inwhich one octet
enclosed both nuclei. He now brought itmore in linewith his new postu
lates, and proposed that the inner group of six electrons consisted of one
free pair associated with each nucleus and one pair lyingbetween the two
nuclei62 (presumably Langmuir had a single bond in mind). A similar
structurewas also proposed forNO, the innergroup ofwhich consisted of
one pair for each nucleus and one electron lyingbetween the nuclei.63 He
tried to devise a similar formula forN02, but this line of reasoning failed
him:

I cannot think of any very satisfactory structure forN02. It is probable

that the two O atoms have their full quota of electrons [i.e., eight] and
that the N nucleus with one electron is between the two oxygen atoms.

0v P
Such a molecule is very unstable and readily polymerizes to N?N

in which N behaves normally [as in N2?] two electrons being held in

common between the two N kernels. Because of this fact that there is a
two not forming any part of a
pair of electrons around the kernels (the
group of 8)64

states
Langmuir broke off at a crucial point, precisely where he explicitly
that the electron pair between the nitrogen atoms does not form part of
any octet. In his letter to Lewis (Letter 7, Appendix) Langmuir particu
in deriving structures for the nitrogen oxides
larly recalled his difficulties
from his threeAufbau postulates, and I strongly suspect that this specific
was the germ of his "octet theory." For it is ob
problem with N02
vious in the next entry of 29 January 1919, five days later, that Lang
muir had understood the role of octets.65 He a table of compounds
gives
of chlorine and oxygen (Cl02, etc.), with each atom represented as a
at the corners by
square, the four edges of Lewis's cubic octet, and joined
shared pair bonds. Langmuir had realized that these free and shared pairs
constituted an octet, for he wrote: "In each of these Cl compounds the
chlorine atom has its surrounding cube of electrons. Thus an oxy
gotten

telbid., p. 127.
telbid., p. 132.
^Ibid., pp. 133-134.
*sibi<L, pp. 142-145.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
54 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

gen atom 0> [with three electron pairs] can add itself to the edge of any
completed cube of eight."66
In the next entry, dated 6 February 1919 and entitled "Theory of Va
lence,"67 the "octet" appears for the first time.68 The
equation following

pages are filled with specific structures derived by the octet rule for com
of nitrogen, sulfur, and carbon. For atoms of normal
pounds phosphorus,
electrovalence (Cl1, O11, Nm, CIV) the new structuresmatched the tradi
tional ones, but for compounds traditionally representedwith pentavalant
or hexavalent sulfur derived new structures with
nitrogen Langmuir only
four shared bonds atom. For all this must have been an
pair per Langmuir

intensely exciting revelation, and rather marvelous too, for he was not

sure why his theoryworked so well: "Why is it that the valence bonds in

organic compounds check with the pairs of electrons held in common but
not so in etc. It is clear that Lang
nitrogen compounds?"69 abundantly
muir's rediscovery of Lewis's theorywas wholly independent of Lewis: it

began with his idea of the structure of N2 in 1916, leading in January

1919 to his over NO2, to his cubic structures of CIO2, etc.,
puzzlement
and finally, in February 1919, to his general octet equation. The form of
Langmuir's rediscovery was distinctively his own; itwas intimately related
to the unique events of his own intellectual
history.

6. THE ANNOUNCEMENT OF THE "OCTET" THEORY,

1919-1920

Throughout February Langmuir expanded and refinedhis new theory of

valence for publication. As usual he lost no time in his enthusiasm
sharing
with the rest of the scientificworld, and it proved contagious. His
diary of
1March 1919 records an interested response from a group of
physicists?
an acid test:

Have been all month on of Electrons in Atoms

working "Arrangements
and Molecules." Most to write I have New
interesting paper attempted.
York: Amer Phys Soc Meeting. I read 10 min paper on Atomic Struc
ture.Am asked at end of
meeting to show slides etc. Talk for V2 hour.70
The real clamor over the "new with address
theory" began Langmuir's

66Ibid., p. 145.
61Ibid, pp. 146-155.
*>*Ibid, p. 147.
*9Ibid.f p. 154.
70A. Rosenfeld, op. cit. (note 11), p. 173.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 55

on 9 April 1919 to the annual meeting of theAmerican Chemical Society

at Buffalo. This meeting was normally held inDecember, but thewar had
led to its postponement?as if to serve as As he
Langmuir's springboard.
noted in his the was so enthusiastic that he was asked to
diary, response

repeat it for those who had missed it the first time:

April 9: I read paper lVi hrs (section Inorg& Phys Chem) on Electrons
in Atoms & Molecules. It arouses very great interest (more than among

physicists).

April 10: I am asked to repeat the reading of my paper. E. C. Franklin

asks me to take Phys Dept at Stanford Univ. but I am not interested.GE
Lab too
good.71

In 1920, when Langmuir again received theNichols Medal for his work on
valence, it was noted that his at Buffalo had set a
repeat performance

precedent.72
The response to his theory led him to consider its application to organic
and it led to more requests for lectures:
chemistry,

April 11: After night on train, back to Schen. L. W. Jones, Clewes &
Others called my attention to agreement of my theory with organic

comps. I will examine them more fully.

April 12: Sat: Tel call fromWashington to give my paper there. Spend
morning in library& afternoon at home studying organic, N, S, & O
Find wonderful with octet
compounds. agreement theory.73

The invitation from was to address the American Association

Washington
at its annual meeting. On April 29 Langmuir also addressed a large and
eager assembly at the National Academy of Sciences, which included
"four hundred chemists from the National Bureau of Standards."74

Throughout the rest of 1919 Langmuir gave lectures at a number of

universities, a series of seven to W. A. Noyes' at the
including department

University of Illinois.
was an effective lecturer; he loved the part, and his
Langmuir extremely
enthusiasm his audiences before him. He was almost
swept compulsively
a boy he had lectured his little brother into a corner until
pedagogic; as

IbicL, pp. 173-174.

72Editorial, "Wm. H. Nichols Award,"/. Ind. Eng. Chem., 12 (1920), 386.
73A. Rosenfeld, op. cit (note 11), pp. 173-174.
"The Structure of the Atom," Scientific Monthly, 8 (1919), 572-573.
74Editorial,

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 56 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

poor Dean had yelled for help.75 His son-in-law has described Langmuir's
passion for educating the neighborhood children:

He was like a . . . There was a

great big magnet. charged atmosphere
around him. Anything he wanted to tell you seemed exciting because he
was excited about it. It couldn't seem dull if he it was
thought important.
If he wanted to tell you about it, you wanted to hear. ... He
always
described things in simple, down-to-earth terms, and always thought up
graphic demonstrations. And he made it all fun.76

Langmuir clearly had the same effect on his adult audiences.

The dramatic success of Langmuir's Buffalo was soon noticed
speech by
the popular scientific press.77 The editor of ScientificMonthly hailed the
"new conception" of chemical combination: "it the ordinary
explains

theory of valence and indicates when this must be modified." It also

corrected Bohr's atomic Bohr's results were correct, the
theory?"while
was Such statements are a measure both of Langmuir's
theory wrong."78
enthusiasm for his and of chemists' enthusiasm for Langmuir. In
theory
like vein the editor of the industrial journal Chemical and Metallurgical
Engineering hailed the "new philosophy of chemistry and matter,"79 and
this editorial wasfollowed by a full-length exposition80 of Langmuir's
paper. The tone of this popular presentation is a curious mixture of diffi
dence and enthusiasm:

The main difficulty is thatwe must take a new point of view, and any
new is to be bothersome. . . . The vast
thing likely theory opens up

possibilities. We can imagine periods of perplexity and dizziness afterwe

have triedwith sticks and balls of wax to build up atoms and molecules,
. . . and we almost dread the task.81

To aid the "dread" task, the editors later published a description of plaster
models illustrating Langmuir's postulates and offered a kit for making
them, designed by Leffert Lefferts, a consulting chemist inNew York.82
75A.
Rosenfeld, op. cit. (note 11), p. 49.
Ibid., p. 180.
77Editorial, "Scientific Meetings," Prog. ofSci. (1919), p. 569.
78Editorial, op. cit. (note 74).
79Editorial, "A New Philosophy of Chemistry and Matter," Chem. Met. Eng., 21
(1919), 57-58.
80E. Hendrick, "The Langmuir Postulates," ibid., pp. 73-81.
*Hbid.t p. 77.
82O.R. Foster, "Some Remarkable Models of Atomic and Molecular Structure,"
Chem. Met. Eng., 23 (1920), 690-692.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 57

Such exhortations to the reader a ambivalence

betray profound among
industrial chemists to the more prestigous world of academic theoretical
chemistry. Industrial chemists were not noted for an interest in new ideas,
and one senses here the companion feelings of diffidence and distrust, as
well as enthusiasm. The attraction of the new theory
was
certainly due to

