Skip to main content

Full text of "Chemistry And Atomic Structure"

See other formats




Ph.D. (Birmingham), B.Sc. (London) 

O.B.E., DSc., F.R.S., F.I.C 



.2. A 






THE lapid development of subatomic physics in recent 
years has been accompanied by the promulgation of many 
theories as to the ultimate structuie of matter, and these 
conceptions, although peifectly legitimate as working 
hypotheses, are nevertheless often regarded as proven facts, 
especially by those chemists who are too much engiossed 
in their piofessional studies to follow critically the trend 
of modern physical lesearch into the natuie of atoms. Yet 
it must be conceded that the acid test of the validity of 
these speculations is, do they classify, con elate and explain 
the known facts of chemical science. 

In Chetmstty and Atomic Stnictwe the author, Di. J. D. 
Main Smith, who has devoted much time, labom, and 
icseaich to the subject, has applied this ciitenon to the 
various theoiies which have at diffeient times received 
acceptance since the foundation of chemistiy to the present 

As the almost inevitable result of the lapid and inci easing 
advance of knowledge in physics and chemistry, students 
are tempted moie and moie to hurry over the fundamental 
concepts of these sciences in order to gam a supeificial 
acquaintance with recently discovered phenomena. The 
authoi coriects this unfortunate tendency by devoting the 
opening chapteis of his treatise to the fundamental topics 
of atoms, molecules, valency, electiochemistry, and the 
classification of the elements. A survey of the giowth of 
modem chemical science shows that on these sound foun- 
dations has gradually been erected the solid structuie of 
tluee-dimensional chemistry, arising out of an intensive 
cultivation of caibon denvatives, then extending to 
compounds of othei elements and embracing more 
complicated types of molecular architecture. 

To the textuie and giam of this molecular fabric modem 
physics has applied the refined methods of X-ray and 
positive ray analysis and the poweiful weapons furnished 
by radioactive matter, with the result that much of the 
innei mechanism of the chemical atom has been icvealed. 

6 Chemistty and Atomic Structure 

These epoch-making discoveries necessitate an amended 
description of the phenomena of chemical valency and 
combination in terms of the electronic theory of atomic 
stiucture. The conception of atoms as the domains of 
dynamic electrons is now shown to be compatible with 
the facts of chemical combination and stereo-isomerism. 

This treatise, which is dedicated to the memory of the 
Founder of the co-ordination theory of atomic valency and 
molecular constitution, furnishes a concise and logical 
exposition of Werner's original hypotheses and brings these 
conceptions into line with recent experimental evidence 
and modern philosophic thought. 


October ', 1924 


UNTIL comparatively recent years the theory of atoms was 
entirely a chemical theory, neither particularly useful nor 
necessary in physics. The last generation, however, has 
witnessed the development of a physical theory of atomic 
structure more complete than ever had appeared possible 
in chemistry, and it has lately been customary to regard 
chemistry as concerned only with the superficial structure 
of atoms. 

Consideration of the enoimously varied and minutely 
diversified chemical properties of the colossal number of 
known chemical compounds brings the conviction that the 
outer structure of atoms is insufficient for an explanation 
and that chemistry is fundamentally concerned with the 
structure of atoms even to the heart of the nucleus. 
Chemistry and physics must, therefore, be regarded as 
complementary sciences, overlapping at the same point as 
the classical and quantum theories of physics overlap, at 
the surface of the atom. 

I have endeavoured to maintain throughout as self- 
detached a point of view as possible in the presentation of 
historical facts and the exposition of views and theories, 
while maintaining as independent a point of view as 
possible in the critical review of hypotheses and interpreta- 
tions of experimental facts. 

I have not scrupled to introduce new and perhaps strange 
ideas, as additional or supplemental explanations of both 
simple and abstruse problems and phenomena. I do not, 
however, wish to appear to have grafted on to the work of 
otheis views foreign to their intention, and, in case it may 
not always appear from the text, I hold myself responsible 
for the following, the generalisations of Werner's co-ordina- 
tion theory ; the classification and tables for the isomensm 
of tetrahedral, octahedral, and cubic co-ordination com- 
plexes ; the criticism of Werner's nomenclature, the 
interpretation of electronic structures from radioactivity 
phenomena ; the criticism of the actinium series atomic 
weights ; the detailed calculations of Moseley's " scieening 

Chemistt y and Atomic Structure 

constants " , the criticism of the fractional atomic weight 
of hydrogen, and artificial disintegrations ; the ethei 
mechanism for energy exchange between light waves and 
electrons and for the non-radiative properties of Bohr 
orbits ; the criticism of the " relativity effect " ; the 
proposal of spatial precession domains for elliptic orbits 
for both free and combined atoms ; the criticism of the 
covalency bond of two electrons ; the proposal of the law 
of uniform atomic plan , the interpretation of chemical 
evidence in terms of electron subgroup structure ; and the 
detailed structures assigned to elements generally. 

I wish to express a great indebtedness to the work and 
ideas of the late Piofessor Alfred Werner and Professor 
G. T, Morgan in chemistry, and of Sir Oliver Lodge, 
Professors N. Bohr and A. Sommerfeld, Sii J. J. Thomson, 
and Sir Ernest Rutherford in physics and the borderland of 
physics and chemistry, and to acknowledge the unfailing 
general and generous assistance of H, M. Department of 
Scientific and Industiial Research. 


Octobei, 1924 



I ATOMS - 15 

Hindu and Greek Hypotheses of Atomic Matter 
Chemical Theories of the Atom after the Renaissance 
Richter's Law of Equivalent or Reciprocal Proportions 
Proust's Law of Constant Composition Dalton's Law 
of Simple Multiple Proportions by Weight Defects 
and Virtues of Dalton's Atomic Theory. 


Relative Weights of Ultimate Particles of Chemical 
Compounds and Element* Gay-Lussac's Law of 
Simple Multiple Proportions by Volume Avogaclro's 
Hypothesis, and Divibible Ultimate Particles or Mole- 
cules Molecular Weight and Vapour Density 
Dulong and Petit's Law of Atomic Heat Joule's Law 
of Molecular Heat Mitscherlich's Law of Iso- 
morphism Canmzzaro and the Distinction between 
Atom and Molecule. Indivisibility of the Chemical 
Atom not postulated Atomic and Molecular Weights 
of the Standard Elements Oxygen and Hydrogen 
Atomicity of Elements Equivalent Weights Basicity 
and Acidity Osmotic Pressure and Molecular Weights 
Atomic Weights and the Foundations of Chemical 
Classification, Valency, and Molecular and Atomic 
Structure Reality of Atomic Weights, Isotopes, and 
the Variation of Mass with Velocity 

III VALENCY ----------- 34 

Mechanism of Chemical Combination Radicals 
Isomensrn Dumas' Substitution Theory Laurent's 
Nucleus Theory Liebig's Theory of Radicals 
Gerhardt's Theory of Residues Dumas' Theory of 
Types Liebig's Polybasic Acid Theory Kolbe's 
Theory of Radicals Frankland and Atomic Equi- 
valency Geihardt's Theory of Types and "Atomic- 
ity " Canmzzaro's, Kekuld's, and Couper's Theories of 
Polyatomicity Chemical Bonds. Valency and Valence. 


Electrolysis of Water Davy's Electrolytic Isolation of 
Potassium and Sodium Davy's Electro-chemical 
Theory Avogadro's Theory of Oxygemcitv Berze- 
hus' Dualistic Theory Grotthus' Hypothesis Fara- 
day and the Laws of Electrolysis. Weber's Theory of 
Electricity Williamson's Theory of Atoms in Motion 
Clausius' Theory of lomsation Clerk Maxwell's 
Molecule of Electricity Johnstone Stoney's Electric 
Atoms or Electrons Helmholtz's Electric Atoms or 
Electric Units of Chemical Affinity Arrhemus' 
Theory of Electrolytes Van't Hoffs Theory of 
Osmotic Pressure Ciamician'b Theory of lomsation 
Werner's Co-ordination Theory of Chemical Affinity 
and Valency 

io Chemist? y and Atomic Structure 


Atoms in Space Growth of Three-dimensional} Latent Bonds of Frankland, Speigel, 
Airhemub, and Friend Werner's Ammonium Theory 
of Tnvalcnt Nitrogen Cham-formula; of Metal- 
ammmes Gibbs' Anticipation of the Co-ordination 
Theory Tetrahedral Carbon Valences. Geometrical, 
Stereo-, and Optical Isomensm Van't Hoff's and Le 
Bel's Theories of the Tetrahedral Carbon Atom 
Stereo-chemistry of the Univalent Elements. Valency 
and Atomic Shape Baeyer's Strain Theory and 
Directional Valences Thorpe and Ingold's Direc- 
tional Valences of Carbon 


Metals and Non-metals Dualism Prout's Hypo- 
thesis Doberemer's Triads Pettenlcofer's Series, 
Odlmg's Family Relationships De Chancouitois' 
Telluric Screw Newhnd s Law of Octal es Mende- 
leeffs Periodic Law and Classification Lothar 
Meyei's Periodic Law Periodic Tables and Curves 
Elements of Zero Valency 


Molecular Addition Compounds FiTed Integral 
Valency. Werner's Theory of Carbon Valency 
Generalisations of kernel's Co-ordination Theory 
Co-ordination Numbers Hydrolysis and Ammono- 
lysis The Four Types of Co-ordination Complexes 
Werner's Nomenclatuie Chelate Groups Valency 
of Ammonium Nitrogen Electrical Polarity of Atoms 
in Co-ordination Complexes. 

Tetrahedral and Octahedral Spatial Atomic Configura- 
tions Astbury's Trigonal Prism Coiinguiation Tetra- 
hedral Symmetry and the Types, Classes, and Isomenc 
Forms of Tetrahedral Nuclear Complexes Tetra- 
hedral Multmuclcar Complexes Number of Enautio- 
morphs and Meso-isomers Cis-trans Isomensm of 
Cyclic Compounds and Compounds with Double- 
bond*. Asymmetric Atoms Octahedral Symmetry 
and the Tjpes, Classes, and Isomenc Forms of Octa- 
hedral Nuclear Complexes lomsation IbOmensm 
Octahedral Multmuclear and Mixed Complexes Iso- 
mensm and the Co-ordination Theory. Cubic Sym- 
metry and the Types and Classes of Cubic Nuclear 
Completes The Co-ordination Theory and the 
Structure of Atoms 

12 Ghemistty and. Atomic Structure 



Ether Mechanism for Quantum Orbits and the Exchange 
of Energy between Light Waves and Electrons Som- 
merfeld's Theory of Elliptic Orbits Total, Radial, 
and Azimuthal Quantum Numbers Relativity Effect 
for Variation of Electron Mass with Velocity in Elliptic 
Orbits Dimensions and Distribution of Orbits Pre- 
cession of Electrons in Elliptic Orbits Ether Mechan- 
ism for Orbital Precession in Planets and Atoms Ellip- 
soidal Precession in Fixed Domains for Elliptic Oibtts 
Chemical Evidence necessary for the Stiuctural Details 
of Atoms 


CLASSIFICATION ...... - - 177 

Rydbeig's Quadratic Groups and Series, Langmuir's 
Theory of Atomic Structuie and Covalency Combina- 
tion Electronic Theones of Chemical Combination 
Bury's Amendment of Langmmr's Theory Sir J J 
Thomson's, Theory of Atomic Structure and Chemical 
Combination Bohr'& Theory of Atomic Structure 
Sidgwick's Extension of Bohi's Theory to Chemical 
Combination The Law of Uniform Atomic Plan 


The Extension of Bohr's Theory to Atoms in Combina- 
tion The Two Subgioups of the Valency Electrons 
of Tnvalent Atoms The Three Subgroups of the 
Valency Electrons of Multivalent Atoms Persistence 
of the Subgroups in the Interior of Atoms The Dis- 
crepancies of Bohr's Subgroup Scheme Deduction 
and Establishment of the Law of Uniform Atomic Plan 
Details of the Atomic Structures of all the known 

Alteration of Mass with Velocity Deformation of 
Orbits oJ Massive Bodies circulating with Varying 
Velocity Sommerfeld's " Relativity Effect " for 
Elliptic Orbits of Electrons in Atoms. Plane Pre- 
cession Bi-nuclcar Orbits of Two Atoms m Non- 
dissociable Chemical Combination Ovals of Cassmi 
and the Lemmscate of Bernouilli. Orbital Precession 
for Bi-nuclear Orbits. Spatial Precession Super- 
position of Simple Harmonic Motions. 

BIBLIOGRAPHY - ........ 207 

Atomic Structure. Chemical Co-ordination Valency 




I Tetrahedral Valency Deflection ------- 64 

II Johnstone Stoney's Periodic Spiral ------- 75 

III Tetrahedral Optical Isomers -------- 96 

IV Bi-nuclear Tetrahedral Isomers ------- jor 

V. Four Octahedral Isomers of Type 2, Class (v) - - - - 106 

VI Octahedral lomsation Isomers, Non-chelate ----- 108 

VII. Octahedral lomsation Isomers, Chelate ------ 109 

VIII Octahedral lomsation Isomers, Chelate - - - - - - m 

IX Cubic Isomers, Chelate - - - - - - - - - 116 

X Curves of Moseley's Constants- - - - - - - - 151 

XI Electron Orbits, 2 Quanta, Unsymmetncal - - - - - 167 

XII. Electron Orbits, 3 Quanta, Unsymmetncal - - - - - 168 

XIII Electron Orbits, 4 Quanta, Unsymmecrical - - - - - 168 

XIV Electron Orbits, Symmetrical -------- 169 

XV Interpenetrating Electron Orbits, Unsymmetncal - 171 

XVI Ellipsoidal Spatial Precession Orbits - - - - - - 175 

XVII Single Relativity Orbit ---------199 

XVIII Successive Relativity Orbits -------- 200 

XIX Figure-of-eight Bi-nuclear Orbits ------- 203 

XX Caseinian Oval Bi-nuclear Orbits ------- 204 



1 Mendele"eff's Periodic Table of 1871 ------- 71 

2 Bayley's Periodic Table ---------- 74 

3 Periodic Classification ---------- 79 

4 Types and Classes of Tetrahedral Isomers ------ 97 

5 Types and Classes of Octahedral Isomers ------ 104 

6 Isomenc Forms of Octahedral Type i ------- 104 

7 Isomeric Forms of Octahedral Type 2------- 105 

8 Isomeric Forms of Octahedral Type 3------- 105 

9 Isomeric Forms of Octahedral Type 4------- 107 

10 Genealogical Table of the Radioactive Elements 133 

11 Bohr's Atomic Structures for the Inert Gas - - - - - - 184 

12 Atomic Structures of " Key " Elements ------ 196 

13 Atomic Structures in the First Transition Series - - - - 196 

14 Atomic Structures in the Second Transition Series - - - - 197 

15 Atomic Stiuctures in the " Rare Earth " Transition Subseries - - 197 
1 6. Atomic Structures in the " Platinum " Transition Subsenes - - 198 

The contractions used in the references to original papers are, in nearly 
all cases, those used in the Abstracts of Chemical Papers issued by the 
Bureau of Chemical Abstracts 



SPECULATIONS as to the ultimate causes and origins of 
material phenomena usually resolve into notions regarding 
infinity, and fall into two classes, related as proposition and 
its converse. Matter, for example, may be regarded as 
continuous or granular. Theories of the continuous natuie 
of matter postulate that any portion of any apparently 
homogeneous substance is throughout as homogeneous as 
it appears, and admits of unlimited subdivision without 
any change in nature appearing. Theories of the granular 
structure of matter, on the other hand, postulate that any 
portion of any apparently homogeneous substance is 
reducible by subdivision to a particle which is eithei 
indivisible, or divisible, only with a fundamental change in 
the nature of the substance into parts consisting of indi- 
visible particles. All atomic theories postulate indivisible 
ultimate particles, though in chemistry atomic theories 
relate to particles which are indivisible except with change 
in properties, and ultimate indivisibility is not postulated. 

The earliest lecorded system of philosophy, based on an 
atomistic view of matter, appears to be that of the Hindu 
philosopher, Kanada, probably earliei than 1000 B.C., who 
postulated the existence of small particles of matter con- 
sisting each of a few ultimate indivisible particles 01 monads. 
Kanada's doctrines were early incoiporated in the tenets of 
the various sects of the Buddhist religion, and became 
widely diffused over Asia. 

It has been suggested, and there is considerable evidence 
of the westward spread of Buddhist doctrines prior and 
subsequent to the rise of the eaily Greek civilisation, that 
the atomistic views of Leucippus, a Gieek philosopher of 
about 450 B.C., and of his pupil Democritus (425 B.C.) are 
to be ascribed to Buddhist influences. However this may 
be, it is certain that the atomic hypothesis, usually associated 
with the name of Democritus, alone has directly influenced 
western scientific thought since the Renaissance. 

1 6 Chemistry and, Atomic Structure 

The atomistic philosophy of Democritus was modified 
by Plato, a pupil of Socrates, in the dialogue, the Ttmaeus, 
about 400 B.C., by grafting on parts of the mathematical 
and geometrical doctrines of the later Pythagoreans. Plato 
assigned to the primordial atoms the shapes of the five 
regular solids, the tetrahedron, octahedron, cube, icosa- 
hedron, and dodecahedron, the four, six, eight, twelve, and 
twenty point symmetrical configurations respectively. The 
Democritan doctrine was further elaborated by Epicurus 
(about 300 B.C.) into a comprehensive system of philosophy, 
though little now remains on record of the Epicurean 
system except in the poem, De Rerum Natura, of the 
Roman, Lucretius, who lived from about 100-55 B.C. 

Aristotle (384-322 B.C.), a pupil of Plato, early abandoned 
the Platonic system, and founded the Peripatetic school, 
which rejected the Democritan atomic hypothesis and 
postulated the infinite divisibility of matter and the tians- 
mution of elementary substances. Owing to the tremen- 
dous influence of the Aristotelian school on contempoiary 
and subsequent thought, atomistic doctiines fell into dis- 
repute. Owing further to the fact that chemistry, in the 
early and middle Christian ages, was almost exclusively an 
Arabian art founded on the teaching of Aristotle, the 
atomic view of matter practically disappeared during the 
sixteen centuries from the time of Lucretius, until the 
Renaissance brought back the study of the pre-Christian 

The Democritan doctrine was first revived by Francis 
Bacon (1561-1626) in his Novum Organum, and Gassendi 
(1592-1655), in his strong opposition to the Aristoteliamsm 
of his time, brought forward and adapted the system of 
Epicurus. Despite a partial set-back, due to the con- 
tinuous-matter theory of Descartes, known as the Cartesian 
Philosophy, atomic views of matter became generally 
accepted. Robert Boyle (1627-1691), in his Sceptical 
Chymist and The Usefulness of Natural Philosophy, applied 
a corpuscular or atomic theory of matter to combination 

Atoms 17 

between substances to form other substances and to a 
dynamic explanation of gaseous piessure. Isaac Newton 
(1642-1727), in his Opticks and the Principia, referred 
chemical changes to atomic combinations, accepted and 
developed Boyle's explanation of the cause of gaseous 
pressure, and applied the atomic hypothesis to the forces 
of chemical affinity, gravitation, electricity, and magnetism. 

In 1776, Bryan Higgins, in his Philosophic Essay con- 
cerning Light, brought the atomic hypothesis more directly 
into relation with chemistry, by his suggestion that two 
diffeient atoms combine singly to form a compound, and 
William Higgins, in his Comparative View of the Phlogistic 
and. Antiphlogistic Theories with Inductions (1789), expanded 
this suggestion into a definite theory of combination 
between atoms in multiple proportions, the simplest and 
stablest combination occuirmg between two different 
atoms to form a binary compound. 

Though T. Nicholson had defined chemistry in 1795, in 
his Dictionary of Chemistry, as a science of the changes pro- 
duced in bodies by the movement of paits individually too 
minute to affect the senses, the atomic view of matter was 
merely a bare hypothesis, neither necessary nor even con- 
venient for an explanation of chemical facts. Indeed, at 
the commencement of the nineteenth century no single 
quantitative phenomenon, relating to the combination of 
substances either by weight or by volume, had been shown 
to be in accord with the necessities of an atomic hypothesis 
Numeious deteimmations had been made by weight and 
by volume of the proportions in which elements combine, 
but the results had not been applied to the support of the 
doctune of atoms, though the Higgmses had postulated a 
theoiy ol chemical combination atom to atom, which was 
cleaily susceptible ot quantitative investigation 

As early as 1630, Jean Rey, in an essay, Investigation, of 
the Cause of the Gain in Weight of lin and. Lead, on Calcina- 
tion, showed that such gain in weight never exceeded a 
certain limit. Hnmberg, in his Observations on the Quantity 

1 8 Chemist) y and Atomic Structure 

of Acids Absotbfd by the Alkaline Eatths (1669), determined 
the amounts of various acids lequired to " saturate " 
(neutralise) a fixed amount of " salt of tartar " (potassium 
caibonate), and his lesults may be regarded as the first 
steps towards the establishment of the Law of Equivalent 

Cavendish, in 1767, and Wenzel, in 1777, indicated that 
equivalency existed between the vaiious weights of metals 
and bases that neutralise a definite weight of any given acid, 
and, further, that this equivalency was independent of the 
particulai acid used. The extensive icsearches of Richtei, 
the originatoi of the distressful term " stoichiometiy " 
the quantitative relations between chemically leactmg sub- 
stances finally established in his New Aims of Cbemistiy 
(1792-1802), the truth of the foiegoing law, frequently 
called Richter's Law of Proportionality 01 of Equivalent 
01 Reciprocal Ratios or Proportions. This law states that 
the ratios between the weights of different substances that 
combine with a constant weight of another substance are 
either equal to the ratios of the weights of the substances 
in their combination with each other or are small integral 
multiples or simple submultiples of these ratios. Origin- 
ally applicable to the combination of acids with bases 01 
metals, it was proved by Beizelius 1 to be equally valid foi 
compounds and elements generally, and it is to-day, owing to 
the extraordmaiy lefinement of atomic weight methods, 
the most firmly established of the laws of chemistry. 

One of the outstanding difficulties in the way of the 
early general acceptance of Richter's Law lay in the well- 
known fact that some substances combine in several diffeient 
proportions with a constant weight of anothei substance. 
It was not even certain at the beginning of the nineteenth 
century that combination between substances did not take 
place in eveiy sou of continuously varying ratio. The 
controversy between Berthollet and Pi oust, as to the fixity 

1 Gilbert's Ann , iSu, 37, 248 and 415, 38, 161 and 227, and 1812, 40, 162 
a nd 235 

Atoms 19 

of composition of compounds, lasted from 1802 to 1808, 
and, though Proust's Law of Constant Composition was 

generally accepted and had in fact been tacitly assumed 
many years previously by Cavendish, Richter, and Lavoisier, 
conclusive experimental evidence foi it in any gieat detail 
was not available until Berzelius fiom 1810 onwards had 
determined the exact composition by weight of nearly all 
the chemical compounds known m his day. 

Richter had noticed in 1792 that a metal could form 
oxides with two different proportions of oxygen , Lavoisier 
was awaie of a number of elements which combined with 
another element in several drffeient propoitions , Caven- 
dish analysed three different oxides of nitrogen ; Proust in 
1799 analysed two difteient oxides of copper ; Clement and 
Desoimes discovered in 1801 that carbonic acid contained 
twice as much oxygen as caibonic oxide ; and Dalton in 
1802 showed that nitnc oxide combined with air in two 
pioportions, one double the oihei 

Whether or not Dalton was awaie of the suggestion of 
W. Higgins, that combination between atoms takes place in 
multiple propoitions, is not known with ceitainty, but in 
1803, fouiteen yeais aftei the publication of Higgms's 
suggestion, Dalton brought forward an almost identical 
theory of atomic combination and devised a system of 
atomic weights, entnely based on the assumption of com- 
bination between atoms in simple multiple pioportions. 
The broad outlines of Dalton's theoiy were first indicated 
in Thomson's System of Chemistty, published in 1807, and 
Dalton's celebrated work A New System of Chemical 
Philosophy appealed in 1808, and almost immediately the 
atomic theoiy met with general acceptance. 

The viitue of Dalton's theoiy was not that it was an 
atomic theory, foi theories of atoms are far older than the 
science of chemistiy, but that it represented the fiist 
definite attempt to place on a quantitative chemical 
footing the doctrine of combination between elementary 
corpuscles first laid down by Boyle and later by VV. Higgins. 

2o Chemistry and Atomic Structure 

Dalton's atoms, as we now know, were not real atoms, and 
more nearly coincided with the modem conception of mole- 
cules, and his so-called atomic weights were usually sub- 
multiples of real atomic weights, and were often in fact the 
modern chemical equivalent weights, which are related to 
the atomic weights by a simple multiple now called valency 
or the combining capacity of an atom measured in terms of 
hydrogen atoms or their equivalent. Dalton's outstanding 
achievement was his recognition of the Law of Simple 
Multiple Proportions by Weight, which states that the 
different weights of an element, in combination with a 
constant weight of another element, are small integral 
multiples of a common factor. 