Langmuir, who belonged to both the academic and the industrial world.
Probably only Langmuir could have aroused such interest in a group that
did not ordinarily bother much about new philosophies. Since the group
of industrial chemists made up a large fraction of the chemical profession,
Langmuir's high reputation in this group was no small element in the
success of the new theory.
for Langmuir's success was his almost
Equally important compulsive
desire to publish and proselytize. Here he differed strikingly from Lewis,
who was content to quietly offer a single brief and polished statement of
his ideas with a few key illustrations. Langmuir covered every detail and
was published in brief form in
spoke from every pulpit. Langmuir's theory
the National Academy's Proceedings83 and in the Journal of theFranklin
Institute84 in June 1919. In the same month a long exposition of it ap
Chemical Society85 followed by
peared in the Journal of the American
two papers on special applications, written inMay and June and published
around the turn of the year.86'87 The second of these also included a
resume of Langmuir's most important ideas. Thus by the end of 1919
ideas were available in detail and in summary to the American
Langmuir's
chemical community, and were being widely acknowledged as a sensational
breakthrough in chemical theory.
new valence theory differed from Lewis's in a
Langmuir's version of the
number of important respects. In sheer bulk and detail itwas farmore im
theory consisted of no less than eleven "postulates,"
pressive. Langmuir's
the first seven of which concerned the structure of atoms and theAufbau
of the elements. Lewis had applied his theory only to the first three rows
of the periodic table, to which the cubic atom applied best. Langmuir, on

83I. Langmuir, "The Structure of Atoms and the Octet Theory of Valence," Proc.
Nat. Acad. Set, 5 (1919), 252-259.
84I. Langmuir, "The Arrangement of Electrons in Atoms and Molecules,"/. Frank

linlnst., 187
(1919), 359-364.
85I. Langmuir, "The Arrangement of Electrons in Atoms and Molecules," /. Am.
Chem. Soc., 41 (1919), 868-934. Received 3 March 1919.

86Langmuir, "Isomorphism, Isosterism, and Covalence," ibid., pp. 1543-1559.

87I. Langmuir, "The Octet Theory of Valence and Its Applications with Special
Reference to Organic Compounds," ibid., 42 (1920), 274-292.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 58 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

the other hand, applied his more complex atom, with its shells and planes
of symmetry and "cells" of electron pairs, to the higher elements, includ
ing the rare earths and transition metals. He took up each element
systematically, discussing its chemical and physical properties.
Langmuir's last four postulates concerned the chemical bond and molec
ular structure.They were essentially the principal points of Lewis's paper;
namely, the pairing and sharing of electrons and the stable octet.88 For
Langmuir, however, these postulates were only the startingpoints for his
new theory; as he put it,
"they lead directly to a new theory of valence
which we may call the Octet Theory."89 From the octet equation Lang
muir calculated the number of shared bonds in a molecule, charac
pair

terizing all other bonds as mobile ionic bonds. To make this distinction
clear he proposed the name of "covalent bond,"90 a choice that, like the
term "octet theory,"
immediately caught the public fancy. Using the
octet discussed of each
equation Langmuir systematically compounds
element in turn.

In method as well as
terminology Langmuir's treatment differed from
Lewis's. For one thing Langmuir paid much more attention than Lewis to
the physical properties of compounds. This attention reflects
Langmuir's
practical concern at General Electric with physical properties of materials
as well as his
deep theoretical interest in adsorption and surface chemistry
and his taste for theoreticalmodels of
physics, notably theAufbau model.
Langmuir's basic idea was that groups of two and eight electrons had
weak "external fields" and conferred chemical
nonreactive, nonpolar
properties on whereas Lewis had relied more on the empirical
compounds,
distinction between and
polar nonpolar compounds. Langmuir's approach
to molecular structure was also based on the mathematical deductive
method of physics, as shown by his octet
equation and his vision of a
"deductive chemistry."91
Although Lewis too spoke of postulates and
fancied himself an amateur his to molecular structure
physicist, approach
was a structural one, like that of an chemist.
essentially organic
In the matter of style, Lewis was farmore informal and intuitive than

Langmuir. Their stylistic difference is well illustrated by their treatment

of the structure of NH4Cl, Lewis's
key illustration of his theory. Langmuir

88I. Langmuir, of Electrons,"

"Arrangement op. cit. (note 85), pp. 887-889.
89I. Langmuir, "Structure of Atoms," op. cit. (note 83), p. 254.
90L Langmuir, "Arrangement of Electrons," op. cit. (note 85), p. 926.
911. Langmuir, "Future Developments of Theoretical
Chemistry," Chem. Met.
Eng., 24 (1921), 553-557.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 59

calculated from e = Sn- 2p (e

= = =
16, n 2) that p 0; i.e., that there is no
covalent bond between the nitrogen and chlorine atoms, but an ionic
only
bond. His long and complex discussion involved physical considerations
such as polarization and residual fields.92 Lewis, by contrast, simply drew
the formula for :NH3 in terms of double dot bonds, with its free pair
available for binding a proton. He pointed to the strikinganalogy between
NH4+ and CH4, both with complete octets, remarking that "it will be
sufficient to write an equation in terms of the new symbols in order to
make the explanation obvious."93 Indeed, Lewis never did spell out that
part of his theory, but merely left it implicit in his examples. In the short
run the significance of Lewis's elegant pictorial formulaswas not obvious,
and more abstract but more familiar deductive treatment, with
Langmuir's
its inexorable was more understandable and
seemingly logic, convincing.
But once the basic ideas were familiar, roundabout "deduc
Langmuir's
tions" became and Lewis's intuitive chemical approach
superfluous,

prevailed.94
One of Langmuir's most original points was his recognition that two
e n should have similar structures
compounds with the same values of and
and physical properties; he termed such compounds isosteres. Thus a table
of a dozen properties of N20 and CO2 (e = 16, n = 3) revealed a striking
made the
quantitative similarity that had gone unnoticed.95 Langmuir
most of this dramatic discovery, and it made a great impression at the
time. It was widely reported that it was now possible to predict the
of even before were and even
properties compounds they synthesized,
Lewis showed signs of envy. Langmuir later devoted a whole paper to
isosterism.

These differences between Lewis's and Langmuir's theories and the great
difference in their reception by the chemical world inevitably led to a
clash over
priority.

7. LEWIS AND LANGMUIR: THE DISPUTE OVER PRIORITY

On 22 April 1919 Langmuir sent Lewis a manuscript copy of his first

paper and mentioned some of his latest ideas (Letter 1,Appendix). Some
92I. Langmuir, "Octet Theory," op. cit. (note 87), p. 276.
93G.N. Lewis, "The Atom and the Molecule," op. cit. (note 1), p. 778.
94An of the limits lack of structural intuition is
example imposed by Langmuir's
his treatment of the water molecule in "Arrangement of Electrons," op. cit. (note
85), p. 893.
95I. Langmuir, "Structure of Atoms," op. cit. (note 83), p. 257.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 60 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

what naively, he also gave Lewis a brief history of the origins of his theory,

beginning with Lewis's own work. Lewis was at first not displeased;
Langmuir's work was, after all, a belated recognition of his own ideas. In a

letter to of 15 June 1919 Lewis treatment

Langmuir approved Langmuir's
of the higher elements and their physical properties, subjects that Lewis
had scarcely touched on (Letter 2, Appendix). Lewis was less impressed
with atom model and his treatment of the nitrogen oxides,
Langmuir's

subjects he had had something to say about. Langmuir replied eagerly on

28 June 1919, looking forward to a stimulating competition with Lewis
in applying their new theory (Letter 3, Appendix). Lewis, however, un
a public footrace,
doubtedly did not share Langmuir's enthusiasm for
since Langmuir was the marathon hiker and had a headstart as well.
Lewis delighted in debate, but also had a strong sense of intellectual
that time the extent of Langmuir's success was
property rights. By begin

ning to be clear to him; in June Millikan had written from Chicago that
seems to have created an
"Langmuir quite impression by adopting your
ideas and them to the I have not gone into them very
re-presenting public.

carefully as yet, but in so far as I have studied them I see littlewhich he

has added towhat you had already presented."96
Lewis's uneasiness over Langmuir's popularity is explicit in his letter to
Langmuir of 9 July 1919 (Letter 4, Appendix). Itwas clear to him by then
that much of Langmuir's work concerned the first two rows of elements,

which his own cubic atom covered. of Langmuir's statements must

Many
have struck Lewis as rephrased versions of his own insights, for which
Langmuir was getting all the credit. In particular Lewis objected to Lang
muir's catch "octet and to avoidance of Lewis's
phrase theory" Langmuir's
term "odd molecule." He might also have complained that Langmuir had

religiously avoided Lewis's double dot symbol for the covalent bond, cer
an intentional omission. Lewis's concern with nomenclature is
tainly
understandable; although he could do nothing about the public enthu
siasm for Langmuir, nomenclature was a concern and, in
professional
to Nomenclature also had an
theory, subject professional agreement.
obvious on if Langmuir's terms were the theory
bearing priority; accepted,
would be linked forever with name. The emotional overtones
Langmuir's
involved are clear in Lewis's acid reflection that "sometimes
slightly
parents show singular infelicity in naming their children, but on thewhole
they seem to enjoy having the privilege." Lewis made it politely but per

96Letter from R. A. Millikan to G.N. Lewis, 19 June 1919. Lewis Archive.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 61

fectly clear that he thought Langmuir had added littlebut a new name to
Lewis's own ideas.