It appears that Dalion had embraced the theoiy of 
atomic combination considerably prior to his iccognition 
of the experimental facts of combination of substances in 
simple multiple proportions by weight, and that his law as 
to such combination was deduced from the necessities of 
an atomic theoiy. It is certain that much of the evidence 
on which he relied to support his theory has since been 
shown to be untrustworthy or susceptible of other con- 
clusions He showed, foi example, that olefiant gas 
(ethylene) contains twice as gieat a proportion of carbon as 
marsh gas (methane). Had ethane been known in his day 
his theoiy would have had to be abandoned as soon as 
proposed, for ethane contains one and one-third times as 
much carbon as methane, to which Dalton assigned the 
formula CH 2 , and two-thirds of that in ethylene, to which 
Dalton assigned the formula CH. In benzene, C 6 H 6 , 
naphthalene, C 10 H 8 , anthracene, C 14 H 10 , and picene, 
C 22 H 14 , Dalton would have found compounds having 
weight ratios of hydrogen to carbon of I 5 0'8, 0-714, and 
o 836, n reconcilable with his law of simple multiple pio- 
portions by weight. It can be legaided only as remarkably 
fortunate that the few compounds known in Dalton's day 
were the simpler compounds of chemistry, to which alone 
the law applies. 

Atoms 21 

The two chief assumptions of Dalton's theory, that com- 
bination takes place in simple multiple proportions by 
weight and that the simplest and stablest compound is a 
binary combination of one atom with one other atom, are 
both unjustifiable. The first assumption is validly applic- 
able only to such compounds as aie not formed by combina- 
tion between atoms of the same element. The second 
assumption is also unwarranted, and in fact led to a con- 
fusion in the subsequent half a century almost unparalleled 
in the history of any science. 

As early as 1814 Wollaston clearly saw that it was 
impossible, in the then existing state of chemical knowledge, 
to ascertain, in the cases to which the law of simple multiple 
piopoitions by weight was applicable, which or if any com- 
pounds consisted of one atom of one element combined 
with one atom of another element. He proposed to discard 
Dalton's purely hypothetical atomic weights and to substi- 
tute equivalent weights, 1 being the experimentally ascer- 
tained combining weights of elements referred to a fixed 
weight of a standard element. Gmelm, in his Handbuch 
der Chemie (1817), was no less cleai as to the insecurity of 
Dalton's atomic weights, and decided in favour of equi- 
valent or combining weights, and Gmelin's system remained 
in more or less extended use on the Continent for nearly 
fifty years. 

Berzelius, probably the greatest experimental chemical 
genius of the nineteenth century, adopted Dalton's atomic 
theory, and proposed the present system of chemical 
nomenclature, 2 in which the initial letter or the initial 
letter and another are used to symbolise elements. Owing 
to the defect in Dalton's theory in that no cei tain criterion 
existed for determining the real relative weights of atoms, 
Berzelius was frequently compelled to alter his scheme of 
atomic weights, and thereby undermined the foundations 

1 A Synoptic Scale of Chemical Equivalents, Phil Tran* R Soc , 1814, 104, 

P J 
z journ. Phys., 1811, 73, 257, and Larbok t Kemten, 1811 

22 Chemistry and Atomic Structure 

of the atomic theoiy they were designed to suppoit to such 
an extent that by about 1850 atomic theories vveie vntually 

Though the ultimate failuie of Dalton's theory was due 
to its inability to define the atoms it postulated, neveithe- 
less the theory contained not only the geim of the modern 
theory of the atom but indeed foreshadowed that atoms 
have definite structmal parts, and it is remarkable that this 
inference from the theoiy should have remained unrecog- 
nised for three-quaiters of a centuiy, and particularly after 
the development of the theory of chemical valency from 
1852 onwards. A single example is sufficient to illustrate 
the point The element manganese forms the following 
oxides, manganous oxide, manganic oxide, manganous 
anhydiide, manganic anhydride, and permanganic anhy- 
dride, in which 100 parts by weight of manganese are com- 
bined with 29, 431, 58, 87, and ioi parts by weight of 
oxygen respectively. These weights of oxygen are the 
simple multiples, 2, 3, 4, 6, and 7, of the common factor 14^- 
Assuming, as Dalton would have assumed, that the first 
oxide contains one atom of manganese, these oxides would 
have the formulas MnO 2 , MnO 3 , Mn0 4 , MnO c , and 
Mn0 7 , identical with the modern formulas if " " be 
taken as a half-atom of oxygen, i.e. a hydrogen equivalent 
of oxygen. The multiples express precisely the various 
valencies or combining capacities of the manganese atom 
measured in terms of atoms of hydrogen or their equivalent, 
and aie unequivocal evidence of the active participation of 
2, 3, 4, 6, and 7 of the manganese electrons m binding 
oxygen atoms, It thus occurs that Dalton's atomic theory, 
detective though it pioved, piesaged the ultimate structuie 
of the real atoms it failed to substantiate. 



IF all atoms have equal combining capacities, measured in 
terms of atoms of any element taken as a standard, atomic 
weights and equivalent (combining) weights are identical ; 
but if atoms have difterent combining capacities, atomic 
weights can not be determined solely from considerations 
of equivalent weights. The cardinal defect of Dalton's 
atomic theory was its failure to provide any means of 
ascertaining the combining capacities of different atoms. 
Dalton assumed that the combining capacities of hydrogen, 
carbon, nitrogen, and oxygen were identical, and that the 
composition of the ultimate pai tides of hydrogen, ethylene, 
ammonia, water, carbonic oxide, and nitric oxide was 
correctly expiessed by the foimulae H, CH, NH, OH, CO, 
and NO, whereas they are in fact H 2 , C 2 H 4 , NH 3 , H 2 0, 
CO, and NO, respectively. A decision, as to which of these 
two seiies of formulae is correct, can however leadily be 
made, if the lelative weights of the pai tides and the formula 
of any one are known. General agreement, as to the com- 
position of the hydrogen particle or molecule, and as to 
methods for determining the molecular weights of elemen- 
tary and compound substances, was not reached for over 
half a century after Dalton's theory was put forward, but 
this half-centuiy of confusion sufficed to elucidate that 
atoms have a definite numerical valency, saturation capacity 
or combining power measured in terms of hydrogen atoms 
or their equivalent. 

In 1805, Gay-Lussac and Humboldt x discovered that 
oxygen and hydrogen combine to form water in the pro- 
portion of two of hydrogen to one of oxygen by volume, 
and in 1808 Gay-Lussac, 2 as the result of further experi- 
ments, put forward the Law of Simple Multiple Propor- 
tions by Volume, which, in modem terms, states that the 
several gaseous volumes, measured under standard condi- 

l journ Phys , 1805, 60, 129 

2 M6m ( Jnc Arcued, iSoq, 2, 207. 

24 Chemistry and Atomic Structure 

tions, of the substances taking part in a chemical change 
and the total change in volumes, if any, are small integral 
multiples either of the smallest or of a common factor of 
these volumes. This law, usually referred to as Gay- 
Lussac's Law or the Law of Gaseous Volumes, was strongly 
opposed by Dalton in 1810, who imagined it was uncon- 
formable with his atomic theory. Dalton, however, 
admitted that the law and the atomic theory weie com- 
patible, if " all elastic fluids (gases) have the same number of 
atoms in the same volume" but regarded such a hypothesis 
as untenable. 

The hypothesis abandoned by Dalton was again pro- 
pounded in 1811 by Avogadro, 1 who showed clearly that 
Gay-Lussac's Law was theieby readily explicable. Avo- 
gadro's Hypothesis states that equal volumes of gases 
under the same conditions of temperature and pressure 
contain equal numbers of independent particles (mole- 
cules) . He inferred from the hypothesis that the ratio of 
the masses of equal volumes of gases is the ratio of the 
masses of their molecules, 01, in other woids, that the 
ratio of vapour densities is the ratio of molecular weights. 
He further showed that, by means of the hypothesis, 
Dalton's arbitrary suppositions as to the relative number of 
atoms in compounds could be rectified or confirmed, and 
that the so-called atoms of hydrogen, nitrogen, and oxygen 
must "each consist of two half -molecules. Avogadro's argu- 
ment was that, as the volume of gaseous water formed by 
combining two volumes of hydrogen with one volume of 
oxygen was double that of the oxygen, the molecule of 
oxygen must be double. Similarly, he aigued that, as the 
volume of ammonia was double that of the nitiogen from 
which it could be formed, the nitrogen molecule must be 
double, and, further, that, as three volumes of hydiogen 
yield only two volumes of ammonia, the molecule of 
hydrogen must be double, i.e. that six half-molecules of 
hydrogen must have been present in the original thiee 

1 Jomn P&y>, 1811, 73, 58 

Atoms and Molecules 25 

molecules. The importance of Avogadro's hypothesis was, 
however, not realised by chemists, and for the next foity- 
nme yeais no ceitain means existed of determining atomic 

Ampere in 1814 1 revived Avogadro's hypothesis, but 
failed to make it acceptable. A similar fate befell Dumas' 
attempts 2 to establish a system of atomic weights based on 
determinations of vapour densities. 

In 1819, Dulong and Petit, as a result of investigations of 
the relations between atomic weights and physical pro- 
perties, proposed 8 the Law of Atomic Heat, which states 
that the heat required to raise the temperature of a 
weight of any element, equal to the atomic weight in 
grams, through i C. is constant for all elements, or that 
the atomic heats of different elements are equal. This 
law is, however, only approximately tiue, the constant 
(6 4 calones) varying for difreient elements fiom less than 
2 to more than 9 caloiies. It is available, in the absence 
of means of determining moleculai weights, for fixing 
atomic weights, and serves as a useful check on atomic 
weight data. The investigations of Neumann in 1831, and 
of Regnault in 1841, enabled Joule in 1844 to piopose the 
Law of Molecular Heat, 4 which states that the atomic 
heats of elements are unchanged in their compounds. 
The exceptions to the heat laws are not only numerous but 
very marked, discrepant atomic heats, for example, being 
almost invariably low It has been frequently suggested 
that the atomic heats of the discrepant elements are more 
closely in agreement with the law at very high tempera- 
tures, the atomic heat of carbon, for example, increasing 
fiom i 8 at 20 to 6-0 at about 2000. This suggestion is 
invalidated by the fact that nearly all elements having 
normal atomic heats, i e about 6-0, have atomic heats as 
high as 9 to 10 at the temperatuie at which caibon has 6-0. 

1 Ann Chun Pbys , 1814, [i], 90, ft 

i Ibid , 1826, [2], 33, 337, and 183^, [2], 50, 170 

3 Ibid , 1819, [2], 19, 350, 

/V>i/ Mag, 1844, [3], 25, 334 

26 Chemistty and Atomic Stiuctwe 

Moreover, as all specific heats tend to zero as zero absolute 
temperature is approached, atomic heats also tend to zeio. 
The fact is that atomic heats are not constant, but that the 
value 6-4 calories is close to the atomic heats of many 

Atomic weights are occasionally determinable on analo- 
gies inferred fiom crystalline structure. In 1820, Mitscher- 
lich proposed the Law of Isomorphism, which states that 
" the same number of atoms combined in the same manner 
produce the same crystalline fo?m ; the crystalline form is 
independent of the chemical natwe of the atoms, and is deter- 
mined solely by theii numlet and mode of combination" If 
this law were rigorously true it would be of service in 
detei mining the combining capacity or valency of elemen- 
tal y atoms by analogy with similai and isomorphous com- 
pounds containing elements of known valency. The law is, 
however, of very limited application, as the number of 
exceptions to it aie numerous, and as no such phenomenon 
as exact isomorphism exists even in the case of compounds 
of elements chemically most closely related, largely due to 
the non-identity of atomic volumes of any atoms. 

Berzelius, who determined the atomic weights of almost 
the whole of the elements known in his day and the exact 
composition of the majoiity of their compounds, was 
guided, in his decisions as to the numerical relation between 
equivalent weight and atomic weight, by geneial con- 
siderations based on the well-asceitained reactivities of 
similar elements in analogous compounds, by Gay-Lussac's 
law of gaseous volumes in so far as he applied Avogadro's 
hypothesis to elementary but not to compound gases and 
vapours, by Dulong and Petit's law of atomic heat, and 
by Mitscherlich's law of isomorphism. That his atomic 
weights were frequently amended, by multiples 01 sub- 
multiples, and were finally often in erroi, is to be attributed 
more to his failuie to accept Avogadio's hypothesis for 
compounds as well as for elements than to the inadequacy 
of contemporary knowledge. This failure undoubtedly 

Atoms and Molecules 27 

held back the development of chemistry for at least a 
generation, as is evident from the fact that, within ten 
years of the regularising of atomic weights consequent on 
Canmzzaio's revival and the immediate general acceptance 
of Avogadro's hypothesis, the obscure facts of valency were 
disentangled and the periodic classification propounded. 

In 1858, Cannizzaro published his Sketch of a Couise of 
Chemical Philosophy., 1 by which he firmly established 
Avogadro's hypothesis and showed that all the known 
physical and chemical facts as to gaseous bodies confirmed 
its validity, and that the distinction between atoms and 
molecules reconciled all the contradictory experimental 
results accumulated in the preceding half century. Canniz- 
zaro attempted to combine the laws of simple multiple 
pioportions by weight and by volume into one law, that 
the vaiious weights of the same element contained in equal 
volumes either of the fiee element or of its compounds aie 
all whole multiples of one and the same weight, that of the 
atom. This, however, is not a law but merely a definition 
of atomic weight, and quite overlooks the important point 
in Dalton's law of simple multiple proportions, in that the 
latter deals with the multiples of a weight which is fre- 
quently less than the atomic weight, and always less than 
the atomic weight in the cases of elements having com- 
bining capacity for hydrogen atoms, i e. valency, greater 
than unity. Canmzzaio emphasised that what enters into 
chemical reactions is the half-molecule of hydrogen, which 
is " indivisible, at least in the sphere of chemical actions 
actually known" and deliberately left open the question as 
to whether or not it was physically possible to divide atoms 
further, a significant disclaimei in view of modem theories 
of the many-electroned atom and its composite nucleus. 

It is a common delusion of waters in the journals of 
popular science no less than in the daily press that the dis- 
covery of the mobile electron and the disintegrate atomic 
nucleus deployed the foundations of chemistry, a science 

1 // Nuavo Cunento, 1858, 7, 321 , Alembtc Club Reprint, No 16 

28 Cbemistty and dtomic Structure 

of indivisible atoms. This delusion lias its oiigin in a mis- 
apprehension of what chemists mean by an atom. The 
integrity of the chemical atom has never been so certain as 
since it was proved to be physically divisible, and, even had 
the atom been pioved in physics to have no real existence 
whatevei, the atom of chemistry would remain, for it is 
the name by which one of the real weight relations between 
different kinds of matter is expeiimentally identifiable. 
The precise meaning attached in chemistry to the term 
atom is involved in the definition, that the Atomic Weight 
of an element is a number equal to thirty-two times the 
ratio of the smallest weight of it, ever found in a gaseous 
volume of it or of any of its compounds, to the weight of 
an equal gaseous volume of oxygen at the same tempera- 
ture and pressure. This number is veiy nearly identical 
with twice the coi responding ratio referred to hydiogen as 
the standaid. From this definition it is obvious that an 
atomic weight is an experimentally determinable number 
which is totally independent oi the reality of the existence 
of physical atoms. Atomic weights are not measureable 
in grams or pounds, but aie mere numbers, expressing the 
ratio of one weight to a specifically selected weight of an 
arbitrary standaid. Cannizzaro chose hydiogen as the 
standard element, but, for reasons of accessibility of 
measurement, the modern standard element is oxygen. 

The definition of atomic weight is so framed that even 
the atomic weight of the standaid element can be experi- 
mentally detei mined. The weights of equal volumes of 
gaseous water and gaseous oxygen are in the ratio 9 to 16, 
and 9 parts by weight of water contain 8 parts by weight of 
oxygen, and, by definition, the atomic weight of oxygen in 
water is 32 times the ratio A, i.e. 16 As this is the smallest 
atomic weight ever found in an oxygen compound, the 
atomic weight of oxygen is 16. The atomic weight of 
hydrogen in water is 32 times the ratio ^, i.e. 2, deduced 
from the fact that I part by weight of hydrogen is contained 
in 9 parts by weight of water, and 16 paits by weight of 

Atoms and Molecules 29 

gaseous oxygen are contained in a volume equal to that 
of 9 parts by weight of gaseous water. This atomic weight 
of hydrogen is not the leal atomic zueight, for compounds of 
hydiogen are known, fiom which an atomic weight about 
half of the foregoing is obtained, I -008 to be exact. 

As equal gaseous volumes of oxygen and water have 
weights in the latio of 9 to 16, and water contains eight- 
ninths of its weight of oxygen, the ratio of the oxygen in 
water to the oxygen in gaseous oxygen is I to 2 by weight. 
By Avogadro's hypothesis the ratio of the weights of equal 
volumes of gases is the ratio of the weights of the particles 
of which the gases aie composed, i.e. the weights of the 
molecules. As the weight of oxygen in gaseous oxygen is 
twice that in an equal gaseous volume of water, the mole- 
cular weight of gaseous oxygen is 32, i.e. twice the atomic 
weight of oxygen deduced from the composition of water. 
In general the Molecular Weight of an element or com- 
pound is a number equal to thirty-two times the ratio of 
the weight of a gaseous volume of the element or com- 
pound to the weight of an equal volume of oxygen at the 
same temperature and pressure. Obviously, if the mole- 
cular weight be so defined, the Atomic Weight of an 
element is the least weight of it ever found in the mole- 
cular weight of it or of any of its compounds. In the 
case of many elements, the molecular weight is not identical 
with the atomic weight, but a simple multiple of it. This 
multiple, known by the term "Atomicity," defines the 
number of atoms in the elementary molecule. Originally 
this term was applied to the combining capacity of an 
element measuied in atoms of hydrogen, but the term 
valency is now used for combining capacity, and atomicity 
applied only to the composition of the molecules of elements. 

A definition, common in text-books of chemistry, is that 
the atom or the molecule represents the least weight of a 
substance that can take pait in a chemical change. This 
definition is untrue for many leasons. The " least weight " 
must be that of the real atom, which, in disease of hydro- 

-? ^ 

3o Chemistry and Atomic S tincture 

gen, is so small that about six hunched thousand million 
billions are lequired to weigh one giam. Apait from 
absolute weights, the least weight of an element that can 
be discerned in a chemical change is not necessarily the 
molecular weight or the atomic weight, but usually a 
smaller weight termed the equivalent weight, and equal to 
the atomic weight and molecular weight only in the case 
of the alkali metals, which have invanable unit valency. 
The Equivalent Weight of an element or compound is 
defined as a number equal to eight times the ratio of the 
weight of it to the weight of oxygen with which it com- 
bines or which it can displace from combination. As 
the ratio of the weights of hydiogen and oxygen in water 
is one-eighth, one equivalent weight of hydiogen is practi- 
cally equal to unity, hence is derived an alternative defini- 
tion, that the equivalent weight of a substance is the 
weight that combines with or displaces 4?om combination 
one part by weight of hydrogen. The atomic weight of 
hydiogen being very nearly unity, the equivalent weight 
represents the weight that combines with or displaces from 
combination one atom of hydiogen, and the atomic weight 
of oxygen being 16 the equivalent weight of oxygen is half 
an atom, i.e. 8. The equivalent weight of water, H 2 O, is 
the half-molecule of weight 9. The equivalent weight of 
orthophosphoric acid, H 3 PO 4 , is a third of the molecule, 
of sulphuric acid, H 2 SO 4 , a half -molecule, and of hydro- 
chloric acid, HC1, a whole molecule. In the case of acids 
the ratio of the molecular weight to the equivalent 
weight is termed the " Basicity " of the acid, mono-basic, 
dibasic, tribasic, and tetrabasic acids containing one, two, 
thiee, and four atoms of hydrogen, respectively, in the 
molecular weight. The equivalent weight of the base 
caustic soda, NaOH, is the whole molecule for it combines 
with one molecule of a monobasic acid, of the base lime, 
CaO, the half-molecule, for it combines with one molecule 
of a dibasic acid, and of the base alumina, A1(OH) 3 , one- 
third of the molecule, for it combines with one molecule 

Atoms and Molecules 31 

of a tnbasic acid. In the case of bases, the ratio of the 
molecular weight to the equivalent weight is termed the 
" Acidity " of the base, and the molecules of mono-acidic, 
di-acidic, tn-acidic, and tetra-acidic bases lespectively com- 
bine with one, two, three, and four atoms of hydrogen of 

The equivalent weight in the case of an element is not 
necessarily a constant, manganese, for example, forming at 
least five different compounds with oxygen. The various 
equivalent weights in these oxides range from 27-5 to 7*85- 
The oxide with the smallest equivalent dissolves in water 
to form an acid in which the ratio of hydrogen to manganese 
is i to 55, and as the acid cannot contain less than an atom 
each of hydiogen and manganese, the atomic weight of the 
lattei cannot be gt eater than 55 Only a knowledge of 
the moleculai weights of all manganese compounds can 
suffice to detcimine the actual atomic weight The mole- 
culai weights of many compounds cannot be obtained, 
howevei, owing to the impossibility of obtaining them m 
the gaseous condition. 

Investigation of the propeities of solutions of substances 
in liquids has disclosed the fact that an internal pressure, 
called the Osmotic Pressure, is set up in dilute solutions, 
such that this pressure is approximately equal to that 
exertable by the same weight of the dissolved substance 
if it existed as a gas and occupied the same volume as 
that of the solution. Companson of the osmotic pies- 
sures of different substances enables lelative molecular 
weights to be detei mined by icference to propeities of 
the solutions which aie dependent on the magnitudes of 
the osmotic piessures Such piopeities are lowering of the 
vapour pressure of the solvent, elevation of its boiling point, 
depression of its freezing point, and lowering of its solu- 
bility foi othei substances. Determinations of molecular 
weights based on the propeities of solutions are seldom 
more than approximately exact, but the results usually 
serve to decide which multiple of the equivalent weight is 

32 Chemistry and Atomic Stmcture 

the true molecular weight, and hence enable atomic weights 
to be fixed in cases where the vapour density method is 

It has often been stated that atomic weights are matters 
of essentially minor importance, and that past generations 
of chemists have wasted incalculably valuable time, energy, 
and skill, in the determination of a mere number, a relative 
weight, which, as some modem physicists have declared, is 
not even a characteristic property of any sort of matter. 
Such statements can be viewed by chemists only as pathetic 
confessions of ignorance. The weight properties of matter 
aie in fact almost the only propeities of matter that have 
enabled the innermost secrets of nature to be unveiled. 
The ratio of atomic weight to equivalent weight is the sole 
determinant of the combining capacity or valency of atoms, 
and the whole of modern knowledge of the complex struc- 
tures of organic and inorganic chemistry rests and abides 
in valency. Without a knowledge of atomic weights and 
valency the classification of the elements in terms of 
periodic properties is inconceivable, and without the 
periodic classification there could have been no hope of 
the explanation of radioactivity, or of the arrangement of 
numerous electrons into the ordered systems of piesent-day 
physical and chemical theories of atomic structure. 

The two factors which are stated to render ordinary 
atomic weights unreal, are the existence of Isotopes (iso- 
meric atoms identical in all properties except mass) and 
the circumstance of the variation of all mass with velocity. 
Whatever the facts may be as to the existence of isotopes, 
there is no doubt that throughout nature atomic weights 
are constant, and isotopic atoms must theiefore be present 
in a constant proportion in any directly weighable amount 
of any element. Consequently atomic weights are real if 
only real averages, and their validity in chemistry is inde- 
pendent of the existence of isotopes. The minute variation 
of mass with velocity is based on known variation with 
velocity in the value of the ratio of electric charge to mass 

Atoms and Molecules 33 

Obviously one of the factois of the ratio must be variable, 
but it is a mere hypothesis that the variable factor is mass 
it may in fact be electric charge, and it is certain that 
the differences in the values of the unit electric charge, 
that of the ion or electron, are greater than the possible 
differences due to variation with velocity, so that the 
hypothesis of the constancy of electric chaige with variable 
velocity has no evidence to support it, and, it may be 
argued, is not a fundamentally necessaiy hypothesis for the 
explanation of the variation of the ratio of charge to mass. 
Even should the variation in the charge to mass ratio be 
ultimately demonstrated to be wholly due to variation in 
mass, this variation is practically zero at all velocities much 
less than that of light, and is extraordinarily minute even 
with attainable velocities comparable with that of light. 
Any possible variations in mass that could occur in any 
chemical reaction, conducted on substances at any practic- 
able velocity, would he far outside any hope of detection 
by any chemical methods of direct weighing, and conse- 
quently the variation in mass with velocity has no real 
bearing on the chemical validity of atomic weights. 



THE mechanism of chemical combination is the root 
problem of chemistry, and theones regarding it were pro- 
pounded even prior to the chemical theoiy of atoms In 
1787, Moiveau suggested that compounds consisted of 
oxygen combined with a base or radical, 1 and Lavoisier 2 
regaided all compounds as oxides, and suggested that 
inoiganic bodies were oxides of simple substances and 
organic bodies oxides of complexes 01 radicals Lavoisier's 
theory was essentially dualistic, for he divided the oxides 
into two contrasting classes, those that were derived from 
acidic and those from sahfiable radicals or bases. To 
Lavoisier, all radicals were bases, but the term " bases " 
was later restncted to the oxides of metals, acid ladicals 
being those whose oxides gave rise to acids. 

In 1815, Gay-Lussac 3 showed that cyanogen was a nevei- 
vaiymg constituent group or ladical of a series of com- 
pounds, and was chemically equivalent to the simple acidic 
radicals (elements). 