By the end of 1919 Lewis no longer felt so diplomatic; two events in

particular finally led him to seek arbitration. In late 1919 W. A. Noyes
submitted a paper to the Journal of the American Chemical Society,
based on a lecture he gave at Chicago that spring,which took issue with
some points of Langmuir's work.97 Noyes had long been interested in
valence theory and had already corresponded with Langmuir at some
a
length about it inMay.98 They arranged that Langmuir would draw up
rebuttal to be published with Noyes' paper. Through William Bray, a
member of the editorial board and a close colleague of Lewis's at Berkeley,
Lewis learned of these plans and was furious. It seems that Noyes had
discussed the same questions with Lewis and that Lewis had told him of
relevant experimental work going on at Berkeley (Letter 5, Appendix).
Lewis also felt that he should have been the one to defend his theory, not

Langmuir. On 13 January 1920 Lewis wrote the editor of the Journal,

Arthur B. Lamb (1880-1952), complaining of his neglect by Noyes and
term "octet theory"
Langmuir and objecting to their use of Langmuir's
and of the popular term "Lewis-Langmuir theory." He also objected to

Langmuir's paper, "The Octet Theory of Valence," which appeared that

month and which included several key illustrations of the "octet theory"
the formation of the ammonium ion and the stereochemistry of
(e.g.,
carbon and nitrogen) that Lewis had explicitly covered in his paper.99
Lamb sent copies of Lewis's letter to several leading chemists, who
was reasonable.100 On 24 March Lamb
agreed that Lewis's position
informed Langmuir that the editors "with much heart-burning" had
decided not to publish Noyes' paper; "the greater part of his paper, while
a graceful and fitting response to the presentation of the medal was not
we feel should be printed in the Journal."
just the type of paper which
rebuttal was also refused.101 Lamb enclosed a of
Langmuir's portion
Lewis's letter, presumably to suggest the difficulties involved in his deci
sion. He hoped that the dispute would be resolved: "I feel that we shall
all suffer irreparably if there is anything but friendliness and harmony
between the two men whom I consider most for the future of our
promise

97Letter from I. Langmuir to G.A, Abbott, 16 March 1920. Langmuir Papers.

98 Letters in the
Langmuir Papers.
op. cit (note 87), pp. 278-280, 286-288.
"I. Langmuir, "Octet Theory,"
100Letter from A.B. Lamb to G.N. Lewis, 2 April 1920. Lewis Archive.
101Letter from A.B. Lamb to I. 24 March 1920. Langmuir Papers.
Langmuir,

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
62 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

science in America."102 To Lewis, Lamb counseled and under

patience

standing: "It would appear to me that if proper recognition is now given

to you for your work in this field, opportunities for discord in the future
will be less than at present. Langmuir is so intensely interested in this
problem, and has a mind of such activity and power, that he is sure to
evolve modifications of this theory sufficiently distinctive so that there
will be no confusion."103

Prompted by Lamb, Langmuir finally replied to Lewis's letter of July

1919; as he confessed in his long letter of 3 April 1920, the question of
priority was extremely distasteful to him (Letter 6, Appendix), His letter,
however, was calculated to smooth Lewis's ruffled
hardly feelings. Lang
muir's forthright remarks on the question of priority must in fact have
seemed calculated to displease; he said that he could not "perpetually be
giving credit," and then proceeded to give credit to Kossel, Parson, and
Bohr?to everyone except Lewis. as we do
practically Knowing Langmuir's
intense interest in 1916 in the basis of atomic structure, we can
physical
see how for him the of Kossel, Parson, and Bohr was more im
physics
pressive than Lewis's hints of a non-Coulombic force. There is no doubt
that Langmuir was perfectly open and honest; he was simply
defending
himself with the same vigor?indeed, relish?with which he would defend
an abstract idea. His
insensitivity to the human context of ideas was well
known; Rosenfeld observed:

was out errors" in other men's work. And

Langmuir always "pointing

occasionally the erring soul would get angry, and Langmuir would be
surprised and He liked other to correct his own errors?
puzzled. people
which were not rare, since he did not mind
sticking his neck out with
assertions. ... He that others would not
positive assumed, therefore,
mind?in fact, would welcome?his criticism of their ideas. Were not
scientists after all in pursuit of truth?
"Having Langmuir 'point out an
one of his old colleagues recalls, "was like
error'," having a boulder land
on
you."104

In a reply to Lamb, Langmuir hoped his letterwould

"help to reconcile
our differences of opinion, for the sake of frankness I have had
although
to criticize his 1916 paper in a way that he may not like."105 He was

103Letter from A.B. Lamb to G.N. Lewis, 2 April 1920. Lewis Archive.
104A. Rosenfeld, op. cit. (note 11), pp. 190-191.
105Letter from I. to A.B. Lamb, 4
Langmuir April 1920. Langmuir Papers.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 63
unaware of the social value of tact and but his
naively restraint, failings
were honest ones, as his final confession "I am rather
touching suggests:
at a loss to know what I can do or what I should do" (Letter 6,
Appendix).
Lewis was not In December 1920 he declined to collaborate
assuaged.
with Langmuir in a report on theories of atomic structure for the National
Research Council?for reasons. But he added that "in
geographical general
itmight be best to ask me to defend such theories as I have originated, and
to ask Dr. to sponsor the theories which are his."106 He stead
Langmuir
refused to the terms "octet," "covalence," and the "Lewis
fastly adopt
a letter to Noyes
Langmuir theory."107 In in 1926 he complained of
reference to a structural formula as the "Lewis
Noyes' (of HCl04)

Langmuir" formula:

This formula . . . should be called the formula of Lewis, Langmuir and

W. A. and so on, the name of every author who it
Noyes, adding quotes
hereafter. Or perhaps itwould be simpler to call it the formula of Lewis,
and not mention the names of all the others who have quoted it.

Perhaps I am inclined to be too caustic in thismatter, but I really do

feel that while people were justified in being carried away a bit by
charm and enthusiasm some years ago, to
Langmuir's personal persist,

especially as they do in England, in speaking of the Langmuir theory of

valence is inexcusable. If anyone feels the contrary I would like to ask
him to mention any one original principle of Langmuir's regarding
valence which has tenable. I do not, however, wish this ancient
proved
me unduly critical of your present paper.108
grudge tomake
then both Lewis's and interests had since turned else
By Langmuir's long

where, and a friendly exchange of letters in 1930 suggests that at least

between them the "ancient was ancient
grudge" history.
In retrospect, what can be said about the question of priority? Langmuir
at the beginning of every paper, noting that
religiously gave Lewis credit
his work was "an extension of Lewis's But having thus discharged
theory."
his obligation, he rarely referred again to Lewis's work. He also spoke of
his "octet as if it were a novel deduction from Lewis's
theory" principles,
when in fact it is clearly implicit, if not spelled out, in Lewis's paper. The
crux of the matter was rediscovery of the octet theory;
priority Langmuir's

106Letter from G.N. Lewis to F.A. Saunders, 24 December 1920. Lewis Archive.
107G.N. Lewis, Valence, op. cit. (note 3), pp. 87, 105.
108Letter from G.N. Lewis to W.A. Noyes, 13 July 1926. Lewis Archive.
109Letters in the Lewis Archive.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
64 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

the sense of revelation was so strong that his new insight felt as if itwere
his own discovery. Psychologically itwas, and itwould have taken a less
ferventman than Langmuir to see that itwas in fact a rediscovery.

8. LANGMUIR, 1921-1923

Despite his intense activity in the subject in 1919 and early 1920, Lang
muir's interest in valence theory per se was really only an interlude in his
career. his main theoretical interest had been atomic structure,
Previously
and in 1920 he returned to it.That fall he attempted to derive, by classical
mechanics, a rigorous theory of the helium atom and theH2 molecule.110
The basic this ambitious effort had been suggested by A.
idea behind
Land?, who had proposed in 1919 that only the angularmomentum of the
electrons in atoms circulated around the nucleus, while the electrons

themselves remained in limited or "octants" to

regions analogous Lang

Langmuir proposed an "oscillating" model of the static

muir's "cells."
atom inwhich two electrons described oscillating semicircular paths. This
model seemed an ideal between the chemists' static atom and
compromise
Bohr's dynamic one. By modifying the quantum postulate he was also
able to calculate the ionization potential of helium.
But quantum was not familiar to Langmuir; in a letter to
theory ground
Bohr in October 1920 he confessed he might not have used the quantum
theory correctly and asked for Bohr's criticism.111He also told Bohr that
he felt Land? and Bohr himself had underestimated the importance of

coupling between electron orbits in the stable groups of two and eight
electrons. Bohr's reply in December 1920 was enthusiastic: he too had
been deeply impressed by Land?'s ideas, and also had come to realize the
need for coupling electron orbits.112 Bohr was already at work developing
these ideas into what was to be his highly successful Aufbau theory of
atomic structure. But Bohr confessed he was dubious about Langmuir's
models: the motion of electrons, he felt,was too complex to be described
by a mechanical model. Despite Bohr's skepticism, however, Langmuir
published his complete, mechanical calculations inNovember 1921.113
U0I. Langmuir, "The Structure of the Helium Atom," Science, SI 605
(1920),
607.
111Letter from I. to N. Bohr, 25 October
Langmuir 1920. American Institute of
Physics Archives.
112Letter from N. Bohr to I. Langmuir, 3 December 1920. American Institute of
Physics Archives.
113I. Langmuir, "The Structure of the Helium Atom," Rev., 17
Phys. (1921),
339-353.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 65