In 1820, Dalton discovered a hydrocaibon in oil-gas 
having the same composition as ethylene, and it was later 
suggested that it consisted of two ethylene particles. The 
composition of this hydrocarbon, now known as butylene, 
was confiimed by Faraday in 1825,* w ^ showed that its 
vapour density was twice that of ethylene, and at the same 
time announced the discovery of another hydiocarbon, 
benzene. Faraday called attention to a number of existing 
cases of substances having the same composition but 
different properties, and cited Liebig's discovery of 1823 
that the latter's silver fulminate and Wohlci's silver cyanate 
had the same composition, 5 and Gay-Lussac's suggestion 

i Lavoisier, Morveau, and Fourcroy, Mitbode de Nomenclature, 1787 
~ Iran's eUmentaire de Chimie, 1789 and 1793 
s Ann Chim Pbys , 1815, [i], 95, 136 

4 Phil Trans Roy Sac, 1825, 461. 

5 Liehg, Gilbert's Ann , 1823, 75, 393 , Ann Cbim Pbys , 1823. [2], 24, 204 
and Gay-Lussac and Licbig, ibid,, 1824, [2], 25, 285 

Valency 35 

that the difference in pioperties of substances with the 
same composition must be due to differences in the arrange- 
ments of the same atoms. 

Wohler's discovery in 1828* that the salt, ammonium 
cyanate, is converted, by simply heating in aqueous solu- 
tion, into the organic substance, urea, having the same 
qualitative and quantitative composition, lent additional 
weight to Gay-Lussac's suggestion. In 1831, Berzelius 2 
showed that racemic acid, discovered in *' tartar " (wine- 
lees residues) by Gay-Lussac, was identical in chemical 
composition and many chemical properties with ordinary 
tartaric acid, but differed from it in solubility and in the 
crystalline form of its salts. To express this difference in pro- 
perties of substances with identical composition, Berzelius 
proposed the term " Isomensm. " He further indicated 
that Mitscherlich's law of isomorphism must be amended 
in a new direction, in that the same atoms may be arranged 
in chemically very similar substances so as to produce 
different crystalline foims. The demonstration of the 
existence of isomenc bodies involved the lecognition of the 
fact that the atoms in a compound were not combined 
togethei " each to all and all to each" but that specific 
atoms must be diffei entry combined together. From this 
time chemistry acquhed a new outlook, and the important 
problem of the science became the determination of the 
details of the combination between atoms. In the subse- 
quent theories of radicals, nuclei, substitutions, and types, 
chemical conceptions of the molecule with definite struc- 
ture and of the atom with delimited and directional com- 
bining capacity slowly clanfied and finally crystallised into 
the modern theories of moleculai stiucture and atomic 

In 1827, Dumas and Boullay 3 propounded the theory 
that ether and the simple derivatives of ethylene all 

1 Fogg Ann , 1828, 12, 258 

2 Jabresber , 1832, 11, 44, and 12, 63 

3 Ann Cfam Pbys , 1828, [2], 36, 294, and 1828, [2], 37, 15 

36 Chemist) y and Atomic Sttnctwe 

contained the ethylene group, C 4 H 8 , for which Beizelius 
proposed the name etheiin Liebig, howevci, adopting the 
dualistic view, regarded ether as the oxide of the ladical, 
C 4 H 10 , which he called etheryl 01 ethyl. In 1832, Wohler 
and Liebig x published their icsearches on the " Radical of 
Benzoic Acid" and descubed a series of closely i elated, 
interconvertible compounds all containing the radical 

In 1834, Dumas, as the result of his researches on the 
action of chloime on alcohol, 2 put forward his " Substitu- 
tion Theory," in which he postulated that hydiogen in 
01 game compounds could be replaced atom for atom by 
chlorine, bromine, and iodine, and the equivalent of 
oxygen, without very material alteiation in the nature of 
the resultant products. This theoiy was expanded by 
Lament 3 into the "Nucleus Theory," in which the 
carbon atoms in compounds were legarded as foiming an 
invariable nucleus of definite shape, the various atoms lound 
the nucleus being substrtutable, equivalent for equivalent, 
without material change in nature of the compounds. 

In 1837, Liebig, 4 and Dumas and Liebig, 5 put forwaid 
their "Theory of Radicals," now called the "Older 
Theory of Radicals, ' ' in which complex groups of atoms 01 
radicals were assumed to exist in organic compounds, and 
to be transferable unchanged from compound to compound 
like the elementary atoms of inorganic compounds. Ger- 
hardt, however, rejected the theory of unchangeable 
radicals and proposed a " Theory of Residues " in which 
he regarded an organic molecule as a single stiuctuie, not 
a binary system of radicals. To Gerhaidt, radicals were 
merely the portions, residues, that took no pait in a paiti- 
cular chemical reaction, and appeared as independent 

1 Ann , 1832, 3, 249 

- Ann. Cbim, Pbys , 1834, [2], 56, 113 and 140 

*Ibid } 1835, [a], 60, 220, 1836, [2], 61, 125, and 63, 27, 42, 207, 
and 377 

4 Ann ,1838, 25, 3 

6 Compt. rend , 1837, 5, 567, and J. prakt. Chem., 1838, 14, 298 
Ann, Chm. Pbys. t 1839, OL 72, 184 

Valency 37 

poitions of molecules only when displaced in a reaction by 
combination between the reacting poitions of molecules, 
such residues or unreactive remnants then joining together 
to foim a copulated or conjugated compound. 

In 1840, Dumas incoiporated his theoiies of radicals and 
substitutions, together with portions of Laurent's nucleus 
theoiy and Gerhardt's theoiy of icsidues, into a compre- 
hensive "Theory of Types" 1 now refeired to as the 
" Older Theory of Types." Dumas legarded substances, 
with similar propeities and analogous composition, as 
belonging to one chemical type, for example, acetic acid 
and tnchloi acetic acid ; and substances, with different 
piopeities but simply dcnvable fiom one another, as 
belonging to the same mechanical or mole tula; type, for 
example, alcohol and acetic acid. He specially emphasised 
the nnpoitance of the arrangements of atoms and 
radicals in the molecule as determinants of chemical 

Beizehus, realising the incompatibility of his dualistic 
theoiy of electro-chemical combination 2 with the substi- 
tution theoiy, attempted to explain the facts of equivalent 
substitution by postulating imagmaiy ladicals, trichlor- 
acetic acid, for example, being regaided as a copulated 
compound of caibon chlonde, C 2 C1 6 , and the caibon oxide, 
C 2 O 3 (oxalic acid), together with water. This explanation 
bioke down when it was shown that on i eduction tiichlor- 
acetic acid yielded only hydrochloric and acetic acids and 
no oxalic acid, and, though Berzehus was unconvinced and 
continued to oppose the type theoiy till his death in 1848, 
his dualistic theoiy, once the most important and widely 
accepted in Euiope, was almost completely abandoned by 

In 1808, Thomson (T.) had shown that oxalic acid com- 
bined with both potash and strontia to yield two series of 
compounds in which one contained twice as much of the 

1 Ann Cbim Phys , 1840, [z], 73, 73, 113, and 74, 5. 

2 See Chap IV, p. 48 

38 Chemistry and Atomic Sttuctwe 

bases as the othei, 1 and in the same year Wollaston 2 showed 
that carbonic and sulphunc acids had saturating capacities 
foi bases similar to oxalic acid. In 1833, Graham 3 proved 
that phosphoiic acid, PO 5 , unites with one, two, or thiee 
equivalents of water, OH, to form separate acids, HPO C , 
H 2 P0 7 , and H 3 P0 8 , (HP0 3 , H 4 P 2 O 7 , and H 3 P0 4 , _in 
modem formulae), charactensed by their ability to combine 
with one, two, or three equivalents, icspectively, of bases. 
It followed, consequently, that equal numbers o molecules 
of acids (acidic oxides) and bases (basic oxides) did not 
necessaiily neutralise one anothei. This contioveited the 
lule set up by Berzelius that the moleculai weight of an 
acid was equal to the lelative weight lequired to neutralise 
the equivalent weight of a base In 1838, Liebig published 
his " Theory of Polybasic Acids "* and showed clearly 
that, befoie the molecular weight of an acid could be 
deteimmed, it was necessary to know its saturation 
capacity for bases, i.e. its basicity. Liebig's interpretation 
that acids are salts of hydrogen (based on an earlier sugges- 
tion of Davy, who argued that hydrochloiic acid was an 
acid though it contained no oxygen), involved that the 
hydrogen atoms, in a radical combined with oxygen, weic 
difteientiable into two classes accoidmg as the hydiogen 
atoms were or were not leplaceable by metals to give salts, 
thus elucidating part of the structure of carbon-hydiogen 
radicals. Gerhardt incorporated Liebig's theoiy of basicity 
into his " Theory of Residues " (seep 36) and showed that 
the basicity of a copulated 01 conjugated compound was 
always one less than that of its constituents, wow-basic 
benzene, for example, and rftbasic sulphuric acid foiming 
monobasic benzene sulphonic acid. 

In 1844, Mitscheihch showed that, though solutions of 
salts of ordinal y tartaric acid rotated the plane of polarisa- 
tion of light whereas those of racemic or paratartaric acid 

1 Thomson, Phil. Trans Roy Soc., 1808, 98, 63 

2 Ibtd , 98, 96. 
*Ibid, 1833, 123, 253 
4 Ann , 1838, 26j 113. 

Valency 39 

were optically inactive, yet the salts exhibited the same 
ciystallme form. Pasteur, in 1848, however, found on 
further examination, that crystals of salts of lacemic acid 
were not truly uniform but were composed of two sorts, 
one identical in all chemical and optical properties with the 
salts of ordmaiy tartaric acid, the other being also identical 
except that solutions rotated the plane of polarisation of 
light in the reverse dnection. This opposition in the 
optical properties was shown to be parallel with the develop- 
ment on one of the sorts of ciystals of small facets (hemi- 
hedial facets) related to the corresponding facets on the 
other sort as object to non-superposable minor-image. 1 
This discoveiy led later to the theory of the tettabedtal 
carbon atom and the asymmetiy of organic molecules, and, 
in the hands of Werner, led to the brilliant elucidation of 
the constitution of complex inoiganic molecules, and, 
infeientially, to a decisive knowledge of the exterior struc- 
ture of the atoms of many elements. 

In 1849, Wuitz discovered the oiganic amines or substi- 
tuted ammonias, 2 and proposed that these bodies all be 
refened to the general type, ammonia, NH 3 , by replace- 
ment of one or more hydiogen atoms by radicals. Wurtz's 
suggestion of the ammonia type was adopted by Hofmann, 3 
who had prepared a similar seiies of substituted ammonias. 

In 1850, Williamson, as a icsult of his researches on 
" Etbetification" 4 suggested that all compounds could 
be refened to the water type, H 2 O, or to the type of 
condensed molecules of water (H 2 0) n . 

In 1851, Kolbe put forward his " New Theory of 
Radicals," supported by his work on the free ladicals 
obtained by the electrolysis of oiganic acids, and by his 
joint researches with Fiankland on the organometalhc 
bodies, which contained leadily identifiable hydrocaibon 

1 Compt rend , 1848, 26, 535, and 27, 101 , and 1849, 28 J 22 3 

2 Ibid , 1849, 28, 223, and Ann Cbim Pby* , 1850, [3], 30, 498. 

3 Ann , 1850, 73, 91, 74, 174, and 1851, 79, 16 

1 Bnt Assoc Rep , 1850, 2, 65 , Phil Mag , 1850, 37, 350 , P/OL Roy Insln , 
1851-1854, 1, 90, and 239 , J Chcm Soc , 1852, 4, 229, 239, and 350 

/j.o Chemist) y find Atomic Sttuctiue 

radicals direct!/ combined with metals Kolbe and Frank 
land regaided numerous oigamc substances simply a 
derivatives of inorganic substances, fioin which they coulc 
often be obtained by simple substitution processes, anc 
suggested that all oigamc bodies could be regaided a 
derived from carbonic acid 1 

In 1852, Franldand (E.) published his classical icsearcl 
on A New Series of Oigamc Compounds containing Metals,' 
and showed that the power of metals to combine wai 
reduced by copulation with oigamc ladicals in organo- 
metallic compounds, in the same way that acids are reduced in 
basicity by copulation (Geihaidt's Theoiy of Residues, p. 38) 
This i eduction in power of combination on copulation was 
shown by Frankland to vary dnectly with the numbci oi 
ladicals copulated to the metallic atom in the same way 
and to the same extent as partial combination with oxygen 
reduced the powei of combination with other atoms, and 
he theiefore regarded combination with radicals as being 
in no way different from oiclmary inoiganic combination 
The views put forward by Franldand as to the definite 
combining power of individual atoms constituted the first 
direct explanation of Dalton's law of simple multiple pio- 
portions and Liebig's and Geihaidt's rules as to the basicity 
of acids and radicals, and represent the first definite steps 
towards the establishment of the theoiy of Valency. The 
fact that Franldand regaided the oxygen atom, in Gmclm's 
notation, as equivalent to an atom of hydiogen, in no way 
detracts from his recognition of the definite combining 
capacity of atoms or valency. The pievailing confusion in 
atomic weights, however, obscuicd the importance of his 
contribution to chemical theory Neveitheless, his method 
of formulation of combining capacity is not merely strictly 
admissible but necessary, foi valency is deleimmable not in 
terms of any atoms but in teims of the equivalents of atoms, 

1 Uelei die cbemiscbe Konstitution mid Natur der orgamschen Radikale, 1851 , 
/f, 1849, 89, 257,71, 171, 1850,75,211, 76,i, 1857,101,257, and 1860, 
113, 293 

2 Phil Tra-ns Roy Soc , 1852,142,417. 

Valency 41 

i c. in terms of hydrogen atoms with unit combining 

In 1853, Gerhardt put forward his " Theory of Simple 
Types M1 in which he included Wuitz's ammonia type and 
Williamson's water type in a scheme of foimulation of 
organic compounds, such that all were regarded as derived 
from the four simple inorganic types, hydrogen, H 2 , hydro- 
chloric acid, HC1, water, H 2 0, and ammonia, NH 3 . He 
further classified radicals, according to their " atomicity," 
or their capability of leplacing one, two, or three atoms of 
hydrogen in the type fiom which their compounds were 
denved. Wuitz 2 similarly classified ladicals according to 
their " basicity," or capacity to replace hydiogen atoms in 
compounds. In the same year Odling intioduced the 
system of placing dashes as an index to the atom symbol to 
convey the numeiical value of the element in replacing 
hydiogen atoms, and indicated his agreement with Fianlc- 
land as to the definite but, within limits, variable combining 
capacity of some elements 3 

In 1857, Keloile pioposed the addition of a new type, 
methane, CH 4 , to Geihaiclt's types, and suggested that the 
HC1 type was ledundant, being a variant of the H 2 type 4 
Kelcule clearly recognised that the caibon atom was equi- 
valent to four hydiogen atoms, thus claiifying Kolbe's 
suggestion that all oiganic compounds could be legarded 
as denved from caibonic acid, C0 2 . He furthei showed 
that Geihaidt's icsidues weie radicals unattaclced in 
chemical leactions, and elaborated Geihaidt's suggestion 
of mixed types, i.e. compounds denved fiom inoie 
than one fundamental type accoiding to the atom selected 
as the type nucleus. 

In 1858, Cannizzaro published his celcbiated Sketch of a 
Comse of Chemical Philosophy? by which he not meiely 

1 Tmit& dtCbimie 0/gamqiie, Pans, 1853, 1, 132, IV, 589 and 600 

- The Atomic Theory, 1855, p 200 
*y Cbur Sue, 1855, 7, i 

* Ann , 1857, 101, 200, 104, 129, and 133 
5 // Nuovo Ciiiieiiio, 1858, 7, 321. 

42 Chemist? y and Atomic Structure 

established Avogadio's hypothesis and the real distinction 
between equivalent, atomic, and molecular weights, but 
demonstrated the existence of " polyatomic " radicals 
among the simple atoms, thus extending Geihardt's con- 
ception of " atomicity " or saturation capacity from 
organic to inoiganic chemistry. He showed that hydrogen, 
potassium, sodium, lithium, silver, cuprous, and mercurous 
atoms were " monoatomic " (univalent) electro-positive 
radicals, that the halogen elements, chlorine, bromine, and 
iodine, were monoatomic electro-negative radicals, that 
zinc, lead, magnesium, calcium, etc., were " diatomic " 
(bivalent) electro-positive radicals, and oxygen, sulphur, 
selenium, and tellurium weie diatomic electro-negative 
radicals, and, furthei, that atoms existed that were 
equivalent to three or moie atoms of hydiogen or 

In the same year, Kekule brought forwaid, probably 
independently, a theory of " polyatomicity," 1 practically 
identical with that of Cannizzaro, though theie is little 
doubt that Cannizzaro's views had been developed and 
imparted to his students many years befoie Kekule's ideas 
had crystallised, as is evident from Kekule's ambiguous 
paper on the methane type of the pievious year. Kekule 
disagreed with Frankland's views as to variable " atomicity," 
and held that the combining capacity of an element was a 
fixed property of the atom, as exemplified by the constant 
" atomicity " of caibon in organic compounds. Both views 
were right, for it is now clearly recognised that some ele- 
ments have fixed and some elements vanable " atomicity " 
or valency. 

Almost simultaneously with the publication of Kekule's 
theoiy of " polyatomicity, 33 Couper z published a paper in 
which he deduced formulae for compounds based on what 
he called the " elective affinities " and the " affinity of 
degree J> of atoms, identical with Cannizzaro's and Kekule's 

1 Ann , 1858, 106, 129 

2 Phil Mag, 1858, [4], 16, 104, and Coinft rend, 1858, 46, 1157 

Valency 43 

"atomicity" (valency). Both Coupei and Kekule lecog- 
nised that in organic compounds the carbon atoms must be 
regarded as combining among themselves so that only part 
of their affinity or combining capacity was available for 
binding other atoms. Couper, to symbolise the combina- 
tion between atoms, introduced the method of joining the 
atom symbols by means of hyphens or dots, and this method 
was amplified by Crum Brown, 1 who le-mtroduced Dalton's 
system of enclosing the symbol in a circle, Couper's hyphens 
becoming radiating lines equal in number to the combining 
capacity. Frankland, in 1866, by omitting Crum Brown's 
circles, steieotyped the method of formulation now in use 2 
and intioduced the term " bond " to signify the unit of 

Gerhardt had laid it down in his Ttaite de Chimie 
Organique, Pans, 1853, IV, 561, that " it is a widely 
ptevalent ei)or to suppose the possibility of representing 
molecular constitution by means of chemical fo?mults, ot 
in oihet words by the actual arrangement of atoms." 

The polyatomicity and elective affinity theories of 
Kekule and Couper had, however, no other aim than the 
determination of atomic arrangements. Defimteness was 
lent to this aim by ButlerofFs pronouncement in 1861 3 
that the chemical nature of a compound was detei mined 
firstly by its qualitative and quantitative composition, and 
secondly by its chemical " stiucture," by which he meant 
the mode of mutual linking of the atoms in the molecule. 
Practically the whole of research in organic chemistiy since 
that date has been devoted to the deteimination of the 
structural 01 constitutional formulae of compounds, by 
means of which we are to-day enabled to elucidate the 
nature of the mechanism of atomic linkages and the super- 
ficial stiucture of atoms. 

In 1863, Erlenmeyer 4 suggested that the term " basicity " 

1 J Cbem Soc , 1865, 18, 230 

2 Lectuie Notes for Chemical Students, 1866 

3 Zai Chan , 1 86 1, 4, 549 

4 Ibid , 1863, 6, 65, 97, and 609 

44 Chemist? y and Atomic Stiuctute 

be reserved for the combining capacity of acids, instead 
for both atoms and acids, and proposed the terms eih 
zwei-, dtei-, and viei -wetting (=woith}, to expiess tl 
numerical value of atomic combining capacity In 186 
Odling suggested the terms monad, dyad, ti lad, teti ad, etc 
and these teims aie still in occasional use. Hofmann 
book, Intiodiiction to Modem Chemistry, published ] 
London in 1865, contiibuted gieatly to the claiification c 
ideas concerning combining capacity, and he proposed fc 
it the teim " quatitivalence" to indicate its dependence o 
equivalent weights and its numeiical value. This ten 
was shortened to tl valency" (or "valence") by Wiche 
haus in i868, 2 and this teim ib now in almost universal us< 
Confusion, however, continues to exist as to the pi cat 
piefixes to be used in connexion with valency, both Giec 
and Latin piefixes being used by difkient wnteis. A- 
however, valency is definitely of Latin origin (valcns - 
stjong), it is moie consistent nomenclature to use the Lati 
piefixes, and, historically, moie accurate, foi Hofmann use< 
" quantivalency " not " polyvalency" when the tein 
originated. The teiras uni-, bi-, quadn-, quinque- 
sexa-, and septavalent aie, therefoie, to be pi ef cued t< 
mono-, di-, tetui-, penta.-, bexa-, and heptavalent, th< 
prefixes tri- and octa- being the same in both Latin anc 

In modem teims Valency is defined as the definite 
limited capacity possessed by an atom for combining wili 
other atoms measured numerically in hydrogen atoms 01 
equivalent atoms or radicals. It is usual to icseive the 
term " valency " to designate the whole combining capacit) 
of an atom, and to use the term " valence " to specif} 
paiticuki paits of valency, it being thus synonymous with 
the chemist's use of the term " bond " or " linkage." The 
" valency " of caibon, foi example, is foiu, but two only oi 
the four "vale?ices" are appaient in carbon monoxide. 

J. JL 

1 Watt's Dictionary^ London, 1864 
a Ann Snpp , 1868, 6, 257 

Valency 45 

Precision has been lent to this distinction between valency 
and valences by the modern numerical identification of 
valences with elections^ valency thus being the number of 
electrons available or lequired for combination prior to 
the fact of combination, and the valences being the indi- 
vidual elections that actually take part in the com- 
bination For example, hydrogen is univalent, having 
only one or requiring another electron, but in the molecule 
of hydrogen two valences exist between the two atoms, 
owing to the utilisation of both electrons in the com- 

A usual definition of valency is that it is the ratio between 
the atomic and equivalent weights of an atom It has been 
indicated that in complex hydrocarbons, this ratio has not 
an integral value, the equivalent weight of caibon, for 
example, in hexane, C G H 14 , being 5^-, thus involving a 
valency of 12 -V- = 2! atoms of hydrogen. It is obvious, 
in fact, that equivalent weights must vaiy continuously 
from compound to compound in cases where several similai 
atoms ai e combined with each other , their valency, conse- 
quently, is not deteimmable by atoms of other elements. 
The extent to which the valency of an element is to be 
icgaided as due to combination with other atoms of the 
same element, may usually be deduced from synthetic or 
analytical reactions Ethane, for example, can be synthe- 
siscd fiom two molecules of methyl chloride and two 
atoms of sodium, two molecules ot salt being the bye- 

CH 3 C1 + 2Na -j- CH 3 C1 - C 2 H 6 + zNaCl 
If the valency of carbon in methyl cliloiide is four, due to 
combination with three atoms of hydrogen and one of 
chlorine, the valency of carbon in ethane may still be 
regaided as foui by the replacement of a chlorine valence 
by one valence of another carbon atom, the structure of 
ethane thus being H 3 C~ CH 3 . Similar considerations lead 
to the formula H 2 C = CH 2 for ethylene and HC = CH 
for acetylene. This conception of single, double, and tieble 

46 Chemistry and Atomic Structure 

bonds, and the constant quadtivalency of carbon, led immed 
ately after 1858 to a remarkably rapid development , 
the theory of the spatial stmctme of organic molecule 
and later to the demonstration of the tetrahedial disti 
bution in space of the valences of the atoms of carboi 
nitrogen 1 , sulphui 2 , selenium 3 , tin 4 , silicon 5 , phosphoius 
and arsenic 7 . 

1 Le Bel, Coinpt. rend., 1891, 112,724, Pope and Peachey, J Chtm.Soc 
1899, 75, 1127 

2 Pope and Peachey, il-id , 1900, 77, 1072 , Smiles, ibid , 1900, 77, 1174 

3 Pope and Neville, tbtd. } 1902, 81, 1552. 

* Pope and Peachey, Proc, Chem. Soc , 1900, 16, 42 and 116 
o Kappmg, J, Cbem. Soc., 1907, 91, 209 and 717, , 1908, 93, 457 
Meisenheimer and Lichtenstadt, Eei , 1911, 44, 356, Kipping and 
Challenger, J, Chem. Soc , 1911, 99, 626. 
7 Burrows and Turner, ifad : 1921, 119, 426. 