In February 1921 Langmuir had another inspiration, the result of his

efforts to write a on recent theories of atomic structure for the
report
National Research Council. He to his mother:
explained

I started to write it [the report] up about two weeks ago and soon
found that I began to get new ideas on the subject. These have developed
into the most important scientific discoveries that I have yet made.
... I
Especially since last Sunday I have been working night and day.
am able to prove now with certainty that the electrons of which matter
ismade are not moving in orbits, but remain relatively stationary and I
can now calculate the forces that act on each electron.114
exactly

Langmuir rejected orbital motion and replaced the centrifugal force in

Bohr's model with a new force" that the electrons in
"quantum kept

equilibrium positions. From Bohr's equation for the hydrogen atom he

calculated the new force law and from that derived all the results of Bohr's
theory, admitting, a little naively, that "of course the law of force was
chosen to give just this result."115 He had already begun to tackle the
helium atom, but his were the promised exten
expectations premature;
sions of his never
theory appeared.
one more on valence a con
Langmuir published paper theory, logically
densed form of his octet theory.116 He reduced his eleven postulates to
three: one concerned the Aufbau of atoms; one concerned the shared pah

bond; and the third stated that the sum of electrovalences and covalences
a
(the "residual atomic charge") for any atom in a molecule tended to
minimum. The latter statement was a abstracted form of the equa
highly
-
tion e = 8n 2p, and from it Langmuir "deduced" all the empirical
valence rules. He was convinced that the "residual atomic had
charge"

important physical significance, apparently having in mind his mysterious

quantum force: "It is felt by thewriter that this postulate is a crude ex
of a very and fundamental law. When we understand
pression important
the repulsive forces between charged particles better we shall be able to
state the law in a more quantitative form."117 Since he wrote noth
nearly

ing further on this subject, Langmuir must have realized after Bohr pub
lished his new theory later in 1921 that both the "residual atomic charge"

114Letter from I. Langmuir to Mrs. Charles Langmuir, 9 March 1921. Langmuir

Papers.
115I. Langmuir, 'The Structure of the Static Atom," Science, 53 (1921), 290-293.
116I. Langmuir, "Types of Valence," Science, 54 (1921), 59-67.
wibid., p. 62.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
66 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

and the "quantum force" were totally obsolete. The long development of
on atomic thus ended in a cul-de-sac.
Langmuir's thoughts physics
chemists the direct influence of Langmuir's "octet was
Among theory"
short-lived. Its mathematical or axiomatic was soon dated as
expression
chemists became more facilewith the double dot bond. Langmuir's failure
to adopt Lewis's convention was a serious tactical error; so was his fre

quent use of the cubic atom in depicting molecular structures. By 1923

the cubic atom was an embarrassing relic (even Lewis did not depict it in
Valence, though he still favored a modified tetrahedral atom model), and
were obsolete. As Lewis out,
Langmuir's papers accordingly pointed
a
Langmuir also insisted too rigidly on the necessity of a full octet,118
no doubt of his dramatic of the "octet
consequence rediscovery theory";
Lewis foresaw that molecules with incomplete octets defined a type of
acid (later termed Lewis's structural ap
generalized "Lewis-acid").119
was probably
proach thus proved to be more flexible and useful. Valence
the decisive influence; although it bears the mark of Lewis's peculiar out
look and the historical context of itsbirth in the controversial years 1921
1922,120 illustrations of the use of the double dot bond
its numerous
provided chemists with a firmbasis for furtherapplications. Langmuir too
had to a on valence but he never
planned prepare monograph theory,
wrote it,121 and in the long run Lewis's outlook was themore influential.

I want to return, to the central the contrast between

finally, problem:
the total neglect of Lewis's theory in 1916 and its immediate recognition
un
only three years later. The initial failure of the theorywas due to the
receptive intellectual context; the theory fulfilled no obviously pressing
need, and Lewis made no effort to others of its
persuade advantages.
and chemists were content with
Organic, physical, inorganic generally
their traditional points of view. Between 1916 and 1919 therewas a slight
in the intellectual context as more chemists?even
change Noyes?recog
nized the need for a dualistic system and a nonpolar bond; and their
concern over atomic models and atomic theories became more
perhaps

open and acute. But on the whole the context had not greatly changed.
What made the difference was, of course, Langmuir himself, and his

118G.N. Lewis, Valence, op. ext. (note 3), p. 97.

n9Ibid., pp. 97-101, 142.
120It was the result of a seminar
presided over by Lewis in these years.
121 Letter from I. to N. Bohr, 3 November
Langmuir 1921. American Institute of
Physics Archives.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 67

The new was well received because it was

personal reputation. theory

Langmuir's theory and not because of its intrinsic intellectual worth.

Once received, of course, its worth became evident. What I wish to stress

is that the advantages of the new theorywere not immediately obvious,

and that had it not been "Langmuir's theory" the rediscovery and adop
tion of Lewis's theorymight well have awaited a real crisis in the theory of
chemical bonding.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
APPENDIX: CORRESPONDENCE OF
LEWIS AND LANGMUIR122

[1] I. Langmuir toG. N. Lewis, 22 April 1919.

My dear Lewis:

Until a couple of weeks ago I was under the impression that you were still
in France, and therefore have not written you in to the that
regard theory
I have been on
working recently.
I am
sending you under separate cover the manuscript of a paper which

I have sent to the Jr. of theAmerican Chemical Society, and which is to

be published in the June number.
You will probably be interested in the history of the development of
these ideas. When I read your paper on the "Atom and Molecule" in 1916,
I was immediately struck by the very fundamental nature of the ideas you

presented, and of their splendid agreement with the general facts of

chemistry, so I very soon began to look upon all chemical phenomena
from the viewpoint that you presented. It seemed tome that it accounted
particularly well with so-called physical characteristics, such as boiling
etc. In to several other about the
points, freezing-points, talking people

theory, and explaining to them how itwould apply to properties of the

elements in the two short periods, I gradually began to understand extend
the theory somewhat, especially in the direction of coming to a realiza
tion that the tendency to form groups of 8 or 2 was nearly without ex
ception, the cause of the formation of compounds, at least in the case of
the first 20 elements. While on war work at Nahant123 I met Sir Ernest
Rutherford, and later, Sir Richard Paget, and I told both of them in detail
the importance of your theory, in showing them how it could be applied
to the prediction and understanding of the properties of substances.

Early in January of this year Dr. Dushman, of this laboratory, asked me

to give a talk at our Colloquium on the subject of "Adsorption." I told
him that I thought a very much more interesting subject would be Lewis'
theory of the Atom and Molecule, and, accordingly, I read your paper

122Lewis Archive. Except for Letter 5, the letters are also in the Langmuir Papers.
The letters of G.N. Lewis are printed here by the kind permission of Dr. Richard N.
Lewis, Dr. Edward S. Lewis, and Mrs. Mary Lewis.
123 at the naval research
Langmuir worked station at Nahant, Massachusetts.

68

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 69

again carefully, and began to study how I could present the matter in a
way that it would arouse the most interest. In so I was
doing impressed
more than ever by the
general applicability of the theory, and was sur
prised that chemists in general seem to have paid so little attention to your
ideas.

My interestwas then so thoroughly aroused that I spent nearly all of my

time in the development of these ideas for about five or six weeks in

January and February. I was especially interested in extending the theory

to cover all the elements, and to broaden out the of valence so as to
theory
cover all types of compounds. I think you will be interested to see how I
was able to accomplish this. It seems to me there is no field of
chemistry
where these ideas are not going to bring about radical changes in present
conceptions.
Since writing the enclosed paper I have thought over the application of
the theory in the field of organic chemistry, and have found, as is really
to be expected, that the theory explains, as far as I can see, all of the facts
of structural in a manner. The stereo
chemistry thoroughly satisfactory
isomers of carbon, nitrogen, sulphur and phosphorus, all fit inmuch better
with the new theory than with any theory previously proposed; also the
oxonium compounds are fully in accordance with the theory.After having
spent weeks in going over the literature I have only found one set of
compounds that really puzzles me, and that is, the various hydrides of
boron, in which the boron seems to act as though itwere quadrivalent. I
refer particularly to the compounds B2H6, B6H12, and B4H10. The only
reasonable explanation that I can think of is that two boron atoms form
an inner compound with two hydrogen atoms, leaving a pseudo atom
with a single octet, (like that which I have found for the gases of N2 and
C02). This would account for the absence of such compounds as BH4,
B3H8, etc. On the other hand, I do not see on this basis how to explain
the compound KBH3O, even ifwe double thismolecule in order to get an
even number of electrons. It seems to me that work on the
experimental

physical and chemical properties of these compounds is needed to give us

a basis for the proper theory.Of course, according to the octet theory, it is
obvious that no gaseous hydrides of boron should exist.
It seems to me that the theory of the atom and molecule which we have
on quantita
developed should be capable of throwing a great deal of light
tive chemical relationships in such fields as molecular volumes, heat, and
free energy, of reactions, dissociation-constants, etc. You will notice that

in the case of compounds like N2 and CO, and again in N20 and C02,

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 70 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

quantitative relations are actually predicted by this theory. In the second

paper that I am publishing I will show that hydronitric acid and cyanic
acid have ions which have similar structures to themolecules of C02 and
N20. As far as I have been able to find the physical properties (solubil
ities, of azides and are the same.
etc.) cyanates
The next time you come East I hope verymuch that you will stop at the
laboratory, for I thinkwe could very profitably spend a day or so talking
over the further extension of this work.