IN 1789, Tioostwyk, Deiman, and Cuthbertson, on passing 
powerful electrostatic charges of electricity through water, 
discovered that gas was produced which re-formed water 
on being sparked. 1 Pearson and Cuthbertson, who repeated 
the experiment in 1797, showed that the gas produced was 
a mixture of hydrogen and oxygen in the proportions in 
which the gases were assumed to combine to form water. 2 
Nicholson (W.) and Carlisle, by collecting the gas formed 
at each electrode, showed that hydrogen was evolved at one 
electrode and oxygen at the other. 3 Ritter, who had for 
some years been working on the chemical decompositions 
produced by electricity, in the same year, independently 
obtained the same results as Nicholson and Carlisle, and 
came to the conclusion that water united with negative 
electricity to foim oxygen and with positive electricity to 
form hydrogen, and that electricity took part in all chemical 
combinations and decompositions. 4 

It had been a common observation that both acids and 
alkalis as, well as oxygen and hydrogen are produced in the 
electrolysis of water, and in 1806 Davy proved that the 
acids and alkalis were deiived from impurities in the water. 5 
In 1803, Berzelius and Hisinger showed that neutral salts 
in solution in water could be decomposed into acids and 
alkalis by electrolysis, and in 1807 Davy effected his memor- 
able electrolytic isolation of the alkali metals, potassium and 
sodium, from fused caustic potash and caustic soda. On 
this foundation Davy erected his electrical theory of 
chemical affinity, in which he assumed that bases (metals) 
are attracted by negative and oxygen by positive electricity. 
Davy so far developed his electrochemical theory as to 

1 Journ Pby* , Nov , 1879 

2 Phil Trans Roy Soc , 1797, 142. 

3 Nicholson's Jaunt., 1800, 4, 179 

* Voigfs Mag , 1800, 2, 356, and Gilbert's Ann , 1801, 9, 284, and 1802, 10, 282 
5 Phil Trans. Roy. Soc , 1807, I 
s Ibid., 1808, i, and 333. 

48 Chemistry and Atomic Structure 

regard all chemical combinations as due to manifestati< 
of electncity between particles, ordinary electrical effe 
being icgaided as the sum of the " paiticular " cffe 
spiead over laige masses. 

Between 1809 and 1811, Avogadio advanced an elect 
chemical theory similar to Davy's, and arranged substan^ 
in a senes in the oidei in which they play the pait of a< 
or alkali towards one another, and showed that this ser 
is the same as that m which they are arranged according 
their development of positive or negative electncity 
mutual contact. He regarded oxygen as being the m< 
electro-negative substance, and ranked other substam 
according to their " oxygemcity " or tendency to play t 
pait of an acid, substances thus ranging fiom stion^ 
acidic thiough neutial to stiongly alkaline. 1 

In 1812, Berzelius put forward an clectiical theory 
chemical combination, known as the " Dualistic Theory 
which dominated the field of chermstiy for ovei quartei 
a centuiy. 2 Like Davy and Avogadio, Berzelius icgaid 
oxygen as the most electronegative element, but he fuith 
assumed that cveiy atom had two electncal poles, o 
positive and one negative, usually of unequal strength, t 
atom thus having a predominating polarity. Chemic 
combination of the fiist older was that between elementa 
atoms by a complete or partial neutralisation of opposi 
electricities, whereas combination of the second order w 
that between two pai tides of the first ordei having u 
neutralised residues of opposite electricities. Baryta, Ba( 
for example, was regarded as possessing a residual positi 
electrification, and sulphuric acid (anhydride), SO 3 , 
residual negative electrification, combination between the 
resulting in completely neutral " heavy spai " (barm 
sulphate), BaSO 4 or BaO.SO 3 . 

Faraday's work on the laws of electrolysis in 18- 

1 Journ Pbys , 1809, 68, 142 , 1810, 69, i , and 1811, 73, 58 

2 Schtoeigger's JoujJi., 1812, 6, 119 , Essay upon the Theory of Chemical Pi 
portions and the Chemical Action of Electncity, Stockholm, 1814, Treatise up 
Chemistry, 1817 

Electto-Ghemistty 49 

disclosed that the chemical equivalents of elements aie 
associated with equal quantities of electiicity, thus definitely 
controveiting Beizelms' assumption of unequal charges. 
Dumas' proof in 1834 tnat electronegative chloiine could 
replace electiopositive hydiogen, equivalent foi equivalent, 
without material change in chemical properties, extended 
the inapplicability of the duahstic theory fiom inorganic to 
organic chemistry, and within a few years the dualistic 
theory was abandoned by all but its author, to be revived 
neaily half a century latei, in a special foim in application 
to ionic dissociation of salts in solutions, by another Swede, 

To explain the formation of hydrogen or alkali or 
deposition of metals at one electiode and of acid or oxy- 
genous substances at the other electrode in electiolysis, 
Grotthus in 1806 suggested that the paiticles in solution 
acted as small dipolar magnets and arranged themselves 
into chains extending from electiode to electiode, the 
teimmal polar parts of opposite sign being split off under 
the attraction of the oppositely chaiged electrodes. 1 
Subsequent woik has proved that infinitesimal amounts 
of electiicity suffice to make evident the decompositions 
associated with electrolysis, and modern theories involve 
that the electiolyte is dissociated into polai paits in the act 
of solution and in the absence of an electric cunent, and 
Grotthus' Hypothesis, that the cm rent causes the dissoci- 
ation, thus appeais to have no foundation. The modern 
theones are essentially a return to Davy's explanation of 
1806 that chemical attraction between the particles of a 
substance is destioyed by imparting electnc charges, and 
conveisely that the paiticles are chaiged before decomposi- 
tion by electiolysis. 

In 1833 and 1834, Faraday investigated the quantities ot 
the products of electiolytic decomposition pioduced by 
equal quantities of electiicity, and discoveied the two Laws 
of Electrolysis, that the amount of decomposition of a 

1 Ann Cbiin. Pbys , 1806, [i], 58, 64., and 1807, [i], 63, 20. 

50 Chemistry and Atomic Stnutiue 

given electrolyte is proportional to the quantity of elec- 
tricity which flows through it, and that the quantities of 
different substances liberated by the same quantity of 
electricity are in the ratio of their chemical equivalents 1 

Faraday assumed that the products of electiolysis could 

not appear at the electrodes without motion of parts of 

trie electrolyte molecules fiom end to end of the liquid, 

and proposed the term "ions" (Greek =" travellers "), 

" amons " (Greek = " upgoeis "), and " cations " (Greek 

= " doiungoers "), the anions being those atoms 01 gioups 

of atoms transported to the "anode" (Greek = " up- 

path ") or positive electrode, and the cations those tians- 

ported to the "cathode" (Greek = " dozunpatb ") 01 

negative electrode, amons thus canying negative and cations 

positive charges. He fmther suggested that the equivalent 

weights of substances were simply those quantities that 

contained equal quantities of electricity, and that the 

quantities of electricity canied determined the equivalent 

number and the chemical combining force. Though 

Faraday failed to discern the different valencies of atoms 

and the " atomic " nature of electricity, his theoiy of 

chemical combination, as due quantitatively to electiical 

forces, is the basis of all modern theones of chemical 

combination and atomic stiuctuie. 

In 1849, Weber put forward a theory of electiicity 
anticipating, by over half a century, a great pait of lecent 
theories of the structure of matter and the natuie of 
electricity. 2 Weber regarded an electric cunent, not as a 
continuous movement of electricity, but as a movement of 
electrostatic chaiges, and assumed the existence of positive 
and negative " atoms of electiicity" and suggested that they 
must possess electromagnetic mass in virtue of their velocity. 
He further assumed that only the negative atoms of elec- 
tricity were associated with matenal atoms, the positive 

1 Phi Trans Roy Soc , 1833, 123, 23, 1834, 124, 77, and Expenmental 
Reseaiches in Electncity> London, 1839, lj ro 7> r 95^ 2I 5 3 230, and 821 

-Leipzig Trans, 1849, *> and l8 57j 5) 2( 5o , Pogg Ann , 1856, 99, 10 , and 
Jrcikc, 4, p 278 

Elects o-Chemistry 51 

atoms of electiicity being devoid of weight and i evolving 
round the massive negative atom. Interchanging the 
electrical signs, Webei's theory of the atom is practically 
identical with the Thomson-Rutherford-Bohr theory of the 
present day. 

In 1850, Williamson 1 advanced a theory of the constitu- 
tion of molecules and the mechanism of " double-decom- 
positions " practically identical with modern views as to 
mass action and ionic dissociation in solutions. He 
icgarded hydiochloric acid, for example, not as immutable 
molecules of HC1 but as undeigoing a continuous process 
of decomposition and lecombination, " each atom of 
hydtogen constantly changing -places with other atoms of 
hyd)ogen, or, what is the same thing, changing chlorine." He 
fuither regarded the theory of static atoms as unsafe and 
unjustifiable, and stated that " the dynamics of chemistry 
will commence by the t ejection of this supposition, and will 
study the degtee and kind of motion which atoms possess, and 
1 educe to this one fact the vanotis phenomena of change"* 
His statement " that metallic zinc is contained in its sulphate 
is only stnctly tnie by absh action from most of the pi ope t ties 
of the metal. The material atom, which undet cettain circum- 
stances -possesses the properties which we describe by the wotd 
' zinc, 7 is no doubt contained in the sulphate, but with dijfeient 
piopetties" is prophetic in view of the modern theory that 
the sulphate group contains two electrons abstracted from 
the atom of metallic zinc, thus converting the latter into a 
doubly-chaiged atomic ion having few of the propeities of 
the oiigmal atom except weight. 

In oider to explain the fact of the small electric forces 
necessary for electiolytic decompositions, Clausius, in 1857, 
suggested that some of the molecules of the electiolyte aie 
split up into ions in the act of dissolving, the phenomena 
of iomsation thus being stnctly independent of electrolysis. 
He was unable to assign any definite value to the extent of 

1 See Chap III, p 39 

2 Alembic Club Reprints, No 16, pp 15, 17, 19, 23, and 48. 

52 Chemistry a?id Atomic St-tuctiire 

ionisation, but assumed that it was infinitesimal, and that 
the ions, lemoved by discharge at the electiodes during 
clectiolysis, were progiessively replaced by further ionisa- 
tion of the electrolyte * 

Clerk Maxwell, in explaining the process of electrolysis, 
suggested in 1873 that "each molecule., theiefoie^ on being 
hbetated fiom the state of combination, parts with a chatge 

whose magnitude is " (N = numbei of molecules in the 

clectiochemical equivalent), " and is positive for the cation 
and negative fot the anwn. This definite quantity of elec- 
tncity we shall call the molecular chatge. If it weie known 
it would be the most natuial unit of elect) uity " z He 
regarded the constant molecular chaige or " molecule of 
electricity " merely as a convenient hypothesis foi the 
explanation of electrolysis, and rejected it entirely in 
developing his " Electromagnetic Theory of Light." 

In 1874, Johnstone Stoney pioposed a system of funda- 
mental units of nature, the natural unit of quantity of 
electricity being that associated with the electrochemical 
equivalent in electrolysis. He stated, " Fot each chemical 
bond niptwed in electiolysis a ceitain quantity of electucity 
tiaveises the electrolyte., which is the same in all cases" Foi 
this constant quantity of electricity he pioposed the name 
" elect] we" B or electromagnetic unit of quantity. This 
name, in 1891, he altered to "electron" 4 in a paper 
suggesting the physical chaiactenstics of the orbits of the 
electiical particles giving use to the light radiation causing 
the shaip lines in optical spectia. 

Apparently unaware of the suggestions of Cleric Maxwell 
and Johnstone Stoney, Helmholtz, in his Faiaday Lecture 
to the Chemical Society in 1881, put forwaid a similar 
explanation of Faraday's laws of electrolysis. He regarded 

1 Clausms, Pogg Ann , 1857, 101, 388, and 347. 

* Electricity and Magnetism, ist Ed 1873 , 3rd Ed , 1892, p 379 

3 Brit Assoc. Rep , 1874; Sect. A, 22 , reprinted, Phil Mag , i8Si, [5], 11, 


4 Scient Proc. Roy Dublin Soc , 1891, 583. 

Electro-Chemistry 53 

the chemical equivalent of each ion in an electrolyte as 
being united to an electric equivalent or atom of electiicity, 
and suggested " that each unit of affinity that an atom can be 
said to possess represents a chatge that may be tetmed one 
atom of electncity" thus identifying valency with units oi 
electncal charge J 

The answer to Clausius' pioblem, as to the extent of 
ionisation of salts in solution, was suggested in 1884 by 
Anhemus, 2 who regarded electrolytes as being laigely 
ionised in the act of dissolving, and as becoming more and 
more completely ionised with increasing dilution of the 

In 1885, van't Hoff put forward a kinetic theory of 
osmotic pressuie based on the analogy between the process 
of vaporisation and the process of solution. 3 He showed 
that the analogy between sblutions and gases is complete 
on the assumption that the molecules of the solute exert 
the same (osmotic) pressuie as they would as a gas occupying 
the same volume as the solvent. This kinetic theory of 
osmotic piessure has enabled determinations of molecular 
weights to be made in the case of non-vaporisable sub- 
stances, by reference to deviations (fiom the normal con- 
stants of pure solvents) in vapour pressure, boiling and 
freezing points of solutions, and solubility of solutions for 
other substances, these deviations being simple functions of 
the osmotic pressure. 

In the case of electiolytes, the osmotic pressures are 
found uniformly to be abnormally high, and Airhenius, by 
the application of his theory of ionisation, was able to show 
that the abnormality was due to the splitting of the electio- 
lyte into ions, each ion contributing to the osmotic pressuie 
an amount equal to that due to a non-ionised molecule. 4 

ij Cbem Soc , 1881, 39, 302 

2 Researches on the Conductivity of Electrolytes, Stockholm, 1884, Bnt Assoc 
Rep., 1886, 357 

S K Svenska Vet-Alad Hand!, 1885,21,38, Atch Neol , 1886,20,239, 
tnd Zeit pbys Cbem , 1887, 1, 481 

1 Zen phys. Cbem , 1887, 1, 631. 

54 Cbewistiy and Atomic Structure 

As determinations of the degree of ionisation can be made 
on the assumption that the number of ions is piopoitional 
to the electrical conductivity of the solution, the inoleculai 
weights of ionisable substances can be determined by means 
of the conjoint hypotheses of van't HofF and Aiihenius 
In the hands of Werner and otheis, this method has given 
an insight into the constitution of complex compounds 
otheiwise experimentally unobtainable. 

In 1890, Ciamician suggested that ionisation of electro- 
lytes (salts, acids, and bases) in solution is due to attraction 
between the molecules of the solvent and the positive and 
negative parts of the molecules of the electrolyte, thus 
bringing about ruptuie of the chemical molecules into 
sepaiate charged paits or ions, each entirely surrounded by 
molecules of the solvent. 1 This suggestion is supported to 
some extent by the fact that those solvents with great 
ionising capacity usually unite with anhydrous salts to form 
crystalline complex molecules, e g. Na 2 C0 3 .ioH 2 O, washing 
soda with zuatei of crystallisation, CaCl 2 6C 2 H 5 OH, calcium 
chloride with alcohol of crystallisation, MgI 2 .6CH 3 COCH 3 , 
magnesium iodide with acetone of crystallisation, 
AlBig.CgHsOCaHg, aluminium bromide with ethei of 
crystallisation, NiSO^CaELjOg, nickel sulphate with gly- 
cenne of crystallisation, CuSO 4 .H 2 0.4NHg, copper sulphate 
with both water and ammonia of ciystalhsation, and 
CrCl 3 .3C 5 H 5 N, chromic chloride with pyiidme of crystal- 
lisation. Van der Waals, m 1891, put forward a similar 
theory of ionisation, and icgarded the heat of hydration of 
salts as the source of the energy necessary to lupturc the 
chemical molecule into oppositely chaiged ions. 2 

In 1891, Werner put forwaid a new theoiy of chemical 
affinity and valency, 3 in which he regarded the foice of 
valency as being distributed over the suiface of the atom 
and utilised in binding other atoms sufficient in numbei to 

1 Zeit phys. Cbem , 1890, 6, 408. 

z lbid, 1891,8, 215 

3 See his Co-ordtttatton Theory, Ch. VII, page 81 

Electro-Chemist? y 55 

form an enclosing shell to the cential attracting or co- 
ordinating atom. Anhydrous salts on solution were thus 
icgarded as combining with the solvent molecules by means 
of some of their atoms, so that each cential metallic atom 
completed its " co-oidination number " of atoms in the shell, 
the shell being formed partly or wholly of atoms deiived 
from the solvent. If all the acidic atoms (or groups) of 
the original salt remain attached in the co-oidination shell 
no iomsation occurs, but, if one or more of the acidic atoms 
(or groups) be not bound in the shell, ionisation occurs, the 
central atom with its co-ordinated shell of atoms forming 
a complex positive ion, the acidic atoms (or groups) foiming 
negative ions. This co-ordination theoiy was applied not 
merely to the phenomena of ionisation and electrolysis but 
to the constitution and reactivities of complex salts in 
general, and constitutes probably the most outstanding and 
fruitful contribution to the general theory of chemistry in 
the histoiy of the science, and an interpretation of the 
theory in terms of the electionic nature of valency enables 
far-ieaching deductions to be made as to the external and 
internal structure of the atoms of practically the whole ot 
the known elements. 



THOUGH Plato, about 400 B.C., had assigned to the pumoi 
dial atoms the three-dimensional sti natures of the five 
regular solids, the tetiahedton, octahedron^ cube, icosahedton 
and dodecahedron, having four, six, eight, twelve , and twenty 
point configurations icspectively, these shapes were noi 
assigned to the atoms on the grounds of any physica 
necessity occasioned by any known pioperty of matter 
Wollaston's recognition of the simple multiple combining 
capacities of caibonic, oxalic, and sulphunc acids, 1 and his 
suggestion as to the three-dimensional geometry of the 
pumaiy particles of matter, piobably constitute the eailicst 
use of steieo-conceptions in the explanation of expenmental 
facts. Wollaston said, " when our views are sufficiently 
extended, to enable us to teason with precision concetmng the 
ptopottwns of elementary atoms, we shall find the aiithmetical 
relation alone will not be sufficient to explain then mutual 
action, and that we shall be obliged to acquit e a geometrical 
conception of then relative arrangement in all the three 
dimensions of solid extension . . a stable arrangement may 
again take place, if the four particles ate situated at the angles 
of the four equilateral triangles composing a regular tetra- 
hedron . . it is pet haps too much to hope, that the geo- 
metrical arrangement of primary pat tides will ever le 
perfectly known" However, foity-six years aftei Wollas- 
ton's death in 1828, van't Hoft and Le Bel simultaneously 
put forward complete explanations of the pioperties of 
organic substances based on the tetrahedral natuie of the 
carbon atom, and, seventy-two years after the same date, 
Pope and Smiles (see page 46) simultaneously demon- 
strated the tetrahedral nature of the sulphur atom, and the 
tetrahedral structure of the caibon and sulphui atoms in 
Wollaston's polybasic caibonic, oxalic, and sulphui ic acids 
is to-day no less fiimly established than the leahty of atomss 
The giowth of the idea that the pioperties of compound. 

3 Phil Trans Roy Soc , 1808 98, 96 

Falencv and Steieo-Chemishy 57 

are largely dependent on the arrangements of atoms is 
coincident with the giowth of the conception of valency. 
Gay-Lussac's suggestion of 1824, as to the different arrange- 
ments of the atoms in silver fulminate and silver cyanate, 
was amplified by Berzelius in 1832 into a theory of iso- 
mensm. Lament's "nucleus theory" of 1835 fore- 
shadowed the carbon atom with a definite shape and 
capacity for spatial combination. Dumas' theory of 
" mechanical types " and " substitutions " and Laurent's 
" nucleus " theory emphasised the persistence and stability 
of the combinations of caibon atoms with one another. 
Pasteur's discovery in 1848 of the hemihedrism of optically 
active crystals and his researches in subsequent years led to 
the theory of molecular asymmetry and the tetrahedial 
stiucture of a large number of atoms. Liebig's theory of 
polybasic acids in 1838, Williamson's "water type" of 
1850, Franldand's theory in 1852 of the definite combining 
capacity of atoms, and Gerhaidt's theory of simple types 
of 1853, paved the way for Canmzzaro's, Kekule's, and 
Couper's theory of "atomicity" or valency, in 1858, and 
within a few yeais definite valencies had been assigned to 
all the known elements. 

In 1852, Frankland (E ) suggested that " the formation of 
a jive- atom gtoup, from one containing three atoms , can be 
effected by the assimilation of two atoms (or two semi-mole cities} 
either of the same 01 of opposite elect-) o-chemical chaiactet" * 
and in 1866 he amplified this suggestion in the statement : 
" This variation of atomicity always takes place by the dis- 
appearance of an even numbet of bonds . . . one 01 mote pans 
of bonds belonging to an atom of the same element can unite, 
and., having satin ated each othei, become, as it were, latent." z 
This suggestion was revived in 1902 by Spiegel, 3 by 
Ai rheums m 1906,* and again by Fiiend in igoS, 5 as a 

1 Phil Trans Roy Soc , 1852, 142, 417 
- Lecture Notes for Chemical Students, 1866 

3 Zeit anorg Chem , 1902, 29, 365 

4 Theonen der Cbenne, Leipzig, 1906 

5 J Clem Soc, 1908, 93, 260 

58 Chemist) y and Atomic Structure 

theoiy of " neutral affinities," " electrical double valen- 
cies," and " latent valency," and involved additional 
ordinary valences capable of being utilised in pairs of equal 
and opposite sign. In this foim, however, the theory has 
proved sterile, and has led to much fruitless discussion, 
owing to the fact that many elements aie known which 
have several valencies cliff et ing not by two units but by one, 
for example, the twenty-four elements of the three transi- 
tion series of the periodic classification. Werner, in his 
" co-oidination theory " adopted Franldand's suggestion 
and clearly elucidated its meaning. He showed that lower 
valency is converted into valency one or two units higher 
only when the element exhibits electropositive valency 
(metal-lite), and that the inciease in apparent valency by 
two units of equal and opposite sign is found only when the 
element exhibits electronegative valency (as a non-metal). 
He regaided the two apparent valences as dissimilai in 
function, one of the atoms added being held by "residual 
affinity " of the cential electronegative atom and the othci 
appealing as an electronegative ion. On this basis he 
elaborated a theory of the constitution of metal compounds 
with ammonia and water, deiiving them from the simple 
ammonium type, the apparent additional valences of the 
combined ammonia groups being utilised not necessanly m 
pairs but even in parts of a single valence. Positively 
trivalent nitrogen in nitrous anhydride, N 2 3 , by the 
addition of two equivalents of oxygen per atom of nitrogen, 
yields nitric anhydride, N 2 O 5 , containing positively quin- 
quevalent nitrogen. Negatively trivalent nitrogen in 
ammonia, NH 3 , on the other hand, cannot be induced to 
yield a negatively quinquevalent derivative. It combines 
with hydrochloric acid, HC1, to yield ammonium chloride, 
NH 4 C1, in which the negative chloride atom ionises fiom 
the positive ammonium group, NH 4 , on solution in watei. 
Werner postulated that the hydrogen atom from HC1 is 
added to NH 3 without alteration in mttogen valency, merely 
by the general attraction between the positive hydrogen 

Valency and Stereo-Chemistry 59 

atom and the negative nitiogen atom, but that, once com- 
bination is effected, all four hydiogen atoms in the NH 4 
gioup are identically bound to the nitrogen atom. This 
conception, difficult as it then seemed, has in recent years, 
by chemical and physical methods, been fully confirmed, 
and, in the electronic interpretation of valency, is seen to 
be lemarkably simple. Wemei fuithei postulated that the 
metal ammines, foi example, hexammmocobaltic chloride, 
CoCl 3 6NH 3 , are also ammonium compounds, one-sixth of 
a cobalt atom taking the place of a hydrogen atom in 
ammonium, NH 4 , so that the complex compound could 
be regarded as 6(NH 3 Coi).Cl 3 , or [Co(NH 3 ) 6 ]Cl 3 In this 
case six ammonia molecules add three equivalents of cobalt 
and three chlounc atoms, or each ammonia molecule adds 
only one-half of a positive valency plus one-half of a negative 
valency. This interpretation rendered impossible the 
existence of latent valencies in pans of equal and opposite 
sign, and confiimed Weinei's view that valency, once called 
into operation, is a foice dibtiibuted ovei the surface of an 
atom and is independent of the atoms to which it owes its 
oiigin This theory necessitated the distnbution of the 
valency forces in space, and led to the octahedral stiucture 
of many atoms, a structure which has been demonstrated 
beyond any possibility of doubt by Werner and others 
from considerations based on geometiical and optical 

In 1854, Hofmann suggested that the ammonia com- 
pounds with metallic salts were derived from the pentad 
nitiogen type, and luteo or hexammonio-cobalt chlonde 
having the following foimula, Co(NHoCl.NH 4 ). J or 


H 4 N-NH 2 -Co-NH a -NH 1 

Cl NH 2 -NH 4 


60 Chemistry and Atomic Structure 

This suggestion was elaborated into a complete scheme c 
chain-formulae by Blomstrand in I869, 1 and was adopte 
geneially, until Werner's demonstrations of the spati. 
structure of these complexes bi ought about the downfa 
of Blomstiand's theory. The authority of Blomstrand an 
Joigensen, to the latter of whom, chemistiy eaily owes th 
preparation of the bulk of the metalammmes, was sufficien 
to maintain the plane chain-foimulse for ovei forty yeai 
and retarded the application of spatial ideas of molecula 
stiucture in inorganic chemistry for at least a generatior 
Even as late as 1910 authoritative adheients were to b 
found in this country for Blomstrand's foimulations, wit 
the result that Werner's theory of molecular stiuctuie wr 
held in suspicion, and its application to atomic structur 
delayed till after Werner's death in 1919. 