Yours very truly

[2] G. N. Lewis to I. Langmuir, 15 June 1919.

My dear Langmuir:

I am sure you will pardon my long delay in acknowledging themanu

script copy of your article on "The Structure of theAtom and Molecule."

The fact is I have been hoping to find opportunity to discuss with you at

length themany interestingpoints which you have brought out; but con
ditions have been so unusual, and it has been so difficult to get the depart
ment back to its status quo ante that the opportunity has not come, and

even now I am not able to discuss the matter at as I should like.

length
But as I am just leaving for a short vacation Iwanted to let you know how
much I appreciated your thoughtfulness in sendingme an advance copy,
and how much I have been pleased by finding that, after the searching

investigation you have made, you find yourself in agreement with all of
the main conclusions which I have reached in my papers.

Although I have speculated a good deal concerning the structure of the

elements of the long periods, I have not yet developed any ideas which
seem to possess a sufficient of to warrant their
degree certainty publica
tion. I have therefore been interested in your re
extremely hypothesis
these elements, and it strikes me as of a con
garding capable explaining
siderable number of facts. The relations which you establish between

atomic structure and magnetism must have some real basis. I also
certainly
feel sympathetic toward your tautomerism of nickel, and think that this
is the direction in which the ultimate solution will be found. I cannot,
however, as the details of your atomic structure, but I
accept yet perhaps
shall after I have thought of it further.
Your treatment of the elements of the two short seems to me
periods

extremely satisfactory, and it had not occurred to me to make the exten

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 71

sive use of from which do so success

arguments physical properties you
to the structural
fully. With respect formulae of nitrogen
compounds, etc.,
therewill always be some difference of opinion, and we must bear inmind
the almost universal of tautomerism in inorganic
phenomenon compounds.
But it is great fun to play with these formulae, and I have worked out a
greatmany which, in themain, agree with your own, and which I should
have published long before this if it had not been for thewar.
Your ideas the of the atoms in the com
concerning drawing together

pounds of some of the elements of the firstperiod with each other and
with hydrogen corresponds entirelywith a view which we have developed
here, and which Bray and Branch have interested themselves in. The

of an idea of this sort seems to me to

independent development point

pretty strongly to its validity.

The of the co-ordination number is one which deserves a
question great
deal of thought, E, Q. Adams124 had a number of important ideas on this
me before
subject which he put into a paper that he showed he leftCali
fornia. I criticized it on the ground of too great brevity, and although he
has promised several times to bring it out, it has not yet been published. I
am afraid from the standpoint of publicity on this subject all of us here
have been rather remiss. I wish that I could have a chance to talk all of
these interesting things over with you, and perhaps before long I shall be
able to get East. In the meantime I shall be very glad to hear ifyou have
any new ideas on the subject. Will you be good enough to sendme one or
two of your I am not sure that mine have been sent out. I
reprints paper?
am them to you, therefore, under cover.
sending separate
With best regards, I am,
Yours very sincerely,

I. Langmuir to G. N. Lewis, 28 June 1919.

[3]

My dear Lewis

I am sending you reprints of all the papers I have published during the
last four years, and a couple of extra reprints of the paper in the June
Journal. Elwood Hendrick with my help has written up a popular account
of this forMet. & Chem. Eng. and I have two papers I am sending to the

124E.Q. Adams was a graduate student at Berkeley (Ph.D. 1914), and had also
worked summers with Langmuir at General Electric in 1909-1913. The paper re
ferred to was never published, and Adams informs me that the manuscript is lost.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 72 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

Journal of the Amer. Chem. Soc. One is on the application of the valence
theory to organic compounds particularly nitrogen compounds also
oxonium, etc It also discusses the differences
sulphonium, compounds.
between salts and subtances whose molecules are held togetherby pairs of
electrons in comolecules, and such as: 1. why are there "weak
subjects
acids" and "weak bases" but no "weak salts." 2. are there no isomers
Why
of HNCO, HN02 etc. or their inorganic saltswhile the organic derivatives
exist in two isomeric forms?

The second paper deals with isosterism and covalence."

"Isomorphism,
isosterism I mean to describe the between two substances
By relationship

(or comolecules125) having equal numbers of electrons and in similar ar

rangements. Examples are N2 and CO; N20 and C02; HN3 and HNCO
- -
and all trinitrides and cyanates; KCIO3 SrS04; CaC03
inorganic
NaN03; SrHP04 and KHSO4; NaF and MgO; Na20 and MgF2 etc. These
of substances have identical forms or are isomor
pairs nearly crystalline

phous as is to be expected by the octet theory (at least as far as can be

determined from available data). This affords direct proof by Mitscher
lich's rule that the valence (better call it covalence) of N inNaN03 is 4,
and the valence of the central atoms of SO4 ,PO4 ,CIO4"" etc. is four.

Note further that salts of C03~~, S03~~ are not isomor

corresponding

phous since in the latter the covalence is three. By the ordinary theory
however both should show valency of four in their compounds, and should
be (and were) expected to be isomorphous.
I will be much interested in knowing your objections to the "details of
the atomic structure."

It is too bad that E. Q. Adams cannot bring things to definite conclu

sions. I hope you can influence him to publish his work. I have had no
plan of publishing anything more for the present on compounds of the
types usually considered by Werner altho I realize that there is a great deal
that needs clearing up?especially from the view-point of most chemists.
I am sorry that you had not published your results on the structuresof

inorganic compounds, and I hope you will be very active in extending the
theory into new fields. Don't you think that competition of this kind
between us will be to us both and assure more
stimulating rapid progress?
We should however keep each other posted as to the particular fields we
are on to avoid
actually working duplication.
I plan to be inCalifornia inAugust and early
September, and hope I can
125"Comolecule" is a term coined by Langmuir to signify groups of covalently
bonded atoms.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 73

spend a day with you. I will let you know more definitely as tomy plans
as soon as I know them and will let you know where to write me.

Yours
Sincerely

[4] G. N. Lewis to I. Langmuir, 9 July 1919.

My dear Langmuir:

I am very glad indeed to have the complete set of your papers which
you were good enough to send tome. You may be interested to know that
in one of our recent seminars a considerable of the time was devoted
part
to a discussion of your paper on Surfaces.

I shall be very much interested in seeing your two new papers on

Structure, and I shall not publish anything further in this line until I have
seen them. Apparently you have found, as I did, that for the present the
easiest progress can be made in the study of elements which I chose for
consideration, being not only simpler in structure, but on account of the
large number of compounds and the extensive study to which they have
been subjected, giving also a much largerbody of experimental. [Sic]
It has been extremely gratifyingto find that after the extended study to
which you have given thematter you have found no one of the numerous
and rather revolutionary conclusions of my paper which you have wished
to amend. You have been remarkably successfull [sic] in applying this
to a number of concrete cases, and I do not know anyone
theory large
who could have done it so well; but to be perfectly candid I think there is
a chance that the casual reader may make a mistake which I am sure you

would be the last to encourage. He might think that you were proposing a
which in some essential differed from my own, or one
theory respects
which was based upon some vague suggestions of mine which had not been
carefully thought out. While I realize what a short distance we have gone
towards chemical it seems to me that the views
explaining phenomena,
which I presented were about as definite and concrete as was
possible

considering the condensed form of publication. I think if any confusion

should arise it would be due perhaps to points of nomenclature. For
example, while I speak of a group of eight, you speak of an octet. I think,
as a matter of fact, your expression is preferable, and I shall be glad to
as the octet
adopt it, but I should be sorry to see thewhole theory known
because it raises of the sort I have
theory, partly questions just mentioned,

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 74 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

but especially because the octet is no more fundamental to the theory

than the electron pair which constitutes the chemical bond. Many years
ago, when I first began working on this subject, and before electronswere
much known, itwas the change of valence by steps of twowhich seemed
to me about themost strikingphenomenon which had to be explained by
a of valence. It was for this reason that I laid stress upon
theory particular
the fact that so few are known odd molecules or
compounds possessing
odd atoms. These are termswhich I believe you have not adopted. Did you
think of anything better? It is of course important in a new development
of this kind that the nomenclature should be as expressive and as simple
as possible. Sometimes parents show singular infelicity in naming their
own children, but on thewhole they seem to enjoy having the
privilege.
I trust that you will not misunderstand what I have said, or think for a
moment that I am not delighted to see you working in this extremely

interesting field. I shall look forwardwith great pleasure to seeing you in

Berkeley, and hope that you will be able to spendmore than the one day
with us. Do letme know just when you can come and give us as much of
your time as possible. We have a number of interesting things to show you,
and I should like verymuch to have our students hear from you.
With best regards, I am,
Yours very sincerely,

[5] G. N. Lewis toA. B. Lamb, 13 January 1920.