In 1857, Gibbs and Genth, in their epoch-makin 
researches on the ammonia-cobalt bases, 2 refened to th 
cobalt atom as " hexatomic," despite their lecognition tha 
the combining capacity of cobalt was tlnee. They appai 
ently distinguished the tnvalency of cobalt in its combma 
tion with oxygen 01 acid ladicals from the six group 
forming the " conjunct " with the cobalt atom, an< 
regarded the molecules as containing only one atom c 
cobalt. Their views were a remarkable foreshadowing c 
Werner's co-oidination theory of thirty-four yeais latei 
and had Gibbs not subsequently embraced Blomstrand' 
theoiy of cham-formulse and two cobalt atoms pci mole 
cule, it is probable that a co-ordination theoiy would hav 
been evolved long before 1891 

The development of ideas as to the spatial anangemen 
of carbon atoms in organic compounds was very rapid afte 
Kekule's exposition of his theory of " atomicity " (valency 
in 1858. Butleroff, m 1860, viewed the "structure" o 
mode of linking of carbon atoms as of prime importance 
and in 1863, Wislicenus, referring to the existence of iso 

1 Cbemie der Jetztzett, Heidelberg, 1869 

8 Smithsonian Contribution! to Knowledge, Washington, 1857, IX - 

Valency and, Stereo-Chemistry 61 

meric foims of lactic acid, 1 stated that the existence of 
isomers was due to change m the order of distribution of 
atoms m space, whereas foimulce lepresent only a picture 
of a molecule in one plane. In 1866, Kekule suggested the 
cyclic foimula for benzene, which, with very slight modifica- 
tion, represents accurately the views of chemists at the 
piesent day. 8 In the following year Kekule suggested that 
the four units of affinity of the carbon atom pi ejected from 
the surface in the diiection of the four faces of a tetra- 
hedron. 3 In 1869, Paterno suggested that, if the foui 
valences of carbon weie directed towards the coiners of a 
tetiahedion (Diagram I, p. 64), the foimula C 2 H 4 C1 2 should 
lepiesent thiee isomenc substances. In 1873, Wislicenus, 
to whom chemistry is largely indebted for the experimental 
elucidation of the aiiangement of carbon atoms in many 
01 game compounds, suggested the term "geometrical 
isomerism " to covei the cases of compounds identical in 
composition and drfteiing only in the aiiangement of the 
atoms m space 4 This term was altered in 1888 by Victor 
Meyer to ' ' stereo-isomerism ' ' to covei all cases of 
spatial isomers, "geometrical isomerism" being le- 
tained foi the special cases associated with double bonds 
between atoms, and "optical isomerism" for those 
cases in which the only diffeience in the properties 
of the isomeis is their opposite rotatory effect on 
polarised light. 

Within a few months of one anothei in 1874, Van1t Hofl 
in Holland and Le Bel in France published almost identical 
theories of the tetiahedial structure of carbon atoms. 
Van't Hofr appai entry regarded the carbon atom as being 
a tetiahedral shaped piece of mattei, 5 wheieas Le Bel 
icgarded the carbon atom as a centie of atti action lound 
which foui other atoms could be ai ranged with tetiahedral 

1 Ann, 1863,128, i. 

- Ibid , 1866, 137, 1 60, and 229 

J Zeit Chem , 1867, new series 3, 217 

* Ann, 1873, 167, 345 

Foorstel tot nitbreiding dei stiuctuiujoi mules in de mimte, Sept , 1874 

62 Chemistiy and Atomic SttuLtwe 

symmetiy. 1 Both theories gave a complete explanation o 
isomensm in oiganic compounds, van't HofPs theory bein, 
more particulaily dnected to the explanation of geometnca 
isomerism and Le Bel's to optical isomerism, the forme 
being influenced largely by the woik of Kekule anc 
Wislicenus and the latter by that of Pasteur. 

In 1869, Blomstrand z pointed out that the most strongl 
electropositive and electionegative elements possessed th 
smallest satuiation capacity (valency), the highly positiv 
alkali metals, for example, combining with great energ 
with the highly negative halogen elements to form com 
pounds containing one atom of each per molecule. N 
explanation of this circumstance appears to have yet bee 
given, and it constituted the chief reason for the leluctanc 
of chemists to accept the electrical theories of Helmholt 
and others that chemical forces are entirely electrical i 
ongm, for the variability in intensity of chemical force 
with valency appeared incompatible with invariable elec 
tucal charge per valence. The variations in intensity c 
chemical forces aie now known to be closely connected wit 
the capacity of elements to form iomsable compounds, o 
in teims of Weinei's co-ordination theory, with th 
capacity of atoms foi stable co-oidmation with other atom 
thus disclosing the features of their superficial structuic 
The alkali metals and the halogens almost alone of th 
elements give rise to atomic ions in solutions, foim easil 
hydrolisable complex compounds, and give no indication 1 
chemically, of possessing any stiuctuie other than that of 
umfoim spherical surface. There is, consequently, pract] 
cally no stereo-chemistry of the alkali metals 01 the negativ 
ions of the halogens. 

In 1 88 1, van't Hoff suggested that valency is a functio 
of the shape of an atom, and that vanable valency 
associated with variation in the atomic shape. 3 He state 

1 Le Bel, Bull. Soc Chem , 1874 (Nov ), 22, 337 

z Loc cit , p. 60 

3 Anstcbtcn Tiber die orgamsche Cbemie, 1881, XI, 3 

Valency and Stet go-Chemistry 63 

that " an atom of such a shape (tiiangular) would therefore 
cumpott itself genet ally as tnvalent and occasionally as 
sexavalent, and, it should be noted in this connexion, that, just 
a* the number and kind of the valences may be deduced f torn 
the shape of the atom, so conversely the shape of the atom might 
be deduced ft om an accutate knowledge of its valency" Ten 
years latei Le Bel 1 showed that methyl-ethyl-piopyl- 
zjobutyl-ammonium chloiide could be obtained in optically 
active forms, thus demonstiating that the central nitrogen 
atom can give rise to optical isomers. Wernei later 
showed that the dissociation of this and similai ammonium 
compounds into a complex ion and a halogen ion necessitated 
that the mtiogen atom, like carbon, possesses tettahedtal 
symmetry, and this conclusion has since been amply con- 

If the foiu valences of the carbon atom aie directed fiom 
the centre to the foui coineis of a icgular tetrahedron (see 
Diagiam I, p. 64), it can be pioved geometiically that the 
angle between any two of the valences must be 109 28'. 
If, fuither, two carbon atoms in combining by means of a 
single bond have the dnection of the two separate valences 
united into a stiaight line between the centres of the atoms 
it is impossible for a chain of carbon atoms to complete a 
ring, for the angles of an equilatcial triangle, squaie, 
pentagon, and hexagon are icspectively 60, 90, 108, and 
120 Compounds are well known, however, in chemistiy 
containing closed ungs of thiee, four, five, and six carbon 
atoms. The smallest deviation from the tetrahedial angle 
of 109 28 is consequently that of the five-membered ung 
of 108. In 1885, Baeyer put foiward a theoiy, known as 
" Baeyer 's Strain Theory," that the valences of the carbon 
atoms are strained out of the position of tetrahedral equili- 
brium by implication of the atoms in cyclic structuies, the 
stiain being least for the five-membeied cyclo-pentane ting, 
and greatest foi the thtee-membeted cyclo-piopane ting, othei 
rings showing an intei mediate degiee of stiain in propor- 

1 Compt rend., 1891, 112, 725 

Chemist) y and Atomic Stnictiue 


Three Tetrdhedral Ang/es 

Four TetraheJral Ang/ 

five Tetrahedral Ane/cz 

Cyc/oorooane Ring 

5w Jetrahedrjl Ang/es 

Gyclohexane or Benzene &ng 

Valency and Stereo-Chemist? y 65 

tion to the deviation of the polygonal angle from the 
tetrahedral angle. This theory accords fairly well with the 
known ease or difficulty of formation and of disruption of 
cyclic structures. 1 This theory has undeigone considerable 
modification in the last five years as a result of the work of 
Thorpe (J. F ) and Ingold and their collaborators, 2 who 
have proved that the strain on two caibon valences impli- 
cated in an open-chain or a cyclic structure can be leheved 
by the utilisation of the other two valences in binding othei 
atoms (or groups of atoms) of appropriate volume, the angle 
between these two valences being 115 3 for hydrogen 
atoms, IC"9'5 for methyl groups, and still smaller angles for 
larger groups or for a ring of carbon atoms. These investi- 
gations promise to throw considerable light on deviations 
of the structure of the surface of carbon atoms from tetra- 
hedral symmetry, and, inferentially, on the nature and 
stability of the elcctiomc orbits in atoms. 

1 Polyacetylene Compounds, 1885, and Her , 1885, 18, 2277 
z y Cbem Soc , 1919, 115, 320, 1920, 117, 591, etc 



THE earliest attempts to classify elements probably date 
back to the discovery of metals and their recognition as 
single substances. The metals were oiigmally classified as 
noble and base, the foimer including gold and silver, and 
the latter coppei, iron, lead, tin, and mercuiy. Until, 
however, the criteiia foi the distinction of elements fiom 
compounds were established towards the end of the 
eighteenth century, no further advances in the classifica- 
tion of elements weie possible. 

In 1789, Lavoisier 1 ananged the elements into four 
classes : (i) light, heat, and the three gases, oxygen, nitio- 
gen, and hydiogen ; (2) non-metallic solids, such as sulphui, 
phosphorus, and caibon ; (3) metals, such as copper, iion, 
lead, gold, silver, and mercury; and (4) " e tilths," such as 
lime, baryta, magnesia, alumina, and silica. The " earths," 
and light and heat aie, however, no longer regarded as 
elements, and the two last aie not even material substances. 

Berzelms based his " Duahstic Theoiy " z on Lavoisiei's 
classification, with such amendments as the inteivening 
years had icndeied necessary, the elements being divided 
into electropositive and electionegative groups, which 
coincided laigely with the gioups of metallic and non- 
metallic elements. 

In 1815, Prout 3 basing his ideas on Dalton's atomic 
weights as amended by the moie accuiate work of Berzelms, 
suggested that the atomic weights of all the elements were 
exact multiples of the atomic weight o hydrogen. Though 
Prout 's Hypothesis was strongly supported, particularly by 
Thomson (T.), it was regarded by Berzelms in 1827 as 
having no foundation. Attempts were made at various 
times (Dumas and Stas in 1841, and Maiignac in 1842 and 
1860) in the nineteenth centuiy to revive the hypothesis, 

1 Irani clementaire de Cbtwie, Pans, 1789 

2 See Chap IV, p 48 

3 Thomson's Ann, Pbil } 1815, 6, 331, and 1816, 7, HI. 

Classification of the Elements 67 

which was finally abandoned by chemists after Stas' 
extiemely lefined deteimmations of atomic weights in 
1860 had demonstrated that many atomic weights were not 
even approximately integeis. A modification of this 
hypothesis was put forward in 1896 by Lothar Meyer, who 
suggested that all atoms are composed of hydiogen atoms 
combined with varying amounts of the " Iwinniferous etbei" 
which he thought might not be entirely devoid of weight. 
Lothar Meyer's suggestion has not been accepted by 
chemists generally, but recent speculations on the structuie 
of matter and the nature of energy have lent considerable 
support to it, in the sense that all mass may be a manifesta- 
tion of the inertia of specialised portions of the ether, 
decrease in ineitia resulting on veiy close juxtaposition of 
such specialised poitions of the ether. The modem theory 
of isotopes and inferences fiom ladio-activity and the 
artificial disintegration of atoms have, however, brought 
about a moie dnect return to Prout's hypothesis, and non- 
integial atomic weights are now regarded as aveiage atomic 
weights due to mixtures of atoms of integral atomic weight. 
In 1817, Dobereinei obseived that gioups of three closely 
related elements have either nearly the same atomic weight 
01 have atomic weights showing an approximately constant 
difference, for example, iron, cobalt, and nickel, with 
atomic weight about 58, and calcium, strontium, and 
baiium, with atomic weights differing by about 48, chlorine, 
bromine, and iodine with difference about 46, lithium, 
sodium, and potassium with difference about 16, and sul- 
phur, selenium, and tellurium with difference about 48. 1 
Dumas extended Dobereiner's Triads by including fluonne 
with chloime, bromine, and iodine , oxygen with sulphur, 
selenium, and tellurium. ; and nitiogen with phosphoius, 
aisenic, and antimony. 2 

1 Gilben's Ann, 1817, 56, 332, and 57, 436, and Pogg Ann, 1829, 15, 

2 Traits de Cbnme apphqni aux Arts, Pans, 1828, But Assoc Rep , 1851, 
Compt. rend, 1857, 45, 709, 1858, 46, 9151, and 47, 1026, Ann. Cbnn Pbys 
1859, [3], 55, 129 

68 Chemistry and Atomic Structure 

Gmelin, 1 in 1843, further extended Dobereiner's sugges- 
tions and devised a system of classification of the elements 
based on family relationships between elements, similarities 
in chemical properties appearing after ceitain increases in 
atomic weight. In 1845, Faraday showed that elements 
could be classified in two series, (i) paramagnetic elements 
attracted by a magnet, and (2) diamagnetic elements 
repelled by a magnet, and, in 1852, he further showed that 
this classification included in the same series those elements 
exhibiting close chemical similarity. 2 

Pettenkofer re-examined the numerical relationships 
between Dobereiner's tiiads and other chemically related 
elements, and in 1850 suggested that chemically similar 
elements formed series in which their atomic weights could 
be derived as a modified arithmetical progression involving 
the lowest atomic weight and multiples of an integei ; and 
modifications of Pettenkofer's Series were proposed from 
time to time until the general law underlying the classifica- 
tion of the elements was discovered. 3 In 1854, Cooke 4 
showed that Dobereiner's triads were merely parts of seiies, 
and that each seiies followed a simple algebraic law in the 
increase of atomic weight. 

In 1857, Odling arranged the elements into thiiteen 
classes according to similaiities in chemical and physical 
propei ties, each class being ai ranged in the order of atomic 
weights. 5 Though this classification included the various 
groups of elements chemically closely related, it failed to 
ehow any relationship between atomic weights and chemical 
properties. It is, however, practically identical with the 
classification now in use for grouping the elements for 
qualitative analysis, and included silver, lead, and meicuiy, 
which give insoluble chlorides ; arsenic, antimony, and 

1 Handbitcb der Cbemie, Heidelberg, 1843. 

2 A Course of Six Lectures on the Non-metallic Elements, London, 1852 

3 Pettenkofer, The Regular Intervals in the Equivalent Weight* of Elements^ 
1850, and Ann^ 1858, 105, 187 

4 Amar J. Set , 1854, [z], 17, 387. 

e Phi Mag., 1857, [3L 13 i 28o and 4 ao - 

Classification of the Elements 69 

bismuth giving sulphides insoluble in dilute acids ; chiom- 
mm, manganese, and iron giving hydroxides insoluble in 
ammonia ; calcium, strontium, and barium giving insoluble 
caibonates ; and lithium, sodium, and potassium having 
soluble salts. It has been suggested that Odling's system, 
levised in the light of existing knowledge, would furnish 
the most convenient system for the classification of chemical 
knowledge of the elements, but its inheient defect is its 
failure to indicate periodicity of chemical properties with 
atomic weights, and, had it been adopted, it is extremely 
doubtful if it could in any way have assisted in elucidating 
the structure of atoms, which has been deduced almost 
solely from the peiiodic classification. 

In 1862, de Chancouitois proposed a classification of the 
elements based on the new values of atomic weights conse- 
quent on Canmzzaio's system of 1858. De Chancourtois 
plotted the values of the atomic weights on a helical curve 
described on a cylinder, such that corresponding points on 
the cuive differed by sixteen, the atomic weight of oxygen. 
This curve, called by him the " vis tellunque " or " tellunc 
screw" brought into the same column many elements 
chemically closely related, and, in consequence, he sug- 
gested that " the propetties of the elements ate the popeities 
of numbets." 1 

In 1863, Newlands proposed a system of classification of 
the elements in the order of atomic weights, the elements 
being divided into seven groups having chemical properties 
analogous to the first seven elements, hydrogen, lithium, 
glucinum (sometimes called beryllium), boron, carbon, 
nitiogen, and oxygen. 2 Newlands teimed this septenary 
relationship the Law of Octaves, by analogy with the 
seven intervals in the octave of the musical scale. He 
further assigned to the elements in the order of their 

1 Compt rend , 1862, 54, 757, 840, 967 , 55, 600 , 1863, 56, 253, 467, 1217 , 
Fi* tellunque, classenient natiuel dcs corps umples oit ladtcanx obtenu an moyeit 
d'un systems de classification hthcoidal et nuinenqne, Pans, 1863 

- Chem Nevis, 1863, 7, 70, 1864, 10, 59, 94, 95, 240, 1865, 12, 83, 94, 
1866, 13, 113, 130 , On the Discoveiy of the Periodic Law^ London, 1884 

jo Chennstty and Atomic Structure 

atomic weights, a senes of numbeis beginning with hydio- 
gen I. This S7Stem of assigning ordinal numbers to the 
elements was revived fifty years later by Van den Broek, 
who identified the oidmal number with the charge on the 
atomic nucleus and with the number of electrons in the 
neutial atom, and the Atomic Number of an element is 
regarded to-day as an atomic constant no less important 
than atomic weight 01 valency. Newlands' classification 
and law of octaves were unfortunately regarded with 
extreme derision, and he abandoned fuither pursuit of the 
lemaikable relationships he had discovered. The attitude 
of chemists in this country to Newlands' proposals is a 
conspicuous example of the truth of the proverb about 
unhonomed native piophets, for an almost identical proposal 
six yeais later by the Russian chemist, MendeleefF, immedi- 
ately gained general acceptance. 

In 1864, Odling proposed a classification of the elements 
very similar to that of Newlands, and suggested that " Doubt- 
less some of the arithmetical relations exemplified in the fore- 
going table me merely accidental, but, taken altogether , they 
are too numetous and decided not to depend on some hitherto 
umecognued law" 1 Even Odling's authority, however, 
failed to make the proposals acceptable, and the law died 

In March, 1869, MendeleefT communicated to the Russian 
Chemical Society a paper, The Cot relation of the Properties 
and Atomic Weights of the Elements? in which he showed 
that " the elements arranged accotding to the magnitude of 
atomic weight show a penodu change of piopetties" and put 
forwaid a table of the elements arranged in groups practi- 
cally identical with those of Newlands and Odlmg. The 
groups in their three tables weie arranged horizontally, an 
arrangement alteied by MendeleefT in 1871 to the now 
familiar form of veitical groups and horizontal rows, or 

1 Watti* Dictiona>y, 1864, 3, 975 , Quart J Sci , [864, 1, 643 

2 J Rttss Cbem Soc , 1869, 1, 60 , 1870, 2, 14, 1871, 4, 25, and 348 




Classification of the Elements 

7 1 



O 3 


J3 bO 
Oi <J 

o o 


OO i- 














o 5 rt 











7 2 Chemistry and Atomic Stnittute 

In this table Mendeleeff diffeientiated the members of 
some of the groups into two classes according to whether 
they belonged to series with odd 01 with even numbei, 
group II, for example, containing two subgroups, the first 
with beryllium, calcium, strontium, and baimm in the 
series of even number, and the othei with magnesium, zinc, 
cadmium, and mercury in the series of odd number. This 
differentiation of the members of odd and even series is 
possibly moie remarkable than the differentiation into 
groups and series, and selves to emphasise the outstanding 
nature of MendeleefFs genius. He thus regarded some sets 
of two senes, one even and one odd, for example, 4 and 5, 
6 and 7, 8 and 9, and 10 and II, as together forming single 
penods of seventeen elements, seiies 2 and 3 being sepaiate 
gioups of seven each This diffeientiation has in very 
iccent years proved to be of fundamental impoitance in the 
determination of the electionic stiucture of atoms, being 
the sole known cuterion for the separation of the atomic 
elections into gioups containing the maximum possible 
number of electrons. Even more remarkable than the 
differentiation of the Periodic groups into even and odd 
series, was Mendeleeft's suggestion as to the so-called 
tiansition elements. It is commonly but wiongly stated 
that the tiansition elements are those of group VIII, and 
that they form a transition gioup fiom the even to the 
next odd series. This was not Mendeleeff's view, but that 
of lesser minds with weaker vision. He actually made use 
of the following words, " the following transition penod'. 
C? = 52 , Mn = 55 , Fe = 56 , Co = 59 , Ni = 59 ; 
Gu = 63 ; Zn 65 " , and the members of gioup VIII, 
the triads, Fe Co Ni, Ru Rh Pd, and Os Ir Pt, he teimecl 
" intei mediate seties of elements" His ' ' transition periods ' ' 
are almost identical with those in Bohr's theory of atomic 
structure, in which the transition periods repiesent the 
extension of an incomplete but stable gioup of electrons 
into a group with the maximum possible number of elec- 
tions. Bohr diffeied from Mendeleefl meiely in the piecise 

Classification of the Elements 73 

extent of the transition period, scandium, titanium, and 
vanadium, for example, being also included, and zinc 
excluded from the first tiansition peiiod. It is, however, 
piobable that Bohi is conect in the extent of the transi- 
tion series Evidence for this will be given in Chapters 
XIII and XIV. 

In December, 1869, ten months after the appearance of 
McndeleefFs first papei, Lothar Meyer x put forwaid views, 
probably independently, almost identical with MendeleefFs. 
Lothar Meyer appeals to have come to his conclusions 
laigely from the consideration of the physical properties ol 
the elements, wheieas Mendeleeff was guided almost entiiely 
by considerations based on the chemical propeities of the 
elements. To this extent they may be regarded as having 
discovered independent laws of the periodicity of the 
properties of the elements with atomic weight, the fiist 
the penodicity of chemical properties and the other the 
penodicity of physical piopeities Lothai Meyer's views, 
however, disclosed a much less cleai picture of the detailed 
nature of the penodicity, and completely disicgaided the 
fundamental difteientiation into even and odd series and 
the importance of the transition penods. The essential 
point common to both theories was the importance ot 
valency in the deteimination of periodic properties, 
and the recognition of the mciease in positive valency 
from I to 8 in passing from gioup I to group VIII, 
and the mciease in negative valency from I to 4 
in passing in the reveise direction from gioup VII to 
gioup IV. 

Following the foimulation of the Periodic Law and 
Classification by Mendeleeff, numeious variations in the 
foim of the periodic table weie proposed. The following 
table due to Bayley 2 is of mteiest as indicating cleaily the 
distinction between MendeleefFs even and odd series and 
his short and long penods. 

1 Ann Supf, 1870, 7, 35.) 
'-Phil Mag, 1882, [qj, 13, 26 


CbeniLtiy and Atomic Structure 

Classification of the Elements 


The peiiodic relationships weie set out by vanous 
wnters on a helix aftei the manner of de Chancourtois and 
on other forms of geometrical cuives, the underlying idea 
being to illustrate some assumed mathematical function of 
the increases in atomic weight. The majority of these 



formulations of the^peiiodic law offered no advantages over 
the tabular foim, with the exception of the logarithmic 
spiral curve of Johnstone Stoney, 1 shown in Diagiam II. 
This cuive has the advantages not only of representing the 

Cheat News, 1888, 57, 163, and Phil Mag , 1902 [6], 4, 411 

j6 Che mist} y and Atomic Structiue 

magnitude of the atomic weights and the peiiodicity < 
properties, especially valency, but left sesqui-iadius I 
unoccupied, this being filled in 1902 by the inclusion c 
the five recently discoveied " ineit gases," helium, neoi 
argon, krypton, and xenon, having zero valency and n 
chemical propeities. It is lemarkable that Johnston 
Stoney in 1888 did not discern something of the meanin 
of the unoccupied sesqui-iadius, which being a continuatio 
of radius 8 might have been predicted for the accommoda 
tion of from five to seven elements of valency zero or eighl 

In 1895, Thomsen, 1 as the result of an argument on th 
change in nature of valency from the electionegative urn 
valent halogen elements to the electiopositive univalen 
alkali elements, put forward the view of " the p>ob ability o 
the existence of a gioup of inactive elements " In the sam> 
yeai, argon, the commonest of the ineit gases, was dis 
coveied by Rayleigh and Ramsay, 2 and within six years th< 
lemaining foui (or five including niton, the ladiuni emana 
tion), ineit gases had been discoveied by Ramsay, whc 
created foi them a new group O in the periodic table 
Recent theoiies of the stiuctuie of atoms, however, indicate 
that these five or six inert gases would be more logical!) 
included in group VIII, despite their lack of valency, at 
being elements ma: king definite stages in the completion oi 
electronic groupings in atoms. In eithei case, MendeleefPs 
groups are increased fiom 7 to 8 for the shoit periods and 
from 17 to 1 8 for the long periods. 