My dear Lamb:

has asked me to write a word a series of statements

Bray you regarding
by W. A. Noyes and by Langmuir which he has shown me. Bray tellsme
that you take the attitude that it is the duty of the editor, as far as he is
able, to prevent which arise unfortunate
misunderstandings may through
statements in papers submitted for publication.
I feel that there is some
danger of misunderstanding in connection with
the new views of valence or molecular structure. When aroused
Langmuir
the interest and enthusiasm of Eastern chemists by his paper on the ar
rangement of electrons, atoms and molecules he a valuable
performed
service. His and his admirable methods
interesting convincing personality,
of presentation, and his of the new views to
opportunity expounding
many audiences gave the subject an impetus among chemists of all ranks

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 75

which otherwise it might not have obtained in many years. I also ap

preciate highly the intrinsic value of Langmuir's work. He not only
introduced new ideas which may prove to be very valuable, but he showed
an unusual degree of acumen in elaborating the ideas which had preceded
his. There is no one in this department, where we had discussed thewhole

subject very fully,who, inmy opinion, could have given so thorough and
so lucid a of the matter.
presentation
In spite of all this I have felt during recent developments a growing
sense of embarrassment. I have avoided, and should like to avoid,
always

anything in the nature of a polemic. Scientific discoveries are of greater

than scientific credit, but there are certain established rules of
importance

prioritywhich ought to be observed.

Langmuir's original article might be divided into two parts. One con
cerns the elements whose consideration I did not include in my paper.
This one, which is new, is not the one which has been referred to
in his and other recent The second part is an
publications. amplification
and an to numerous concrete cases of the contained in
application theory

my paper. Some of these had not occurred to me; others

developments
had been already fullyworked out in this laboratory and would have been
published except for the intervention of the war. But for all of these
Langmuir, who first published them, is entitled by the established rule of
priority to complete and sole credit.
However, in recent discussions by Langmuir and others that which is
termed the octet theory, or the Langmuir octet theory is simply the
theory of which I gave a complete though concise exposition inmy paper,
and does not ordinarily refer to the extensions or amplifications made by
He has in no case any subtractions from the
Langmuir. suggested theory
which I advanced, and his additions, in which he pushes the theory rather
farther than I am inclined to, are not ones which are involved in the ap

plication of the theory tomost chemical problems. Looking at thematter,

then, as impartially as possible I feel that it is not just to speak of the
valence nor do I like the term octet for
Lewis-Langmuir theory, theory,
while the octet (whichwas called by Parson and myself the group of eight)
is an important part of the theory it is by no means thewhole or even the
essential part.
To return to the communications from Noyes and I think it
Langmuir,
was unfortunate that Noyes' paper was sent first to
Langmuir for a reply,

and I am surprised too, by one or two of the statements in the paper by

For I had several hours last summer with
Noyes. example, spent discussing

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 76 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

him the derivatives of the amine oxides,126 showed him the explanation of
this phenomenon from the standpoint ofmy theory of structure, and told
him of an experimental investigation,which we have since got under way,
which he admitted would show conclusively which of the two types of
structure is correct.

With best wishes for theNew Year, I am,

Yours very sincerely,

[6] I. Langmuir to G. N. Lewis, 3 April 1920.

My dear Lewis:

Three times I have started to write you in to your letter

separate reply
of July 9, 1919 but in each case I have been interrupted and have so little
liked what I had written that I put offwriting. The fact is, that the points
that you raise are ones which are most to write about and
unpleasant
could be much better straightened out ifwe could talk things over. I am
therefore sorry that I was not able to see you when I was in
particularly
California last summer. With my brother, Mrs. and I were
Langmuir taking
an automobile trip thru California and had been to the Yosemite Valley
and were staying for a few days at Lake Tahoe. I had fully planned to stop
a day at Berkeley but a couple of days before we expected to leave Tahoe
Mrs. Langmuir seriously injured her hand by a large falling stone while we
were mountain climbing, and we had tomake a hurried trip to Los Angeles

(my brother's home) for medical attendance.

This morning I received a letter from Lamb in regard to a paper I had

sent him some time ago, and he enclosed a copy of a
portion of a letter
which you had written him, and suggested that itwould be unfortunate "if

anything but friendliness and harmony" should exist between us. With this
I fully agree and I therefore
plan to write you in detail and with the ut
most frankness in to the matter in your two letters.
regard brought up
I have already written you at length about the part that your
publica
tions in connection with on atomic structure
played my work and valence.
Ever since your paper appeared in 1916 I have
thought of chemical phe
-
126According to Lewis, nitrogen in amine oxides had a valence of four
(R^N*
O ), whereas Noyes believed it to have a valence of five (R3N = O). See W.A. Noyes,
"Ionization of Trimethylammonium Hydroxide," /. Am. Chem. Soc, 47 (1925),
3025-3030.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 77

nomena in terms of your theory. In January and February 1919 I did

work, and developed ideas which at that time seemed tome (and still do)
to constitute an and extension of your the
important developement [sic]

ory. I therefore have felt thatwhat I have published is not merely an ex

position of your theory but is distinctly a joint theory, and that I have
contributed at least as much as you have to the final result. I know from

your letter to Lamb that you differwith me in this opinion. I hate to dis
cuss matters of this kind and especially wish to avoid all public expressions
in such matters, but I feel that the best way to straightenout the present

difficulty is to discuss these things frankly. Let me then give my reasons

formy opinion.
In the first place we must all realize that no one can or should have a

proprietary right in a theory for all time.We must always hope that the
theorywill grow and thatmany chemists will contribute to it. It therefore
soon becomes impossible to attribute the whole theory to the one who
took the first steps. For this reason I think that it is highly desirable that
theories should not be known by the name of those who propose them.
For example, I think it ismuch better to speak of the Relativity Theory
than of Einstein's theory; of the Quantum Theory rather than Planck's
For
theory; of the Principle of Equipartition rather than that ofWaterson.
the same reason it seems to me better to speak of the Octet Theory than
of Lewis' theory of valence. (In my mind Parson is the originator of the
Octet theory). (Or possibly itmay be J. J. Thomson). The point is, the
quantum theory growsmuch bigger than Planck.
Furthermore new is based upon and is an extension of
every theory
work that has gone before. We cannot perpetually be giving credit to all
those upon which our work is based. A great many of the features of the
else
theory which you proposed in 1916 had been previously published
a
where. Stark127 had identified pair of electrons held in common between
had a pair of electrons
adjacent atoms with the valence bond. Bohr had
not between two atoms in the molecule, but he proposed a
only hydrogen
structure formethane which is essentially identical with thatwhich both
of us now assume. This is shown by the following quotation from Bohr's
paper of 1913 (Phil.Mag. Nov. 1913, p. 874):
"In such as the molecule CH4 we cannot, however, assume the
systems
existence of an axis of symmetry, and we must in such cases
consequently
omit the of circular orbits. The configuration sug
assumption exactly

der Atomdynamik III. Die Elektrizitat im chemischen

127J. Stark, Prinzipien
Atom (Leipzig, 1915).

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
78 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

a tetrahedron
gested by the theory for molecule of CH4 is of the ordinary
type; the carbon nucleus surrounded a
by very small ring of two electrons
situated in the centre, and a nucleus in every corner. The
being hydrogen
are two electrons each rotat
chemical bonds represented by four rings of
the lines the centre and the corners."
ing around connecting
This inmy opinion is a much clearer and more nearly correct picture of
the methane molecule than in your I do not mean that
any given paper.
Bohr gave a complete valence theory but it is clear that in the eyes of the
world, Bohr must be given priority over you in identifying the pair of elec
trons held in common between atoms with the valence bond. You may
claim thatBohr did not suggest that all bonds in organic compounds are of
the same nature, but on the other hand Bohr himself might say that it is

only an obvious extension of his theory to apply it to other substances.

Similarly I feel that I am justified to a large degree in claiming that in your
1916 paper you did not show that your theorywas of general application
to inorganic compounds but merely gave a few illustrations and that itwas

by no means obvious that the theory laid the foundation for a comprehen
sive theory of valence of the kind that I believe I have developed.
Kossel's paper128 appeared a couple of months (much tomy regret) be
fore yours and therefore "by the established rules of priority" should be
given credit for a number of points which form an essential part of your
theory. Thus the first, second and fifthpostulates (page 768 of your 1916
paper) of your theory are fully developed inKossel's theory. Furthermore
one important feature of your
theory, viz: that the properties of the ele
ments adjacent to the inert gases are determined by their ability to take up
or give up electrons, is
brought out in great detail inKossel's theory.As a
matter of fact Kossel's whole is the same as in so
theory practically yours
far as the application to polar compounds is concerned. Your theory of
course is far ahead of Kossel's in that it also takes into account the pairs of

electrons held in common between atoms, but as we have seen this is

just
not exclusively a feature of your
theory. Kossel emphasizes the stability
of the pair of electrons in helium and in the kernels of other atoms, and
also the octets in the heavier elements. The only important point that he
misses is that these stable can share of electrons with each
groups pairs
other.
The ideas of the octet or stable group of eight is [sic] clearly advanced
by Parson (perhaps by J. J. Thomson in some of his earlywork). Would it
128W. Kossel, "?ber Molek?lbildung ais Frage des Atombaus," Ann. d. Physik, 49
(1916), 229-362.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 79

not be perhaps as logical to speak of theOctet theory as the Parson octet

as the Lewis or the Lewis-Langmuir octet It seems to me
theory theory?
we must all frankly admit that the present theory is the outgrowth of the
work of many.