Dm ing iccent years many vanations of Mendelcefr's 
periodic table have been put forwaid Most of these 
tables, being based on non-chemical theoiies of atomic 
structuie, in their anangement of the elements, cease to 
exhibit the fundamentally important point of classification 
of the elements accoiding to similarity in chemical and 
physical pioperties The piimaiy object of any system of 
classification is to condense a multitude of facts into gioups 

1 Zeit anoig Chew, 18911,9, 283 

2 Proc Roy Sot, , 1895, 57, 265 

Classification of the Elements 77 

containing closely related facts, in order to facilitate appre- 
hension, memory, and the utilisation of knowledge This 
object cannot be attained in chemistry by any system of 
classification which fails to include in a group the maximum 
number of elements exhibiting marked similarity in qualita- 
tive and quantitative chemical and physical relationships 
MendeleefFs classification involved the division of each long 
period into two chemically similar series, distinguished as 
even and odd seiies, this division thus bringing closely 
related elements with the same valency into the same 
periodic group. The variations of MendeleefFs table 
usually involve the extension of the long pet wds, so that the 
table is duplicated, as in Bayley's form, utilised by Bohr 
and otheis to exhibit resemblances in the structuie of atoms. 
This divorcement of the even and odd series fails to make 
apparent the extraordinary resemblance between phos- 
phates and vanadates, sulphates and chromates, perchlorates 
and permanganates, and between the " noble metah " of the 
eighth group and the " noble gases" and fails to make any 
contribution to the theoiy of atomic structure that is not 
more clearly apparent in MendeleefFs table. Table 3 
on page 79 very closely follows MendeleefFs arrangement 
of 1871, and enables not only the whole of the chemical 
iclationships between elements to be clearly discerned, but 
also serves as a basis for the interpretation of atomic struc- 
ture even moie complete than any inferred from the 
separation of the even and odd series of the long peiiods 
consequent on mathematical functions based on atomic 
numbers or atomic weights. 

The anangement of the elements shown in Table 3 is 
based solely on the classification of the elements according 
to similarities in chemical and physical properties, and is 
independent of any theory of the structure of atoms. 
Most of the chemical classifications of elements, based on 
atomic structure, give undue pi eminence to the " inert or 
noble gases? which are, in fact, relatively of little importance 
in nature and in chemical theory and practice. With the 

78 Chemistry and Atomic Structure 

exceptions of helium and neon, none of these elements is 
of direct importance in modern theories of the structure 
of atoms, foi only these two aie regarded as maiking com- 
pleted groups in electionic structures. The " noble metals" 
gold, platinum, indium, osmium, silver, palladium, rho- 
dium, and ruthenium, with which may be included copper, 
nickel, cobalt, and iron, are, on the other hand, not only 
important in nature and in chemistiy, but fundamentally 
important in theories of atomic stiucture, and mark 
decisively the points at which groups of electrons are com- 
pleted in the structure of atoms. The inclusion of both 
the " noble gases " and the " noble metals " in the same 
periodic group 8, is therefore necessitated in the classifica- 
tion of the elements accoiding to chemical pioperties and 
according to atomic structure. MendeleefT's division of 
the " long periods " into even and odd series of over fifty 
years ago is to-day abundantly justified, and, though he 
later failed to appreciate the close relation between the 
" noble metals " and the " noble gases" the soundness of his 
chemical instinct and the biilliance of his genius are almost 
unique in the history of the science. 

Classification of the Elements 



s ss a o 

J rG "aj <U 

P 60 



bo . fi V 

c a ^ - 

c u n p 

3 "^ 5 

*-* ra 3 r i 

Is 1 1 





" C Ji 

OT * "5 

I) J 



P O ~ oS 

t" l-i "U i-i 

>-. CB "i in 

T3 ^ ^ 


i i 


> u 1 



"S rt 

PH < w 







u ^ 

h N o ^ 








w a 



f 1 

JH ^ 1 

c/J r i-l 1 







OS fn 

cs b eo 

O w M 



" fi" 

FTJ rJ^ *^3 n r c3 *Td O r O 
o w o uj o r 5o 'o 

c r oC'r)C 1J n'2c 


p^CI P,C1fl(M P^Pl PH 




FRANKLAND'S recognition in 1852 of the definite numerical 
combining capacity of metallic and organic radicals laid the 
foundation for Gerhard t's theory of " hydiogen atomicity " 
in 1853, elaborated in 1858 by Cannizzaio, Kekule, and 
Couper into the modern theory of integral valency. Owing 
to the fact that Kekule's woik lay almost entirely in the 
field of carbon chemistry, his views on valency were coloured 
very largely by the evidence of the invariable quadnvalency 
of carbon, and he icgarded the valency of all elements as a 
constant atomic property. Confronted with trie existence 
of such compounds as the trichloride and pentachlonde ot 
phosphorus, he revived Berzehus' conception, of second 
01 dei combinations, phosphoius pentachlonde, PC1 5 , being 
regarded as an addition compound of a molecule of the 
trichloride, PC1 3 , and a molecule of chlorine, C1 2 . This 
conception ot " moleculat corn-pounds " was used to explain 
the existence of all chemical compounds to which a rational 
formula could not be assigned on considerations of fixed 
integral valency. Kekule's views continued to dominate 
the theory of complex compounds, hydrates, ammmes, 
polyhalides, double salts, complex acids, and molecular 
addition compounds generally, for over thirty years. 

The preparation in 1876 of the very stable gaseous 
phosphorus pentafluoride, PF 55 by Thoipe (T.E ), defi- 
nitely refuted the idea that valency is an immutable 
property of atoms, and thereafter, though valency con- 
tinued to be regarded as a numerically integral property of 
atoms, definite valency formulations were assigned to many 
compounds previously regarded as molecular addition 

One of the chief criteiia foi the distinction of true 
valency compounds from molecular addition compounds 
was the alleged instability or ease of dissociation of the 
latter as compared with the former. Many facts, however, 
accumulated to show that this criterion was invalid. 

Weme^s Co-oidination Theory 8 1 

Hexammino-cobaltic chloride, CoCl 3 .6NH 3 , for example, is 
veiy stable, whereas cobaltic chloride CoCl 3 is non-existent ; 
sodium platinichloride, PtCl 4 .2NaCl, evolves no hydro- 
chloric acid on treatment with concentrated sulphuric 
acid, though common salt, NaCl, a typical valency com- 
pound, is instantly decomposed , potassium ferrocyanide, 
Fe(CN) 2 4-KCN, yields no prussic acid, HCN, on treat- 
ment with cold acids, wheieas potassium cyanide, KCN, 
evolves prussic acid even with so feeble an acid as car- 
bonic , aluminium fluoride, A1F 3 , is decomposed instantly 
with violence by water, whereas sodium alurmnifluoiide, 
AlFg^NaF, is the inert mineial cryolite , boron tri- 
methide, B(CH 3 ) 3 , is a spontaneously inflammable gas, 
wheieas its ammine, B(CH 3 ) 3 .NH 3 , is sufficiently stable to 
be volatilised unchanged. Such instances of the stabilisa- 
tion of true valency compounds, by conversion into mole- 
culai addition compounds, can be multiplied a thousand- 
fold, and there is no leason to suppose that the natuie 01 
intensity of the forces causing combination in molecular 
addition compounds is difreientiable from those in tiue 
valency compounds. 

Consideiations such as the foregoing led Werner in 1891 x 
to propose a theory of chemical affinity and valency in which 
all chemical combinations were assigned to the same ulti- 
mate cause, the cieation of atomic attractive forces by the 
biinging into operation of integral valency. On the setting 
up of mtegial valency, forces arise on the atoms concerned 
in the valency exchange, each atom becoming the centre of 
a field of force gravitational in type if not in land. Sui- 
rounding atoms are atti acted to the suiface of a largei 
atom in number (the Co-ordination Number) sufficient to 
foim an enclosing shell of atoms, which might or might not 
be identical in numbei with the numencal value of the 
valency of the central atom. Werner's theory was oiigm- 
ally applied to the explanation of the valency of caibon, 
but, in 1893, he extended the theory to moiganic chemis- 

1 Beit/ age %ur Tbeoiie dei AJpmfai tiad Valenz, 1891 

82 Chemistty and Atomic Structure 

try, 1 and his " Co-ordination Theory " has found extende 
application m neaily every field of theoretical and practic 
chemistiy. Werner was not a meie thconst, foi every poir 
in his theoiy, accessible to experimental veiification, w. 
subjected to the most rigid examination. As Piofessc 
G. T. Morgan has stated, in his obituaiy notice of Wernei 
" he spated himself no pains to confirm these preconceptions I 
reference to cognate facts , 01, if these were lacking, by crucu 
laboratory tests. His Revision was so exact and well focusse 
that in the majonty of cases it was confirmed by e^penmeri 
. . . Werner's generalisation has all the attributes of 
scientific theory of first-rate importance. It is in cloi 
accordance with known facts, which it explains and summat 
ises in a logical and comprehensive manner. It endows it 
exponents with the gift of prophesy, and many fat -reachin 
predictions based on its simple hypotheses have been venfie 
subsequently by ditect experiment" 

The following genei ahsations emerge from the manifoL 
details of the work of Werner and his school : 

(a) valency is the numerically integial measuie of th 
capacity of an atom for combination with hydrogen eqm 
valcnts of atoms, being negative in combinations wit] 
electiopositive atoms, and positive m combinations wit] 
electronegative atoms , and is therefore the measure of th 
extent of reduction or oxidation of an atom , 

(b) chemical affinity is the attractive force brought mt< 
existence on atoms by operation of valency exchange, anc 
attracts and binds atoms whether 01 not concerned m th 
valency exchange , that portion of the distributed foio 
not utilised in binding atoms concerned in the valenc 
exchange is termed Residual Affinity, which, though of th 
same electrical polarity as valency, is measui cable, not n 
terms of hydrogen equivalents, but in teims of the numbc 
of atoms directly so bound ; 

1 Zeit anorg Chem , 1893, 3, 267, 294, 310 , Lehrbuch der Stereochemie, 1904 
Neure Anscbauungen auf dem Gebiete det anorgamschcn Cbemte, Brunswick too 
and 1909 , New Ideas on Inoiganic Chemistry^ London, 1911 

J J Chem Soc , 1920, 117, 1639 

Werner's Co-ordination Theory 83 

(<f) atoms bound to a central atom in virtue of valency or 
residual affinity may 01 may not dissociate fiom the central 
atom on solution of the compound in solvents ; 

(J) atoms, or groups of atoms united to a central atom 
thiough the medium of one of theii atoms, can dissociate 
from a cential atom to give lise to electrically charged ions 
only when such atoms, 01 gioups, owe their combination to 
their valency ; 

(<?) the maximum number of electrically chaiged uni- 
valent ions, having polarity opposite to that of a cential 
atom, that can dissociate from a central atom, is equal to 
the valency of the central atom ; 

(/) the number of atoms, 01 groups of atoms, non- 
ionisably associated with a central atom, is the Co-ordina- 
tion Number of the atom ; 

(g) the maximum number of univalent ions of the same 
polarity as that of a cential atom that can dissociate from 
a central atom is equal to the sum of the valencies of all 
the atoms attached other than by valency, and is usually 
equal to the diffeience between the co-ordination number 
and the valency of the central atom ; 

(ti) the intensity of the forces of attachment of co- 
ordinated atoms is a function of the valency of the co-ordi- 
nating atom, of the atomic volumes of the co-ordinated and 
co-ordinating atoms, and of temperature ; 

(i) the co-ordination number is usually a constant, but 
may be variable, and a function of atomic volumes and 
temperatuie ; 

(j) co-ordinated multivalent atoms may be simultan- 
eously the co-ordinating atoms of subsidiary co-ordination 
complexes, which may have ions associated therewith in 
addition to the ions associated with the main co-ordination 
complex ; 

(k) ionic dissociation of a co-ordination complex, not 
decomposed in solution, is possible only when the atoms 
combined exceed the co-ordination number ; if the number 
of atoms combined be less ihan the co-ordination number 

84 Chemistry and Atomic Structure 

the latter may be attained by addition of molecules of tl 
solvent, and any ions lesultmg may be derived fiom tl 
original compound 01 fiom the added solvent molecules 
but the number of univalent ions, not including the con 
plex ion, and not derived from subsidiary co-oidmatic 
complexes, cannot exceed the valency of the cential atom 

(/) co-ordination compounds may be decomposed o 
solution and some or all^of the co-ordinated atoms leplace 
by molecules of the solvent, such decomposition dependin 
on the relative intensities of the residual affinities of th 
co-ordinated atoms and the reactive atoms of the solver] 
molecules , 

(m) a multivalent atom 01 group, co-ordinated to 
central atom in virtue of valency, can yield the complet 
multivalent ion only on replacement by a molecule of th 
solvent , unco-oidinated valences of a multivalent atom o 
gioup may be associated iomsably 01 otherwise with othe 
atoms or with the co-ordination complex , a multivalen 
atom, or group, co-ordinated with a central atom by oni 
valence and associated with the central atom by an unco 
oidmated valence, is a bound ion ; 

(;/) a gioup of atoms may be co-ordinated to a centra 
atom by means of two of its atoms, the gioup thus occupying 
two co-ordination positions ; such doubly co-ordmatec 
groups may be bound by two valences, by one valence anc 
residual affinity, or by residual affinity alone ; 

(o) co-ordinated atoms are arranged in space round the 
central co-oidinatmg (nucleai) atom with the symmetry 
of one or other of the five regular solids, each co-ordinated 
atom occupying one corner of the imaginary regulai 
solid , 

(p) isomeric forms of a co-oidination complex are 
possible according to the symmetry of the imaginary solid 
from which they aie derived and according to the nature 
of the co-ordinated atoms ; 

(q) isomeric forms o co-ordination complexes aie optic- 
ally active when the complex does not contain a centre, 

Wetness Co-ordination Theoty 85 

plane, 01 an alternating axis of symmetry, and optically 
active isomeis exist in pairs which are i elated to one another 
as object and non-siipeiposable mm 01 -image ; 

(?) equal amounts of each of two oppositely optically 
active isomeis of a co-oidination compound may give lise 
in the solid state to a racemic compound, and such a com- 
pound may be icsolved into its generators by spontaneous 
crystallisation or by combination with another optically 
active compound , 

(/) apait fiom the capacity of atoms co-oidmated by 
valency to become charged ions on replacement by solvent 
molecules, no difference is detectable between the properties 
of atoms co-oidmated by valency and those of atoms 
co-ordinated by residual affinity. 

The foiegoing generalisations do not constitute an 
exhaustive presentation of the conclusions derived from the 
Co-ordination Theory, but suffice to present its broad 
outlines. Only a paitial statement of the evidence relative 
to these generalisations can be given heie, and the readei is 
refened foi fuller details to Weinei's publications and to 
Thomas' lecent book on Complex Salts. 

Though Werner's theory in its ongin was entirely non- 
electrical, it was eaily obvious that his views on valency as 
against residual affinity were scaicely more precise than 
those of the foundeis of the valency theoiy. Wernei, 
however, distinguished clearly that the bringing of valency 
into opeiation was not the immediate cause of the con- 
tinued existence of chemical compounds, but merely a 
process of exchange between atoms which resulted in the 
cieation of the actual forces causing and maintaining com- 
bination. Eventually he identified the exchange piocess 
of valency as the tiansfer of elections from 01 to an atom 
to or from othei atoms, the number ot electrons trans- 
ferred or acquiied being the integral measuie of valency, 
this transfer or acqunement of elections resulting in the 
creation of electric chaiges, recognisable only on ions 
formed in solutions. He identified chemical affinity with 

86 Chemistry and Atomic Structure 

the attractive force centred on an atom, and resulting fion 
the electron transfei, residual affinity being that poitioi 
of the force not utilised in binding atoms concerned in th< 
electron transfer. He postulated no definite mechanism bj 
which atoms bound in a co-ordination complex becam< 
incapable of ionisation without further chemical reactioi 

The existence and persistence of the co-ordination com 
plex was the central hypothesis of Werner's theory, but < 
masterly seiies of extending over thirty year 
conclusively demonstrated the reality of the three-dimen 
sional co-ordination complex in highly diversified types o 
complex chemical compounds, and his central postulate 
is to-day no longer a hypothesis but a verified law oi 
chemical combination. 

The types of compounds, to which Wernei's theoiy ha< 
been most widely applied, are the complex salts of metal 1 
and hydiogen with acidic atoms 01 groups such as the 
halogens, oxygen, sulphur, nitrogen, caibon, and boion 
and the oxy- and thio- acid radicals. Examples of these 
types aie HF.BF 3 ; HC1.NH, ; AlF 3 . 3 NaF ; PtCl 4 .2KCl 
CrCl 3 .6H 2 O; CuCl 2 .3Cu0.3H 2 O ; Sb 2 S 5 .3Na S 
BF 3 .NH 3 ; - CoCl 3 .6NH 3 ; Co,O 3 .Cl 2 Br 2 "8NH 3 .H~O 
Co 9 O 3 .2CoBr 3 .6HoO.i2NH 3 ; Fe(CN)o.CO 3 KCN 
B(CH 8 ) 8 .NH 8 ; A1C1 3 .H 2 S; AlBr a .3C 8 H 8 ; CoCO 3 N0 3 , 
5NH 3 . 

In order to symbolise the fact of the integrity of the 
co-ordination complex Wernci adopted the symbol of 
square Itackets to enclose the formula of the complex, a 
method of symbolisation long in use for the partition of 
formulas into radicals or compound ions. The co-ordina- 
tion formulas for the foregoing compounds thus become 
[BF 4 ]H ; [NHJC1 ; [AlF 6 ]Na 3 ; [PtCl b ]K 2 ; [Ci(H 2 O) c ]Cl 3 ; 

[Cu(^ Cu ) 3 ]ci 2 ,[SbS 4 ]Na 3 ;[BF 3 .NH 3 l[Co(NH 3 ) fl ]Cl 3 , 
[CoOH.Cl(NH 3 ) 4 ]Br , [GO {^ C (NH 3 ) 4 } , ] Br e , 

Werner's Co-ordination Theory 87 

[FcCO(CN) B ]K 3 , [B(CH 3 ) 3 NH 3 ] ; [A1C1 3 .SFL] , 
[AlBr(C 6 H 6 ) 3 ]Br 2 ; and [CoCO 3 (NH 3 ) 5 ] NO 3 

Inspection of these formulae shows that their pre- 
dominant feature is the presence of four or six atoms or 
groups associated with a central atom, metallic or electio- 
positive in natuie with the exception of the nitiogen atom 
in [NH 4 ]C1. Examination of the many thousands of 
co-ordination compounds now known reveals the facts that 
the piedominant co-ordination number is 6, and that the 
next most common number is 4, usually associated with 
the elements of small atomic weight such as Li, Be, B, C, 
and N, or elements of veiy large atomic weight such as 
Au, Hg, Tl, and Bi. 

The progiessive ionisation of co-ordination compounds 
by replacement of co-ordinated atoms or groups by mole- 
cules of a solvent (hydrolysis) is well illustiated by the 
triammme of cobaltic chloride [CoCl 3 (NH 3 ) 3 ]. On solu- 
tion of this compound in ice-cold water, ions are not at first 
detectable. On standing, one-third of the chlonne atoms 
become ionised, and the mono-hydrate, aquo-triammmo- 
cobaltic chloride, [CoCl 2 H 2 O(NH 3 ) 3 ]C1, can be isolated 
fiom the solution. On long standing, two-thirds of the 
chlorine atoms become ionised, and the dihydrate, 
diaquo-tnammino-cobaltic chloride, [CoCl(H 2 O) 2 (NH 3 ) 3 ] 
C1 2 , can be isolated from the solution. On raising 
the temperature, the whole of the chlorine atoms 
become ionised and the tnhydiate, tii-aquo-tnammino- 
cobaltic chloride, [Co(H 2 O) 3 (NH 3 ) 3 ]Cl 3 , can be 

The three molecules of ammonia can not be progressively 
removed by leplacement by the solvent owing to the 
instability of the resulting hydrates. The reverse process 
of replacement of the water molecules can however be 
accomplished, either by substitution by atoms previously 
ions or by replacement (ammonolysis) by further ammonia 
molecules, the following series of compounds being obtain- 
able by treatment with ammonia, diaquo-tetiammmo- 

88 Chemist? y and Atomic Structure 

cobaltic chlonde, [Co(H 2 0) 2 (NH 3 ) 4 ]Cl 3 , aquo-peut, 
mino-cobaltic chloride, [Co H 2 O(NH 3 ) 5 ]C1 3 , and h 
ammmo-cobaltic chloride, [Co(NH 3 ) 6 ]Cl 3 

The complete senes of compounds containing chloi 
atoms and water or ammonia molecules in the co-ordmat 
complex is also obtainable, tetiammino-cobaltic chlor. 
[CoCl a (NH 3 ) 4 ]Cl, aquo-tetiammino-cobaltic chlori 
[CoClH 2 O(NH 3 ) 4 ]Cl 2 , and pentammino-cobaltic chlori 
[CoCl(NH 3 ) 5 ]Cl 2 . All of the foregoing ten compou 
can be arranged into four types, according to the number 
acidic atoms in the co-ordination complex, tri-acido-ty 
[CoX 3 A 3 ], diacido-type, [CoX 2 Ai], mono-acido-ty 
[CoXA 5 ], and the non-acido-type, [CoA 6 ], " X " being ; 
univalent acidic atom or monobasic acid radical, and " j. 
any molecule containing one atom with residual affinity 
electronegative type. 

No other types of acido-co-oidination complexes 
possible with co-ordination number 6 and a trivalent a 
tral atom, though it is customary to legard complexes si 
as [Co(N0 2 ) ffl l [Co(N0 2 ) 6 NH 3 ], and [CotNO^NH, 
as separate types. They are however merely variants 
the tri-acido-type, in which three molecules of ammoi 
are replaced by three molecules of a simple salt such 
sodium mtnte, NaNO 3 , the complexes thus being asso 
ated with three ions, two ions, and one ion of a m 
valent metal or basic gioup, the compounds thus havi 
the formula [Co(NOo) 6 ]Na 3 , [Co(NO 2 ) 5 NH 3 ]Na 2 , a, 
[Co(N0 2 ) 4 (NH 3 ) 2 ]Na. 

The fact that the three, two, and one molecules of t 
added salt aie capable of independent lonisation folio 
from the circumstance that only one of the two atoms 
each added molecule is bound in the co-ordination cor 
plex. The hydrogen atoms of ammonia aie co-ordmati 
to nitiogen and co-ordinated ammonia molecules cons 
quently furnish no hydrogen ions. Molecules of c 
ordmated water, however, almost invariably fuimsh hydi< 
gen ions, and the aquo-salts are consequently always acid 

Wetness Co-otdination Theoiy 89 

in reaction. The compound, [Co(NH 3 ) 5 .H 2 O]Cl3, known 
as roseo-cobalt chloride, or aquo-pentammino-cobaltic 
chloride, can in fact be titrated with alkalis in precisely the 
same way as a mono-basic acid, and the result of neutralisa- 
tion (or hydrolysis) is the basic salt, [Co(NH 3 ) 5 .OH]Cl 2 , 
known as basic-purpureo-cobalt chloride, or hydroxo- 
pentammmo-cobaltic chloride. 

In order to indicate the fact that the hydroxide not a 
chloride radical is bound non-ionisably in the co-ordination 
complex, Werner adopted the nomenclature used in similar 
cases in organic chemistry, the acidic atom substituted in a 
" nucleus " being given the suffix " 0," for example chloro- 
methyl-cyanide, C1CH 2 CN, bromo-acetyl-chlonde, 
BiCH, -COC1, and nitro-benzoic acid, NO 2 C G H 4 COOH. 
Hence the use of such names as chlor0-pentammin0-cobaltic 
chloride for [Co(NH 3 ) 5 Cl]Cl 2 , and hydrox0-aquo-tetram- 
mino-cobaltic chloride for Co[(NH 3 ) 4 .H 2 O.OH]Cl 2 . 

However useful this nomenclature may be for these 
classes of compounds, it fails when applied to the tri-acido- 
type of co-oidination compounds. [Co(NH 3 ) 3 Cl 3 ], for 
example, must become trichloro-triammino-cobalttf. To 
obviate this, Werner was driven to regard the cobalt as 
acting acidically, and called the compound trichloro- 
triammino-cobaltwte. This nomenclature is quite un- 
necessary, for the simple name triammino-cobaltic chloride 
conveys all that need be understood of the chemical nature 
of the compound. The old name potassium ferrocyanide, 
Fe(CN) 2 4KCN, in Werner's system, becomes potassium 
hexacyano-ferroate, [Fe(CN) 6 ]K 4 . Werner's nomenclature 
is unfortunate in that it conveys nothing that the older and 
much simpler nomenclature does not convey, and he was 
frequently obliged to suppress the veiy facts that his 
nomenclature was invented to provide for, the lomsation of 
paits of a complex. Strictly his name aquo-pentammino- 
cobaltic chloride should be hydroxo-pentammino-cobaltic 
dichloride hydrochloride, for the aquo-group furnishes a 
hydrogen ion in solutions. 

cp Chemistty and Atomic Structure 

Methyl ferrocyanide, Fe(CN) 2 4CH 3 CN, termed in 
Wernei's system methyl hexacyano-fenoate, should be 
tricyano-tn-cyanomethyl-fenoate, foi the foui methyl 
cyanide molecules no moie furnish methyl ions than 
ammonia molecules furnish hydiogen ions. The non- 
ionisation of the methyl groups explains the existence of 
the two isomeis a- and p-methyl ferrocyanides, as the cis 
and trans forms of an octahedral complex. It may be 
stated that any system of chemical nomenclature based on 
ionisation is bound to fail, foi the simple reason that 
lonisation may have any value from infinitesimal to com- 
plete, depending on the solvent and the extent of dilution 
of the solution 

A stionger objection to Weinei's nomenclatuie is the 
fact that it is not consistent with his co-ordination theoiy 
in which metallic ammmes aie regarded as ammonium 
derivatives. If ammonium chloiide is validly describable 
as the pioduct of the addition of the electro-positive pait 
of a simple salt, hydrogen chloride, to the electro-negative 
nitrogen of ammonia, then metallic ammmes are similar 
ammonium compounds, consisting of the product of the 
addition of the electro-positive part of a simple salt, 
cobaltic chloride, to the electro-negative nitrogen of 
ammonia. CoCl 3 3NH 3 is consequently cobaltic-tn- 
ammonium chloride, just as CH 3 C1.NH 3 is methylammon- 
lum chloride in oiganic chemistry. In the strict ammon- 
ium nomenclature CoCl 3 .3NH 4 Cl would be ammonium 
cobaltic-dilonde, CoCl 3 .NH 3 .2NH 4 Cl would be diam- 
monium cobaliic-ammonium chloiide, and CoCl 3 .2NH 3 . 
NH 4 C1 would be ammonium cobaltic-diammonium 

Werner's use of the term ammino is a convenient short- 
hand form for co-ordinated ammonia, but no justification 
exists for the use of such teims as cobaltiate and feiroate, a 
method of nomenclatuie already in common use to desig- 
nate oxy-acids. The confusion ensuing from Werner's 
nomenclature is illustrated by the fact that the common 

Werners Co-ordination Theory 91 

oxy-salt potassium chloro-chromate, KCrO 3 Cl, would 
scaicely be differentiated by name from the double salt, 
CrQ 3 .3KCl, at present known as potassium chromi- 

The foiegoing three geneial types of co-oidination 
complexes derived from tnvalent metals include those in 
which a single group of atoms contributes more than one 
to the six co-ordination positions, and numerous cases are 
known of doubly-bound groups, and at least one trebly-bound 
group has been identified by Werner. 