So much for the question of priority.

In all that I have said so far and inwhat follows I am discussing only that
part of my publications which covers a similar ground to that of your pa
I understand that your criticisms are not directed towards those
per. parts
of my papers which deal with the eighth group and rare earth elements etc.
In one of [the] paragraphs of your letter to Lamb you say that the sec
ond part of my June paper is an application of your theory to numerous
concrete cases and that some of thesewould probably have been published
war. I assure you that I had no intention
by you if it had not been for the
of stealing march on you in this connection and that if I had thought
a
that you were planning to publish further papers on this subject in the
near future, I would have written beforehand. As a matter of fact,
you
since many papers had appeared from your laboratory during the interval
of nearly three years, and since your paper before the American Chemical

structure of atoms did not touch

Society in Sept. 1916129 dealing with the
this of valence in a very indirect way, it was natural
upon theory except
that I should believe that you were not working very actively on this sub
not yet re
ject. Furthermore I had been told in Jan. 1919 that you had
turned from France, and it did not seem to
right hold back any contribu
tion that Imight make.
In all ofmy publications on this subject I have endeavoured in every rea
sonable to be what I have considered fair. In every paper
way scrupulously
I have in as a as a statement to the effect
placed conspicuous place possible
that my work has consisted in a development and an extension of your
In many cases I have referred to your paper in the
theory. repeatedly
course of one of mine. In all the personal letters I have written in connec

tion with this work I have gone out of my way to be fair in giving you
credit. As one illustration I enclose that firstpage of a letterwhich I wrote
assure you that this is typical of all ofmy
recently to Sommerfeld, and I
letters of this kind. I am really at a loss to know what more I can do. For
the reasons I have already givefn] and will give, I cannot honestly say that
the that I am is Lewis' of valence. In order to
theory using theory explain
without needless I must refer to my own papers
repetition frequently
rather than to yours, but in so I do not mean to claim
doing priority.

129The meeting was in December; see Lewis, "Static Atom," op. cit. (note 46).

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
80 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

In fact I have given little attention to priority except in so far as it seemed

necessary for the sake of fairness. It seems to me that it is for other chem

ists to decide to what extent each of us has contributed to the present the

ory. We are necessarily prejudiced and can only with great difficulty form
we should do is to refer reason
impartial judgements in such a matter. All
ably frequently to thework of the other and not make any claim as to pri
ority. As far as I can remember I have not made any claim at
specifically
all in regard to thematter but have frequently disclaimed priority as far as
I fairly could. I am willing to go as far in this direction as you think I
should.

For I have noticed in the last number of Nature a note

example referring
to the theory as "Langmuir's Theory of theAtom."130 This note then goes
on to discuss some which are contained in your I have
points clearly paper.

just received a letter from H. S. Allen saying that he is also writing a letter
to Nature and suggests that I write a reply. This I intend to do and in so
are contained
doing I will point out that all the features in the discussion
in your and that the should never be referred to as
paper theory Lang
muir's of the atom. But in those cases where the subject has
theory except
been covered in your it does not seem reasonable that I should
paper pro
test the term altho I would to have
against Lewis-Langmuir, greatly prefer
the theory referred to as the octet theory and have it forgotten (as will
soon be the case) who first the term octet.
proposed
The last paragraph of your letterwhich was quoted by Lamb contains
many statements with which I cannot Let us consider the statement
agree.
"that which is termed the octet theory?is simply the theory of which I
a
gave complete though concise exposition inmy paper." I have the high
est admiration for the theory which you proposed in 1916 and admit

frankly that you had a remarkably clear view of the nature of chemical va
lence. To me itwas themost stimulating paper that I had ever read and (as
I have repeatedly stated publicly in most of my lectures) was the direct
cause of the work that I
subsequently did in this field. But inmy opinion
we have
progressed essentially beyond this point now, and I doubt ifyou
will find any impartial student of the subject who will agree with you that
a of the as in my recent
you gave complete exposition theory given papers.
I will discuss your paper frommy point of view and I thinkyou will un
derstand the basis of my opinion. I hate verymuch to be forced to criti
cize a as admirable as yours, but I consider that it is very
paper honestly

130Langmuir was probably referring to S.C. Bradford, "On Langmuir's Theory of

Atoms," Nature, 105 (11 March 1920), 41.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 81

unfair not to recognize some originality in the work I have done on the

theory of valence. I believe in thismatter you are doing me an injustice

greater than thatwhich you believe I am doing you.
In the first place letme ask you whether you have
carefully tried to ex
clude from your mind all the developements [sic] which you have
thought
of since the spring of 1916? This is a particularly difficult thing to do. In
over now I am able to see a in it
reading your paper great many things
which I did not get from it at all in Jan. 1919. Furthermore I see many
statements which seem to me to be inconsistent with view
my present

point, and I find a vagueness inmany points which explains why no-one as
far as I know recognized in your work a general and important new theory
of valence. You will understand my point better frommy own experience
inworking with your theory.
Prior to Jan. 19191 had read over your paper at least twice and had been
most struck by the ease with which the ordinary valence relations (valence
and were from fundamental
contravalence) explained very few assump
tions. Also by the ease with which I was able (but thiswas not done in
your paper) to derive the physical properties. But it did not occur to me
that the theorywas a perfectly general theorywhich covered all classes of

compounds much better (not only with fewer assumptions) than any other
of valence.
theory
In Jan. 1919 (or perhaps itwas Dec. 1918) in preparing to give a collo

quium on this subject, I spent a couple of weeks in rather intensive study

of the subject. I gave the colloquium without any thought of having con
tributed anything new to the subject. Several of those present who had
read your paper urged me to publish an article on the subject for they in
sisted that they failed to get any such view of chemical phenomena from
reading your paper, and claimed thatmany of the ideas which I had pre
sented were not in your at all. For two weeks or so after I had no
paper

thought of writing a paper on the subject as it seemed tome that I had too
little new matter to present.

In connection with this colloquium, and the discussion which followed,

many arose as to the structure of I re
questions particular compounds.
member especially having a discussion in regard to oxides of nitrogen and
nitric acid. I did not see that the theory gave any definite picture for the
structure of compounds such as N20, HN02, HNO3, HC102, etc. And I
had no idea as to how to apply the theory to most of the acids of sulfur
and phosphorus.
A couple of weeks later, after I had developed the ideas in regard to the

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
 82 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

arrangement of the electrons inwhat I call the third and fourth shells, I be
came impressed with the importance of the group of
eight in the second
shell rather than the group of 2,4,6 or 8. Up to this time I had kept in
mind the view that the only requirement was that the total number of
electrons was to be an even number. I had in mind such state
particularly
ments as the following from your 1916 paper:

top p. 770 "chlorine has 8 electrons in the outer shell in chlorides, 6 in

hypochlorites, 4 in chlorites, 2 in chlorates and none in perchlorates.',
Middle p. 770 "In every substance inwhich each element has either its
highest or its lowest polar number, E will appear inmultiples of 8."131?
"In in which the elements have numbers intermediate
compounds polar
between the and the lowest the number of electrons is not as a
highest
"
rule a of 8, but is in almost all cases an even number
multiple

The one from these statements is that the numbers

impression gets polar
are to be assumed, as the valences of and +7 are as
just -l,+l,+3,+5,
sumed in the ordinary theory.There is no thought expressed in your paper
of counting up all the electrons and deducing how many pairs of electrons
must be held in common. The choice of
polar number and number of elec
trons in the shell seems to be quite arbitrary. In this respect itwas not at
all obvious from your theory that your theorywas anymore definite than
the ordinary theory. In fact it seemed much less definite, and in our dis
cussions we were the number of that
discouraged by great compounds
seemed to be possible according to your theory.
I started to say that I had
developed what seemed tome to be new ideas
as to the
stability of 8 (instead of 2, 4 or 6) electrons in the second shell
and then for the first time began to see the necessity of
having all the elec
trons take part in the formation of octets. I do not see that this idea is
stated anywhere in your paper. I consider it an essential part of what I call
the octet theory, and without itI do not see how the equation 2p = 8n - e
can be derived.