Groups of the type of ethylenediamine, NH 2 .CH 2 .CH 2 . 
NH 2 , and metallic hydroxides, e.g. Cu(OH) 2 , are equivalent 
to two molecules of ammonia. Groups of the type of the 
aminoacetate group, NH 2 .CH 2 .CO 2 , and the acetyl- 
acetone radical, O : C.CH : C.O , aie equivalent to one 

CH 3 CH 3 

molecule of ammonia and one univalent atom. Groups of 
the type of the oxalate gioup, O 2 C.CO 2 -, and the 
pentamethylene radical, CH 2 .CH 2 .CH 2 .CH 2 CH 2 , are 
equivalent to two univalent atoms. 

Piofessor G. T. Morgan has proposed to designate all 
groups, contributing two positions to a co-ordination com- 
plex, Chelate Groups. Complexes containing thiee of 
the first type of chelate group belong to the general non- 
acido-type, [MA ]. Those containing one chelate gioup 
of the second type and four molecules of ammonia belong 
to the general mono-acido-type, [MXA 5 ]. Those con- 
taining two such chelate groups and two molecules of 
ammonia belong to the general di-acido-type, [MX 2 A 4 ] 
Those containing three such chelate groups belong to the 
general tri-acido-type, [MX 3 A 3 ]. Complexes containing 
one chelate group of the third type and four molecules ot 
ammonia belong to the general di-acido-type, while those 
containing two or three such chelate groups belong to the 
general tri-acido-type. 

It is obvious that a chelate group by combination with a 
central atom gives rise to a cyclic system, for example, 

92 Chemistry and Atomic Structure 

,NH 2 -CH 2 

Co^ . In cases wheie two or moie chelale 

groups combine with a cential atom, the latter becomes 
the common membei or j^z?0-atorn of two 01 more cyclic 
systems, for example, in cupric aniino-acetate, and basic 
cupric chloride, 


r /OR- ,-OH\ 
c / ^ . \, 


H2 \CO,/\0,C/ 2 OH OH 


In the first case the valency of copper is expended in the 
co-ordination complex, whereas in the othei the two 
chloride ions associated with the complex ion account for 
the valency of copper, the complex ion thus being a diacidic 
basic radical, in accordance with the lule that the maximum 
number of univalent ions associated with a complex is equal 
to the valency of the co-oidmating atom. 

It was indicated, page 87, that ammonium chloride 
differed from the co-ordination compounds of metals in 
containing a central electronegative atom, its co-oidmation 
formula being [NH 4 ]C1. This compound may, howevei, 
be regarded as an ammine of electiopositive hydiogen 
[H.NH 3 ]C1, and is thus strictly analogous to a metallic 
ammine. Either method of formulation involves that tlic 
valency of nitrogen is unchanged in ammonia and ammon- 
ium salts. This can readily be demonstrated experi- 
mentally. On oxidation of ammonium chloride to watci, 
free nitrogen and hydrochloric acid, only three equivalents 
of oxygen are necessary. 

2 NH 4 C1 -f 3 = 3 H 2 + N 2 + 2 HC1 
On oxidation to water, free nitrogen, and fiee chlorine, 
four equivalents are necessary. 

>2 NH 4 C1 + 4 = 4 H 2 + N 2 + C1 2 
On oxidation to water, nitric acid, and hydiochloric acid, 
eight equivalents are necessary. 

Werners Co-ordination Theory 93 

NH 4 C1 + 40 - H 2 + HNO 3 + HC1 

On oxidation to water, nitric acid, and fiee chlorine, nine 
equivalents of oxygen aie necessary. 

2 NH 4 C1 + 90 = sH 2 + 2 HN0 3 + C1 2 

If ammonium chloride contained quinquevalent nitrogen 
the equivalents of oxygen necessary for these oxidations 
would be increased by two in all cases. If the nitrogen 
atom were electropositively quinquevalent as in nitric acid, 
then no oxidation would be necessary, and hydrolysis alone 
would yield nitiic acid, hydrogen, and hydrochloric acid. 
Ammonium chloride, however, on complete hydiolysis 
yields only ammonia and hydrochloric acid. 

It must therefore be concluded that ammonium salts 
contain not merely tuvalent nitrogen but electronegative 
mttogen, and that the nitrogen atom is completely reduced 
just as is the oxygen atom in water and the chlorine atom 
in hydrochloric acid. It may thus be inferred that the 
nitrogen atom in ammine combinations is equivalent in 
type of affinity to oxygen, chlorine, and othei acidic atoms, 
and that the metallic atoms of co-ordination complexes 
possess electropositive valency and electropositive residual 
affinity. Hence it is that the co-ordinated atoms in metallic 
co-ordination compounds are almost invariably nitiogen, 
oxygen, fluorine, chlorine, biomine, and iodine, and that 
metallic co-ordination compounds are usually ammmes, 
hydrates, complex oxy-salts, oxides, and double halides. 
Co-ordination compounds of the ammonium type, on the 
other hand, aie usually hydiides 01 carbon compounds with 

The so-called boronium, thallonium, carbonmm, silicon- 
ium, titanonium, stannomum, stibonmm, bismuthonium, 
and iodomum compounds exhibit no real analogy to 
ammonium compounds, but contain as truly electiopositive 
central atoms as do the metallic ammines, and the same is 
probably true of many sulphonmm, phosphonium, and 
arsonium compounds. Even the nitrogen atom, so fre- 
quently highly electionegative, becomes an electropositive 

94 Cbemistty and Atomic Stnictuie 

cential atom in combination with oxygen and peiliaps th 
halogen elements. 

An interesting example of the direct combination c 
electropositive nitiogen with electronegative nitiogen, bot 
in the trivalent state, may be regarded as furnished by on 
of the nitrogen iodides, N 2 H 3 I 3 , to which the co-oidinatio 
formula [NI 8 .NH 3 ] may be assigned, each nitrogen conti. 
buting one position of opposite polarity to the four cc 
ordination positions of the other. The leactions of th 
compound, however, indicate that this elegant formulatio 
is doubtful, as nearly the whole of the nitrogen is evolve 
fiee on hydrolysis, and the compound reacts with moi 
ammonia to form NI 3 .2NH 3 and NI 3 .3NH 3 . It seem 
probable, therefore, that the nitiogen atom is electic 
negative even in combination with the halogens, and the 
the nitrogen iodides contain electropositive iodine, th 
constitution of the last compound thus being [(NH 3 .I) 3 N 
only three of the four electionegative nitrogen aton 
attaining the co-oidination number four. 



THE spatial distribution of the forces of chemical affinity- 
consequent on valency exchange, necessarily involved m 
Werner's theory the spatial distribution of atoms about a 
central atom. The spatial distubution of atoms united to 
the quadrivalent carbon atom was the essential feature of 
the theories in organic chemistry associated with the names 
of van't Hoff and Le Bel, and Werner's co-ordination 
theory fuithei extended steieo-chemical conceptions in 
organic chemistry and to inorganic chemistry generally. 
Werner postulated that all atoms with co-ordination 
number four possessed a tetrahedral configuration, and those 
with co-ordination numbei six an octahedral configuration. 

It is extremely doubtful, however, that Werner's postu- 
late is universally correct, for he was compelled to admit a 
planar distribution of co-ordination positions for the 
bivalent platinum atom with the co-ordination numbei 
four, though the six co-ordination positions ot quadiivalent 
platinum were definitely demonstrated to be not only 
spatial but octahedral. It is more than possible that many 
atoms with co-ordination number four possess a planar 
distribution of valency forces. Vernon 1 has contnbuted 
considerable evidence foi the planar distribution in the case 
of the dimethyl halides of bivalent tellunum, and Morgan 
and Diew and their co-workers have detected no stereo- 
isomeis m the case of the di-selemum acetylacetones, 
though such should exist if the selenium atoms in these 
compounds have not a planar distribution of valency foices. 
With legard to atoms having the co-ordination number 
six, Werner was able to demonstrate their octahedral 
stiucture in all the cases examined, and until lecently it 
appeared that his generalisation of octahedral structure for 
six co-ordination positions was correct m all cases. The 
tnacetylacetones of the tnvalent metals however, should 
exist in two optically active forms, and, though no evidence 

1 J Chew Soc , 1920, 117, 86, 889 , 119, 105, and 687 


Opt/oaf Isomers 

C/ass /(v) w/thout Chefete Groups 


C/ass?(iv)w/th one unsymmefcr/cd/ Che/ate GroLp 



C/ass 30v)wtth two s/m//dr unsymmebr/ca/ Che/ate Groups 

C/ass 3(v)wth two d/ss/mi/ar unsy/nmetncdl Cheldte Croups 

Co-ordination Stereo-Chemistry 97 

has been obtained to this effect, it has been assumed that the 
known compounds are meiely unresolved racemic compounds. 
The researches of Mr. W. T. Astbury, however, prove 
that this explanation is untenable. His examination of 
numerous triacetylacetones by ciystallographic and X-iay 
analysis indicates that the octahedral stiucture is impossible, 
and that the spatial arrangement is that of the trigonal 
prism. This deduction in no way affects the reality of the 
octahedral stiuctures assigned by Werner to chromium, 
iron, cobalt, rhodium, indium, and platinum atoms as the 
result of his chemical investigations. It merely proves that 





None One Two 

Type i 2 3 

Class (i) aaaa AA bb AA AA 

(n) aaab AB cc AA BB 

(m) aabb AA be AA BC 

( 1V ) aabc ABcd* AB AB* 

(v) abed* ABCD* 

* Represents two optically active isomers 

the octahedron is not the invariable six-point structure for 
these and other atoms. The new evidence indicates that 
the regular octahedron and the trigonal prism are limiting 
types of atomic six-point structure, and that Werner erred 
only in a too-inclusive formulation of his postulate as to 
spatial distribution of co-oidmation positions. The impor- 
tance of Werner's contribution to the theory of moleculai 
and atomic structure is not theieby materially affected. 

Examination of the symmetiy of the tetiahedron indi- 
cates that tetrahedral co-ordination complexes may be 

98 Gbemistiy and Atomic Sttuctwe 

classified into three main types and fourteen classes 
according to the number of chelate groups associated wit! 
the cential atom. Representing a co-ordinated chelafr 
group by two capital letteis to symbolise its symmetiica 
or unsymmetrical nature, and atoms or groups, occupyin; 
only one co-ordination position, by small letters, the thre< 
types and fourteen classes of tetrahedral co-ordmatioi 
complexes are shown in Table 4, page 97. 

Of these fourteen classes only I (v), 2 (iv), 3 (iv), am 
3 (v) represent two isomeric complexes, i.e. complexe 
related to one another as object and non-supeiposabl 
mirror-image, and at least one of these complexes must b 
present in any compound exhibiting optical activity. Th. 
eight optically active complexes, in four mirror-image pairs 
are shown in Diagram III, p. 96. The fust three optical! 
active classes have been demonstrated to exist, though th 
two isomeric compounds of the 3 (iv) class (symmetiica 
spiro-nuclcus) have only iccently been obtained for th 
first time by Mills and Noddci * by the optical resolution 
of the lacemic ortho-dilactone of benzophenone tetia 
carboxylic acid. As yet no compounds of the 3 (v) clas 
(asymmetric spiro -nucleus) have been obtained optical! 
active, though little doubt can be thrown on their possibl 
existence. The fact that thirteen of the fourteen possibl 
classes have been proved to exist and to give rise to th 
theoretical number of isomers rendeis impossible any othe 
than the tetrahedral nature of the quadiivalent carboi 

Some of the foregoing optically active classes have bee] 
proved to exist containing nitrogen, sulphur, tin, silicon 
phosphoius, aisenic and selenium as co-ordinating atoms 
and the tetrahedral nature of these atoms is theiefor 
(see p. 46) certain. Most of the optically inactive classe 
have been pioved to exist containing lithium, beiyllium 
boron, oxygen, chlorine, biomine, iodine, gold, mercury 
thallium, and bismuth, and the tetiahedral nature of thes 

1 J Chem. Sac., 1921, 119 a 2094 

Co-ordination Stereo- Chemistry 99 

atoms is scaicely less certain than in the cases of the fore- 

The foregoing three types and fourteen classes of tetra- 
hedral co-ordination complexes relate only to uni-nuclear 
compounds, i.e. those containing one central atom, but all 
compounds containing any number of centi al co-ordinating 
atoms are simply denvable from these types and classes by- 
combination. As no tetrahedral nuclear atom regaided 
separately can give lise to more than two isomeric forms, 
both of which are equally and oppositely optically active, 
two such nuclear atoms cannot give rise to more than 
2x2=4 binuclear isomers, three nuclear atoms to no 
more than 2X2 x 2 = 8 trinuclear isomers, four to no 
more than 2x2x2x2 = 16 quadrimi clear isomeis, and 
in general n cential atoms to no more than 2" multinuclear 

If two similar nuclear atoms are situated symmetrically 
in a chain of atoms, the chain can be divided into two equal 
poitions each containing one of the nuclear atoms. One 
uninuclear half-chain is capable of existing in two isomeric 
forms, and each of these forms is identical with an isomeric 
form of the othei half-chain, or, in othei woids, the two 
uninuclear half-chains yield only two sorts of forms, related 
to one another as optical isomeis. These two half-chains 
can be combined together to form a complete binuclear 
chain in three different ways, the first consisting of two 
identical half-chains of one sort, the second of two identical 
half-chains of the other sort, and the thiid of a half-chain 
of each sort. The first two binuclear chains are equally 
and oppositely optically active, and the third optically 
inactive. These three isomeiic foims aie the maximum 
number of isomers possible with two similar nuclear 
atoms symmetiically situated in a chain of atoms. 

Such inactive forms are described m organic chemistry 
as mesa-forms, and are said to be inactive by internal com- 
pensation^ meaning theieby that the half-molecules, of 
which they may be regaided as composed, are equally and 

roo Chemistty and Atomic Structure 

oppositely optically active and consequently reduce the 
activity of the whole molecules to zero 

It is obvious that the number of meso-foims correspond- 
ing to any formula must be equal to half the number of 
optical isomers, that is, equal to the numbei of pairs of 
such isomers. (Optical isomers are usually described as 
enantiomoipbs, or enantiomorphic pairs.) Meso-forms can 
only exist when a formula can be divided into two equal 
portions, and, in general, meso-forms are equal in number 
to one quarter of the number of isomers theoretically 
denvable from the total number of nuclear atoms in the 
formula. The number of corresponding optically active 
isomers is equal to one-half the theoretical total number of 
isomeis. If the numbei of nuclear atoms, capable of 
giving rise to isomers, be n and the formula be symmetrical, 
the number of enantiomoiphs is equal to one-half of 2", 
= z"' 1 . As the number of meso-foims is one-half of the 
number of enantiomorphs, the number of meso-foims is 
one-half of 2"" 7 , = 2 n ~ 2 , and the total number of isomers, 
optically active and inactive, is thus equal to 3 x 2 n ~ 2 . 

The simplest case of two nuclear atoms, symmetrically 
situated in a chain of atoms, is illustrated by the thiec 
isomeric taitaric acids shown in Diagram IV, two of these 
isomers being enantiomorphs and the other an inactive or 
meso-form. It is usual to add to these three forms a 
fourth form known as racemic acid, a compound of equal 
numbers of molecules of the two enantiomorphic isomers. 
Strictly, racemic tartanc acid is not an isomeric taitaric 
acid, for its molecular weight, were it ascertamable, must 
be at least double that of any of the three truly isomeric 
taitaric acids. Racemic compounds do not owe their 
existence to different airangements of the atoms in mole- 
cules, but to combination between two diffeient molecules 
which aie enantiomoiphs, the number of racemic foims 
thus being inevitably equal to the number of enantio- 
morphic pairs of isomers. 

A special type of isomerism arises in the case of com- 

Co-otdination Steteo-Chemistry 


pounds containing two nuclear atoms, each of the sym- 
metrical mono-chelate class 2 (m), when the chelate group 
is common to both nuclei, and each nucleus is symmetrically 


Meso or 
T&rtsr/c Actd 

C/S form Trans Form 

Isomers of 

Two 5/milar Tekrahecf&l Systems wtth one Chelate Group in Common 

Cis and Trans Isomers of 
Two Dlsstmtbr Tebnahedra/ Systems w/th one Che/ate Group m Common 



C/S form 
Ma/e/c Ac/cJ 


Trans form 

situated in the chelate group of the other. Owing to the 
fact that class 2 (m) involves the nucleai atom being 
attached to two unlike atoms or groups, the chelate group 
can be applied in two different ways to the nuclear atom, 

IO2 Chemistry and Atomic Structure 

the two dissimilar atoms in the nucleus in the chelate group 
in one case being reversed in " orientation " (relative 
position in space). If the two nuclei are identical, these 
two positions coi respond to bringing similar atoms on the 
same side of the molecule, called the cis position, 01 on 
opposite sides of the molecule, called the trans position. 
If the two nuclei are not identical, the terms cis and 
trans can be applied only arbitiarily. The two cases are 
illustrated in Diagram IV, page 101. 

As the nuclear atoms belong to class 2 (iii), each nuclear 
system must separately contain a plane of symmetry, and 
as the symmetrical chelate group is common to both 
systems, the plane of symmetry must be identical for both 
nuclear systems. As complexes containing a plane of 
symmetry cannot be optically active, cis and trans isomers 
cannot be optically active, unless another nuclear atom is 
present belonging to one of the classes I (v), 2 (iv), 3 (iv), 
and 3 (v), in which event there would exist at most four 
isomers, two cis-enantiomorphs and two trans-enantio- 
morphs. Numerous cases are known in the cyclic structures 
of organic chemistry. 

The limiting cases of the cis and trans type of isomerism 
is reached when the chelate group is reduced to one atom, 
the isomers then consisting of two nuclear atoms united by 
a double-bond This is the most familiar case of cis and 
trans isomerism in organic chemistry, and includes open- 
chain and cyclic ethylene derivatives, oximes, hydrazones, 
and diazonium compounds. The ethylene type of iso- 
merism in maleic acid and fumaric acid is illustrated in 
Diagram IV, page 101. 

In organic chemistry it is usual to lefei to an atom, such 
as a carbon atom, as an asymmetric atom, when, acting as a 
nuclear atom of a tetiahedral system, it is so combined as 
to give rise to optical isomers 01 enantiomorphs. This 
term, due to van't Hoff, has no justification on the chemical 
evidence, for phenomena of optical activity can aiise even 
if the carbon atom be as perfectly symmetrical as the 

Co-otdwation Stereo-Chemistry 103 

regular tetrahedion The only asymmetry that has ever 
been detected in chemistry is the asymmetry of molecules 
as a whole, such asymmetry arising merely from the spatial 
distribution of atoms 01 groups about a central atom, which 
may itself have any symmetiy elements whatever. There 
is in fact good reason to suppose that the carbon atom in 
all of its combinations possesses at least one of the six planes 
of tetrahedral symmetry. 

The symmetry of the octahedron is much more complex 
than that of the tetrahedron, and octahedral co-ordination 
complexes aie consequently much moie varied in type. 
Four main types are possible accoiding to the number of 
chelate groups united to the central atom, the fiist type 
being non-chelate, the second mono-chelate, the third 
di-chelate, and the fouith tri-chelate. 

These gioups may be further divided into foity-one 
classes, eleven of the non-chelate type, and ten each of the 
three chelate types. The eleven non-chelate classes give 
rise to seventy-five forms, of which fifty are optically active. 
The ten classes of the mono-chelate type give rise to 
seventy-two forms, of which fifty-eight are optically active. 
The ten classes of the di-chelate type give rise to eighty 
forms, of which seventy are optically active. The ten 
classes of the tri-chelate type give rise to fifty-foui foims, 
all of which are optically active. The octahedral con- 
figuiation thus gives rise to two hundred and eighty-one 
forms, of which two hundred and thirty-two are optically 
active in one hundred and sixteen enantiomoiphic pairs. 
The following tables 5, 6, 7, 8, and 9, detail the various 
classes and foims. 

Of the forty-one classes of octahedral complexes, twenty- 
two have been experimentally leahsed, the first nine of 
type I, the first five of types 2 and 3, and the fiist two and 
the fourth of type 4. None of the classes of type I have 
yet been obtained in optically active forms, but the inactive 
foims correspond to the theoretical numbci in nearly all cases. 

None of type 2 had been obtained in optically active 


Chemistry and Atomic Stnicture 





Class (i) aaaaaa 

(n) aaaaab 

(in) aaaabb 

(iv) aaaabc 

(\) aaabbb 

(vi) aaabbc 

(vn) aaabcd 

(vm) aabbcc 

(ix) aabbcd 

(x) aabcde 

(xi) abcdef 








AA bbbb 



AB cccc 

AA BB cc 


AA bbbc 

AA AA be 


AB cccd 

AA BB cd 


AA bbcc 

AA BC dd 


AB ccdd 

AB AB cc 



AB CD ee 


AB cede 

AB AB cd 


AA bcde 

AA BC de 


AB cdef 

AB CD cf 





















Optically Active 
Trans Other Total 







Optically Inactive 
Trans Other Total Total 


Total 20 Nil 30 50 

2 25 75 

Co-otdmation Stereo-Chemistiy 




Classes Optically Active 

Cis Trans Other Total 


Optically Inactive 
Trans Orher Total Total 



I I 

i i 
i i 



I I 





i _ *n- 






_ , sy 


. *> 





















12 12 



24 24 

2 4 




36 58 


2 14 











Optically Active 

Optically Inactive 



Other Total 

Cis Trans 

Other Total 






^ y 




. fj 

_ ,_.. T 

__ T 






^ y 






__ y 








































Nil 70 

Nil 10 

Nil 10 



Chemistiy and Atomic Structure 

forms, until Thomas 1 in 1923 obtained the cis-dmitio- 
oxalato-diammino-cobaltic complex, of class 2 (v) in two 
optically active and one trans-dinitro-inactive form, 

Stereo -Isomers 

** 'C-*m 

c/s - 

Opt/cat/y snack/re 
C/s-amm/no -/somer tnans-ammmo -/somer 

being three of the foui theoretically possible cis and tians 
isomeric forms of this class. The correct number of iso- 
mers has been obtained for classes (i), (ii), (ih), and (v) of 
type 3, by Wernei, but the optically active isomers of 
class 3 (iv) have not yet been obtained. Werner was also 

1 J Soc , 1923, 123, 617 

Co-ordination Stereo-Chemistry 








Cis. Trans 

2 2 

2 4 
4 4 


Other Total 

2 2 
2 2 
2 2 
2 2 

4 4 



8 8 
16 16 


Optically Inactive 
Trans Other Total 




Total 8 10 36 54 Nil Nil Nil Nil 54 

successful in obtaining the conect number of isomers foi 
seveial lepiesentatives of classes 4 (i), 4(11), and 4 (iv), but 
in the case of the last class Werner was unable to resolve 
into the optical isomeis the racemic compounds containing 
only one nucleai atom, which were howevei obtained by 
synthesis from optically active compounds. Racemic com- 
pounds of this class, with only one nuclear atom, have only 
this year been lesolved by Moigan and Main Smith 1 into 
the coirect number of optical isomers, the compounds 
resolved being salts of the sahcylato-diethylenediammino- 

(/Nlrf \ /Q C 6 ri4 
C 2 H 4 \ xTTT 2 |2C\ j (see Dia- 

^JNH 2 ; _ co _ 

gram VIII page ill) 

The great majonty of the foiegoing classes actually known 
aie complexes containing the cobalt atom as the central 
co-ordinating atom, but many of the classes are also known 
containing chromium, iron, rhodium, and indium as 

1 J Chem Soc , 19245 125 


Chemistry and Atomic Stiuctme 

nuclear atoms, and all of them have been proved to give 
rise to optically active isomenc forms These atoms have- 
therefore quite definitely a six-point spatial stiucture, which 
is octahedral. 



type? C/ass(/r) 


Bromide of 
C/-S - di - ch/oro -/somer 

Chfor/de of 

C/s ~ 

Bromide of 
trans - d/- ch/oro- /somer 

Ch/or/de of 
trans-ch/oro bromo- /somer 

/^ ac/cfo - tetr&mm/no coba/c/C ha/fdes 

Though no optically active isomeis are known in the 
cases of magnesium, aluminium, silicon, phosphoius, t,ul- 
phur, scandium, titanium, vanadium, manganese, nickel, 
copper, zinc, gallium, geimanium, selenium, >ttrmm, 
zirconium, columbium, molybdenum, ruthenium, palla- 

Co-ordination Stereo-Chemistry 


lonisdtton Isomers 

. X>&-f# 



OptfCd//y active 



Opt/cal/y active bromide of 



Type? C/ass(///J 

Optical fy act we c/)/or/cfe of 

I ^ . ** - 

Optically active ch/orideof 
c/s-chloro - 

Opttcd!Jy inactive bmm/de of Opk/cd//y tnact/ve chhr/de of 

trans- dhchloro-fsomer trans-ch/oro-bromo~/somer 

Dt-audo^diethylened/ammino-cobatUc hafides 

no Chemistry and Atomic Structure 

dium, silver, cadmium, indium, tin, antimony, telluriur 
lanthanum and the rare earth metals, celtium, tantalur 
tungsten, osmium, platinum, lead, thorium, and uianiur 
a sufficiently large number of the inactive forms of many 
the classes are known to render practically certain that z 
these elements can possess a six-point spatial atomic stru 
ture, which is probably octahedral. 