It has seemed to methat the derivation of this equation or the view

point which underlies it is a vital step inmaking the theory a general the
ory of valence. I know that in my own case the aid of this I
by equation
was enabled to derive definite structures for many
immediately substances,
whereas without it I felt I had been groping in the dark. It seems to me
that it is just this definitness [sic] which I have given to the
theory that
131 "Polar number" refers to the number of electrons transferred to or from an
atom;? refers to the total number of valence electrons in a molecule.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 83

has more than any other fact led to its whereas I have
general recognition,
found no cases in the literaturewhere your work had been considered at
all as a general theory of valence. By making it possible to calculate p the
number of pairs of electrons shared between atoms I have very greatly de
creased the number of possible structureswhich may be assumed to exist,
and the fact that even with this limitation there is agreement with the
chemical data is inmy mind the strongest evidence in favor of the theory.
There are a few other points in your 1916 paper which I would like to
call attention to. You seem to take a very indefinite stand in regard to the
structure of such substances as sodium chloride. In several em
places you

phacise [sic] the "continuous transition between themost polar and the
most of substances." to the octet as I under
nonpolar According theory
stand it there is nothing corresponding to a pair of electrons which could
a chlorine and a sodium atom together.The structure of so
possibly hold
dium chloride and an organic molecule are different in kind, not in degree,
altho in a series of I can see how it is to
compounds possible get transitions.
The statements in postulates 2 and 3 are quite at variance with what I
have called the octet These statements are:
theory.

"the number of electrons in the shellmay vary during chemical change

between 0 and 8." "The atom tends to hold an even number of electrons

in the shell."

In the theory as I have presented it, there is a tendency to have 8 elec

trons in the shell (or 18 or 32 for the heavier elements) but little or no ten
num
dency to hold 2, 4 or 6. The fact that there is nearly always an even
ber of electrons follows from the fact that only pairs of electrons are
shared between atoms. This you fail to emphacize [sic] (postulate 4 does
not mention it). On p. 772 (top) where you discuss the interpenetrability
of the shells you do not mention the importance of pairs. I think this lack
of emphacis [sic] on this important point would have made it very diffi
chemist to extend to any
cult for the average your theory large number of

cases. In fact I have met many chemists who were familiar with your pa

per who tried to put atoms together so that only electron [sic]132 would
be held in common between atoms.
adjacent

I would also criticize the following statements:

p. 777. "Ammonium ionmay be regarded as a loose complex due to the
electrical attraction of the two polar molecules."

undoubtedly meant "only one electron.'1

132Langmuir

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
84 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

p. 778. 'The union of sulfur trioxide to oxide ion to form sulfate ion is
similar to the addition of ammonia and hydrogen ion to form ammo
nium ion."

These of course are minor

points.
These are all the statements in your paper with which I seriously dis

agree. I think it is a remarkably small number and itmeans of course that I

have not felt any necessity of modifying any part of your theory. But I do
think that I have really extended and developed the theory in an essential
way and have not merely been expounding the Lewis theory.
In regard to the work that I have done I thinkyou must also agree that
in my postulates 9, 10 and 11, I have given a definiteness to the theory
which was largely lacking in your presentation. You may not agree with all
my ideas but ifyou do not then you must stillmore admit a trace of origi
nality in the work I have done. Consider for example themathematical
proof inmy paper of Feb. 1920 that the octet theory is identicalwith the
ordinary [theory] of organic chemistrywhenever we take valences equal
to 8 - E. This seems to me to furnish the strongest
proof we have yet had
of the soundness of the theory. Yet I have not been able to see how this
derivation could possibly be derived from your paper.
In the last sentence which Lamb quotes from your letteryou say: "I feel
that it is not just to speak of the Lewis-Langmuir valence theory, nor do I
like the term octet theory, forwhile the octet is an important part of the
it is no means the whole or even an essential
theory by part."

Strictly speaking if the originators of the theorymust be mentioned in

connection with it I think the theory should be called theThomson-Stark
I think you
Rutherford-Bohr-Parson-Kossel-Lewis-Langmuir theory. really
would have a hard time findingwhich one of these names has most reason
to be attached to this theory. But
joking aside don't you think I have done
enough work and contributed sufficiently to even be mentioned in con
nection with the theory ifnames are to be mentioned? The question arises;
is it physically possible to preventmy name from being mentioned in con
nection with it even if thatwere desirable? I am afraid that no protests on
my part would be of much effect. I assure that to whatever extent
you my
name has been associated with the theory
recently has resulted solely from
the fact that I have published a series of papers on it and have brought the
theory
to the attention of a number of In every case I have
large people.

spoken of your work and have said that I have merely extended and devel
oped it. But you know that people are inclined to associate (to too great a

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 85

degree) such ideas with those who develop them instead of those who orig
inate them. It seems to me that the only effective way to counteract such

a tendency is for you to publish more work in this field. If you do not
continue to be active these lines you cannot to be so
along surely expect

intimately associated with thework as are those who publish in this con
nection. In my last letter to you I expressed the hope that you would con
tinue to develop this theory.
(I am getting very sleepy and I notice that I have been rambling rather
aimlessly but I will continue until I finish so as to get the letter sent off
this time. I am doing my own typewritingas you have surely guessed from
the mistakes.)

In regard to the term "Octet Theory." When I got your letter of last

July I tried hard to think of some way of meeting your objections to this
term. But I have never succeeded in
doing this. I did not then any more
than I do now think it at all right or feasible to call it the Lewis theory.
Suppose I should at any timemake an addition to the theory thatwould
be fundamental, should I then stop calling the theory by your name or
would I always have to name each part of the theory separately in order to
prevent someone else frommaking a mistake in regard to priority? If then
we are to have a name for the not to be associated with the origi
theory
nator, it seems to me that the term Octet Theory
is more than
satisfactory

[any] other. You will notice fromwhat I have said that according tomy
view the octets play a much more important part than in the theorywhich
you presented. Thus when you consider that all the electrons in the shells
of the atoms form octets (with the single exception of H2) whereas only
certain ones as a rule form stable it seems to me that the term octet
pairs,
a
theory is justified. Iwould like to get name which would also express the
importance of the group of two, but I have not been able to think of any
use the term in spite
thing satisfactory. For this reason I have continued to
of your objections. I hope you will letme know your present opinions in
the matter for I think we should come to an agreement if possible. I would

be perfectly willing to change the nomenclature if you can suggest any

better.
thing
If other people insist in connecting our names with the theory have we
any right to object? It seems to me the case is quite similar to that of the
Stefan-Boltzmann Law, or the Bohr-Sommerfeld In both cases the
theory.
second name is that of the one who made the theorymore definite by ex
case was it necessary tomodify the
tending it or developing it. In neither
original theory.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
86 IRVING LANGMUIR AND THE "OCTET" THEORY OF VALENCE

Well, I have now fully unburdened my mind on this subject. I think it is

probably a good thing to have discussed all the points that I can think of
which may have led to our differences of opinion. It is hardly reasonable
to expect you to go to such length in explaining your views, but I do hope
you will find time to tellme frankly and in some detail just what unfair
statements I have made and in what way I can counteract their effect. Do

you think we could publish a joint note or paper in which we could

out any misconceptions? I am rather at a loss to know what I
straighten
can or what I should do. But I believe that no real ever comes from
good
friction of the kind thatmight develop between us ifwe letmatters drift
along. I have always had the highest regard for you and will continue to
value your and would hate to see any under
friendship, misunderstandings
mine that friendship.
A couple of times while I have been writing this letter I have thought
that certain parts of it (pages 5, 6, and 7)133 perhaps ought to be sent to
Lamb. It might be better to do this however, along with your letter tome.
In any case I thinkwe should let Lamb know what we finally agree upon
as the fair thing.We might even ask Lamb to arbitrate in connection with
any points that we cannot agree upon. I would be perfectly willing to
adopt any policy inmy publications etc. which he would advise. I doubt
however whether such action will be necessary.

You will be interested in a general theory of chemical reactivitywhich

Dushman (in our laboratory) has recently developed. An abstract will ap
pear shortly in the Jour, of the Franklin Inst. and Dushman will read a
paper at St. Louis. It is simply this:
The velocity of any monomolecular reaction is given by the expression

VeRT

where v is related to Q by the quantum relation Q = Nhv. Thus there is

only one empirical constant needed to express the reaction velocity at all
temperatures. From this it is easy to derive the condition for
equilibrium.
The relation holds with remarkable accuracy for all known cases of dissoci
ations into atoms,
decomposition of PH3 etc. I have also found that it [is]
to the rates of and the vapor of all sub
applicable evaporation pressures
stances even the extreme cases such as He, H2, W, Mo, Pt, Hg, UF6 etc. It

also to dissociations such as CaC03 etc. In the cases of so called as

applies
sociated substances there are the usual deviations. It is a
beautifully simple

133See pp. 52-55.

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions
ROBERT E. KOHLER, JR. 87

relation, but what does it mean? It seems to even where the mecha
apply
nisms seem very different. Dushman is now it to solu
applying successfully
bilities and to reactions in solution.

Well, it is now 2:30 A.M. and I am getting hopelessly sleepy. Iwill mail
this just as it is and will trust that you will forgiveme for not reading it
over before I send it. It is probably full ofmistakes but can nevertheless be
read.

Yours
sincerely,

This content downloaded from 70.89.78.209 on Mon, 27 Oct 2014 17:46:34 PM

All use subject to JSTOR Terms and Conditions