The foregoing forty-one classes of octahedral co-oidin 
tion complexes do not include the whole of the possib 
isomeric compounds, but represent only those derived fro 
one co-ordination complex ii respective of the ions wil 
which the complex may be associated. Many dozens i 
cases of isomeric compounds are known in which the is< 
merism is a consequence of the position of atoms or grou] 
in 01 lonically associated with a complex ion ; dichlor< 
tetrammino-cobaltic bromide, [Co(NH 3 ) 4 Cl 2 ]Br, for e: 
ample, lepresents two isomers of class i (in), the cis AT 
trans forms of the complex ion, both of which are isomer 
with the two con esp ending isomers of class I (iv) of tl 
complex ion of chloio-bromo-tetrammmo-cobaltic chloiid 
[Co(NH 3 ) 4 Cl.Br]Cl. These pairs of isomcis, lonisatu 
losmers, derived from two different complex ions, are shott 
in Diagram VI, page 108. 

If the four ammonia molecules be replaced by two moL 
cules of ethylenediamine, the dichloio-bromide can exi 
in three isomeiic forms, two cis and one trans, of class ^ (\ 
which are ionisation isomeis of the corresponding thn 
isomers of class 2 (m) of the chloro-bromo-chloride, the c 
pairs in each case being optically active as enantiomorph 
pairs. These six isomers are shown in Diagiam VII, page I0' 

Representatives of type 4 are incapable of exhibitir 
ionisation isomerism due to un*valent ions, but may exhib 
ionisation isomerism due to bivalent ions if one of tl 
chelate groups is bivalent. The two optically active isc 
mers of salicylato-diethylenediammino-cobaltic bicarbonai 
of class 4 (iv), for example, are ionisation isomers of the tv\ 
optically active isomers of carbonato-diethylenediamrninc 

Co-ordination Steieo-Chemistry 


cobaltic salicylate of class 4 (ii). These four isomeric com- 
pounds are shown in Diagram VIII. 

The whole of the foregoing cases of isomerism in com- 
pounds containing octahedral atoms relate to compounds 

lonrsdtrpn Isomers 

Type 4. C/ass f/t) 



of carbonate -complex 

containing only one octahedial co-oidination complex, but 
many cases have been discovered by Werner in which the 
isomerism is due to combinations between e severalloctahedral 

a i 

co-oidination complexes, and a few cases were elucidated 
by him in which the obseived isomeiism is due to combina- 
tions of octahedial with tetrahedral complexes. Strictly 

112 Chemistry and Atomic Structure 

all metallic ammines are combinations containing tetia- 
hedral complexes, for the nitrogen atom of each ammonia 
molecule is the focus of a tetrahedral co-ordination complex, 
which, however, owing to its symmetry, is not alone capable 
of giving rise to isomeric forms (see class I (ii) p. 97). 

Similarly the nitrogen atoms of ethylenediamme aie the 
foci of tetrahedral complexes, which, being of class I (iv), 
are aLo incapable of giving rise to isomeric foims. One 
of the carbon atoms of propylenediamine (tetrahedial 
class I (v) ), CH 3 .CH CH 2 .NH 2 , can give rise to optically 

NH 2 

active isomeric forms, and octahedral complexes containing 
propylenediamine consequently give rise to at least two 
optically active isomers, even if the octahedral complex 
itself has no possibilities of isomerism. Werner was success- 
ful in obtaining, for example, 1 the whole of the ten optically 
active isomers of dinitro-piopylenediammino-ethylenedi- 
ammino-cobaltic salts, five from one isomer of propylene- 
diamine and five from the other isomer. Reference to 
Tables 5 and 8 shows that the compounds belong to class 
3 (v) and that four of the five isomers should be derived from 
the cis and one from the trans octahedral complex. The 
four cis and one trans optically active isomers were actually 
obtained with each form of propylenediamine., thus demon- 
strating conclusively not only the octahedral structure of 
the cobalt complex but the tetrahedral structure of only 
one carbon complex. 

It is also of interest to observe that Werner at the same 
time obtained the complete set of partially and completely 
racemic compounds, i e. racemic with respect to propylene- 
diamine only, to cobalt only, and to both propylenediamine 
and cobalt. This series of optically active and racemic 
compounds must be considered one of the most outstanding 
of Werner's triumphs, and one of the most conclusive pi oofs 
for his co-ordination theory. 

Most of the cases of combination between octahedial 

1 Hdv Chim. Acta, 1918, 1, 5 

Co-otdination Steteo-Chennshy 113 

complexes i elate to those containing two cobalt atoms, the 
so-called " (j-bmucleai " complexes, in which one complex 
acts as a chelate group to the other by means of two atoms 
in the cis position. A typical example, due to Werner, 1 
is immo-nitro-tetra-ethylenediammino-dicobaltic complex, 
which contains the nnino and nitio gioups in the cis position 
in each cobalt complex together with two ethylenediamine 

(" en ") molecules, eri z Co/ ,j~ 2 \Coen 2 . Each cobalt 

complex is separately of the general class 4 (iv) and should 
give use to a paii of enantiomoiphs, which are the same pair 
for either cobalt atom. As in the case of a bi-nuclear 
tetrahedral complex, only three combinations are possible 
of two um-nuclear forms, the fiist combination consisting 
of two of the same enantiomorph, the second combination 
of two identical enantiomoiphs of the other sort, and the 
third combination of an enantiornoiph of each sort. The 
fiist two combinations are optical isomeis wheieas the third 
is optically inactive by internal compensation, i.e. a meso- 
form. These three isomers were all obtained, the inactive 
form being the fiist meso-compound to be obtained con- 
taining an " asymmetric " atom other than carbon. These 
three isomers are completely analogous to the three isomenc 
tartanc acids, and dirfei structuially only in containing two 
octahedial instead, of two tetrahedral co-ordinating atoms. 
In 1914, a long-known complex basic salt was shown by 
Werner 2 to contain four octahedral complexes, in which 
each of the three cis positions of a cobalt complex aie 
common to one of thiee other cobalt complexes, the three 
last acting as thiee symmctiicdl chelate gioups to a cential 
atom, the quadnnuclear complex thus repiesenting the 
two optically active foims of class 4(1), both of which weie 
obtained, the compounds being the enantiomorphic paii 

conespondmg to the foimula Co| ~^>Co(NH 3 ) 4 J 3 Br 6 . 

L ei , 1913,48, 3674 
2 Ibid , 1914, 47, 3087. 

114 Chemistry and Atomic Sttuctuic 

The elucidation of the constitution of this complex salt 
threw a flood of light on the constitution of basic salts 
gen ei ally and of many complex mineials, paiticularly those 
of the apatite group, basic copper cadmibiomide, for 

example, having the stiucture, ^df QTT/Cul a ^ 1 2' anc ^ 

apatite the stiuctuie ^(r\ ' pr) 3 /'' 33 / ? r 2 ' 

The phenomena to which Beizclms gave the name 
" isomensm " in 1832, but which had been noticed by 
Faraday in 1825, constitute the foundation for the whole 
theoiy of steieo-chemistiy, and to its manifestations can 
ultimately be assigned the whole of the existing knowledge 
icspectmg the constitution of molecules and the chemical 
stiuctuie of atoms. Weinei's co-ordination theory is 
essentially a theory relying foi its evidence on isomensm. 
Though Werner had shown in the early yeais of lut, theoiy 
that it was competent to explain the constitution and 
isomensm of hundieds of classes of chemical compounds, 
his theoiy was not veiy widely accepted until in 1911 he 
demonstrated the existence of optically active isomeis, of 
cobalt co-oidmation compounds 

This tardy recognition of the co-oidination theoiy has 
been due without doubt to the lather abnoimal impoitance 
attached to optical activity by many of the leadeis o 
modem chemical thought, who aie specialists in the com- 
paratively restncted field of oiganic chemist ly. The 
carbon atom being of a tetiahedial nature is icpresentativc 
of a type to which only a few of the lightest elements con- 
fonn, and Table 4, on page 97, shows that only foui of 
the fourteen possible types of tetiahedial complexes can 
give rise to isomers, and that the whole of the isomeis ate 
optically active Optically active compounds aie theiefoie, 
necessarily, the outstanding feature of isomensm in organic 
chemistiy. Even the optically inactive meso-isomers of 
organic chemistry are the icsult of two optically active foci 
in molecules, as has been shown on page 99 in connexion 

Co-ordination Steteo-Chemistry 115 

with meso-forms. The uninucleai stereo-chemistry of the 
01 ganic domain is thus largely the stereo-chemibtry of 
cnantiomorphic co-ordination complexes. This is no longer 
true in the domain of inorganic chemistry, for a single 
octahedral co-ordination complex can give use to isomeis 
which are optically inactive, and then existence necessitates 
modification of the conception of the importance of optical 
activity in theories of isomerism. To the general failure 
to realise this fundamental difference between tetrahedral 
and octahedral isomerism, is to be attributed much of the 
common failure to appreciate the decisive nature of Wernei's 
experimental evidence for his co-ordination theoiy 

Many complex compounds aie known in inoi ganic 
chemistry to which it appeals impossible to assign a co- 
ordination structuie of six or less atoms or groups about a 
cential atom, and it is usual to regard such compounds as 
complexes containing a nuclear atom with a higher co- 
ordination number than six. The commonest complexes 
are those to which the co-ordination number eight has been 
assigned to the cential atom, for example, the zncomum, 
cerium, and thorium atoms in their tetia-acetylacetones 
and related compounds. However atti active this analogical 
formulation may be it must be remaiked that no case 
is known of the existence of isomenc forms of com- 
plexes to which the co-ordination number eight may be 

Examination of the symmetry of the eight-point structuie 
of the cube shows that it is capable of giving rise to five 
mam types of complexes, non-chelate, mono-chelate, di- 
chelate, tn-chelate, and tetia-chelate, and that these five 
types include one hundred and nine classes, of which 
twenty- two aie non-chclate, twenty- two mono-chelate, 
twenty- five di-chelate, twenty tn-chelate, and twenty 
tetra-chelate. The tetra-acetylacetones should give rise 
to two optically inactive isomeis, which may be desciibed 
as syn and anti in orientation, while unsymmetncal dike- 
tones should give use to eleven isomers, six anti enantio- 

Chemistry and Atomic Structure 

moiphs, and one anti and tour syn inactive foims (see 
Diagram IX) In no case have isomeis been detected. 


Comp/exes wtt) Cub/c 
j Symrn 

A*r- 9 A 

Anb Form 
Unsymmch icel Chela ke G/oups 

pair Indnttomoribh/cpj'r Cnant/omortihic pair 
Qptrca/ty active Anbi-tscmcrs ' 

fnactwe An/-/somer 

A 9^- 

/nact/ve Syn-Isomers 
/eren Isomers of Propionyfocetones of puadnfafenb Meh/s 

The evidence foi cubic co-oidmalion complexes is at 
present nil, and though it cannot be stated that such com- 
plexes cannot exist, it is at least a cunous coincidence that 
all modem theories of atomic structuie involve that no 
atomic structure, elemental 01 ionic, has eight elections in 

Co-ordination Stereo-Chemistry 117 

or neai the atomic surface, and that six is the maximum 
number of subgroup elections that can appear neai the 
surface of an atom or an atomic ion. 

The evidence contiibuted by Werner's co-ordination 
theory to the theoiy of the stiucture of atoms may be 
summarised as follows. Atoms on combination gain or lose 
negative electiic charges in accordance with their electro- 
negative or electiopositive valency , the exterior of atoms 
is characterised by the possession of a definite number of 
symmetrically distributed points at which combination 
with other atoms occuis ; these characteristic atomic 
superficial features aie, in numbei, two for the hydiogen 
atom, four for the atoms from lithium to oxygen, four for 
most of the atoms of the periods of the peiiodic classifica- 
tion which include the halogen elements ; commonly six 
for most of the atoms of the periods which include the 
" transition elements " ; and four for most of the atoms in 
combination with oxygen as acids. 

The inteipretation of this evidence in terms of electrons 
will be given in latei chapters on the electronic structuie 
of atoms. It may be lemarked that the evidence of the 
co-ordination theory is almost the only experimental 
evidence available at the piesent day for the determination 
of the number of electrons in the various sub-gioups of 
atomic structure, just as the periodic classification is the 
sole expeiimental evidence for the deteimination of the 
number of electrons in the larger gioups of atomic structure 



CAVENDISH had shown as early as 1767 that the equivalency 
between an acid and a base on neutiahsation was not the 
equality of weight, and this fact is incoipoiated in 
Richter's law of equivalent pioportions, applying to com- 
binations between elementary substances as well as to com- 
pounds. The equivalent weight of a substance is the unit 
for the measurement of chemical operations. That the 
unit is a diffeient weight foi different substances is no 
hindrance in chermstiy to the use of the unit, any more 
than the difference in weight of commodities is any 
hindiance to commercial operations equated in cash. 
Had the equivalent weights of substances ultimately been 
clemonstiated to be identical with the weights of their 
ultimate paiticles, atoms or molecules, chemistiy would 
have been a science of extieme simplicity, a science of the 
stiucture of molecules instead of, as it is and always 
has been, a science of the structure of atoms and 
the i elation of the paits of atoms to the stiuctuie of 

The facts that the equivalent weight of an atom is not 
always identical with the atomic weight, and that the 
various equivalent weights of an atom aie integral multiples 
of the smallest equivalent weight, were summaiised by 
Dalton in 1803 in his law of simple multiple propoitions 
by weight. Though the precise values of atomic weights 
weie not a matter of general agieement until aftei 1858, 
when Canmzzaro had secuiely established Avogadro's 
hypothesis, no doubt existed from the eailiest days of 
Dalton's atomic theory that most atoms enteied into 
chemical combinations in virtue of their possession of moie 
than one equivalent weight. This combining capacity of 
an atom, called by Gerhaidt in 1853 its " hydiogen 
atomicity," was identified, immediately aftei Cannizzaro's 
regularisation of atomic weights, as a chaiacteristic of the 
atom no less definite and impoitant than its weight, and, 

Valency and Sub-Atomic Cbemistiy 119 

under the name of valency, lias been the foundation stone 
of the whole structure of modern chemistiy. 

Faraday's work on electiolysis in 1833-1834 disclosed that 
the equivalent weights of all elements are associated with 
the same quantity or chaige of electricity, negative for some 
elements and positive for others, and led Cleik Maxwell in 
1873 to infer that each molecule liberated in electrolysis 
paited with the natural unit of electric charge. Clerk 
Maxwell's suggestion was made moie precise by Johnstone 
Stoney in the following year by the proposition of the 
natural unit of electrical charge as the quantity associated 
with the equivalent weight of an element or the rupture of 
one chemical bond. This unit, at first called the " elec- 
trine " but altered in 1891 to the now familiar " electron," 
is strictly an electrical unit derived from chemical con- 
siderations, and remained an electio-chemical hypothesis 
not admitted geneially for twenty-three yeais, until its 
identification by Sir J J Thomson in 1897 as the negative 
electucal charge on a material unit about eighteen hundred 
times lighter than the hydiogen atom. Since that date 
the name electron has been definitely transferred from the 
negative electrical charge to the material particle cairymg 
the negative charge. 

In 1 88 1, Helmholtz suggested the identification of units 
of chemical affinity, valency, with units of electrical charge 
or " atoms " of electricity, each chemical atom being 
assigned as many " atoms " ot electricity as it possessed 
units of affinity or valency. Clausius had aheady, in 1857, 
suggested that molecules were lo a small extent ionised m 
solution into parts cairymg positive or negative chaiges, 
and it consequently followed irom Helmholtz' suggestion 
that atomic ions in solution carried a number of positive 
or negative " atoms " of electricity in proportion to the 
valency existing in the combination. 

In 1901, Neinst clanfied Helmholtz' views by the 
suggestion that the laws of constant composition and of 
simple multiple proportions applied with equal foice to 

izo Cheimshy find Atomic Structure 

the quantities of electncity as to the quantities of matter 
taking pait in a chemical leaction, and deduced fiom the 
fact that elements are electrically neutral in the fiee state, 
that the ci cation of charges on atoms in combination must 
arise by the exchange between atoms of positive for negative 
" atoms " of electricity or elections, common salt, NaCl, 
for example, being a compound in which the sodium and 
chloiine atoms on becoming ions have exchanged negative 
and positive electrons. 

In 1902, Sir Oliver Lodge, in an addiess to the Institute 
of Electrical Engineers, 1 suggested that only negative 
electrons were transfened in chemical combinations, 
negative charges on atoms representing the acquirement of 
elections, and positive chaiges the loss of electrons, 
and his suggestion is the basis of all present theories 
of electronic reactions in chemistry. He further suggested 
that electrons in one atom could be utilised by another 
atom prior to their separation to foim appropriately 
charged ions, and this may be regaided as the ongin 
of the many subsequent theones of election-sharing in 
non-ionised compounds. 2 Two yeais latei 3 he suggested 
that, though the e]ectric chaiges 01 electrons, to the transfei 
of which chemical combination between atoms is due, aie 
indivisible, the forces resulting aie not indivisible, and that, 
though the bulk of the lines of force might be utilised in 
binding the atoms together, some of the lines of force could 
be utilised in binding other molecules, thus giving lise to 
complex aggregates, and ofrenng a feasible mechanism foi 
the explanation of residual affinity. Some time latei, 
P. F. Franldand 4 showed that Lodge's suggestions could be 
applied to the phenomena of ionic dissociation and cata- 
lysis, and to the combination of water of crystallisation 
and to molecular compounds generally. 

1 y lust Elect Eng , 1902-3,82,45 

~ See als>o his Modern Views on Matter, Rormnes Lectuie, CKford, 1903, and 
Elect rons, 1906 

3 Nature, 1904, 70, 176. 

4 Ibid , 1904, 70, 221 

Valency and Sub-Atomic Chemistry 121 

In 1908, Ramsay put forwaid the suggestion that elec- 
tricity is one of the chemical elements, the atoms of which 
aie the elections, on the grounds that the electrons are 
" atomic " in nature, have definite mass, exist free as mole- 
cules, and foim compounds by combination with ordinary 
atoms, and crystallised Lodge's electron-sharing views in 
the statement that elections " serve as the bonds of union " 
between atom and atom " * 

In the same year, Staik elaborated the conception of 
electron-sharing between atoms into a theory of chemical 
combination, 2 in which he icgarded an electron in an atom 
as sending lines of foice not only to its own positive part 
but also to the positive part ol the chemically bound atom. 
This theoiy was considerably altered in 1915 3 in its 
application to organic compounds, Stark regarding the 
shared elections as in two atomic structures simultaneously, 
one such electron being equal to one chemical bond, except 
in the combination between caibon atoms and between 
carbon and hydrogen when two shaied electrons were 
considered equivalent to one chemical bond. 

In 1913, Bohi, in extending his dynamic theory of 
atomic structure to the combination between atoms, 4 
suggested that the shared elections in a bond between 
atoms existed as a rotating ring of electrons perpendicular 
to the line joining the atoms, two electrons in the ring 
being equivalent to the bond between two hydrogen atoms 
or between a hydiogen atom and a carbon atom, and similar 
dynamic views were put forward by Kossel in I9i6. 5 In 
Parson's magneton theory of atomic structure, 6 two elec- 
trons were regarded as foiming one chemical bond between 
two like atoms, the two shared elections contributing to 
the formation of a gioup of two electrons m the atomic 

1 J Chem Soc., 1908, 93, 778 

" jfabrb Radioaktiv Llectionik, 1908,6, 125 

3 Die Elektnzitdt im chemischen Atom, Leipyig, 1915 

iJfV.1/ Mag, 1913, [6], 26, 857 

5 Ann P/JJ.M , 1916, [4], 49, 229 

6 SmithsuHiaa lust Publ , 1911;, 65, No n 

122 Chemist) y and Atomic Structure 

stiuctuie of hydrogen and a gioup of eight elections in 
the structme of othei atoms. 

In 1916, Lewis put forward a theoiy of valency based on 
static electrons otherwise almost identical with those of 
Parson and Kossel, 1 in which the outer or valency elections 
of atoms were shared between two atoms so as to complete 
outer groups of eight electrons in each atom, two electrons 
being regarded as equivalent to one chemical bond. 

The outstanding feature of nearly all these electron 
valency theories was the postulate of an outei completed 
group of eight electrons, after combination, in each atom, 
except hydrogen, the electron group in this case consisting 
of two electrons. The presumed necessity foi this lay in 
the fact, recognised by Mendeleefl in 1871, that the maxi- 
mum valency of any element is eight, illustrated by the 
tetroxides of ruthenium and osmium and the oclafluoiide 
of osmium. Reference to MendeleefPs 1871 table (on 
page 71, Chap. VI) shows that this maximum valency of 
eight is exhibited only by elements which do not yield 
compounds with hydrogen, and that the highest valency 
of other elements is equal to the difference between eight 
and the valency towards hydrogen, or, in other words, that 
the sum of the electropositive and electronegative valencies 
of elements is equal to eight. This fact, recognised by 
Abegg and Bodlander in 1899 z in their suggestion that 
the sum of normal and contra valency is eight, is of impor- 
tance in theories of atomic structure, but it is not justifiable 
to conclude that the whole of the eight electrons discern- 
ible in valency .combinations appear in the outer structure 
of any one atom, with the possible exception of osmium in 
the octafluoride. In this connexion it is significant that, with 
the solitary exception noted, all deductions, as to the 
presence of eight electrons in the outer paits of atoms, are 
made from compounds containing multivalent elements, 
particulaily oxygen, and there is good leason to believe 

1 Proc Nat Acad Set , 1916, 2, 586 
z Zeit anorg Chem , 1899, 20, 453 

Valency and Sub- Atomic Chemistry 123 

that in most of such compounds each multivalent element 
acquires outright at least one of the electrons, thus halving 
the numbei of elections. If it were universally true 
that two shared electrons are equivalent to each chemical 
bond, it would be necessaiy to assume that the eight 
chemical bonds m osmium octafhionde and tetroxide 
repiesented sixteen elections in the outer structure of 
osmium, an assumption definitely at variance with the 
eight-election postulate on which the two-electron bond 
claims to be founded. The same discrepancy obviously 
anses for every element having valency greater than four. 
It is permissible, therefore, to conclude, if the maximum 
number of electrons in the outer part of an atom be eight, 
that the chemical bonds of all elements having valency 
gieatei than four consist of not more than one electron 
per bond. If this must be conceded, it applies to neaily 
half of the known elements, and renders more than doubtful 
the general existence of two-electron bonds. 

Examination of chemical compounds geneially leads to 
the conviction that no simple rule exists foi the determina- 
tion of the number of electrons in a chemical bond, and 
that this number is a problem for experimental lesearch on 
eveiy bond in eveiy chemical compound The evidence 
gathered fiom compounds of valency greater than four 
appears to point to the existence of one shared electron per 
bond, and, as all these elements give rise to compounds 
more 01 less readily hydrolysable, it may be surmised that 
easily hydiolysable bonds contain only one electron what- 
ever the valency of the atom concerned. The only ele- 
ments that give lise to compounds unhydrolysable, or 
hydrolysable only with great difficulty, are the feebly 
electi negative non-metallic elements, such as carbon, 
nitrogen, phosphorus, and sulphur, and these together with 
oxygen, the element to which hydrolytic reactions are 
attributable, may be regaided as the elements in which 
two shaied elections may be expected to exist in a chemical 

124 Chemistry and Atomic Structure 

Theoiies based on the postulate of two electrons per 
chemical bond have multiplied to an extraordinary extent 
since 1916, and it is almost impossible to give an account 
of the various uses to which the hypothesis has been put 
by numerous wiiteis and workeis in the fields of both 
organic and inorganic chemistry as well as in physics The 
most outstanding of the theoiies based on the two-electron 
junction is that of Langmuir, 1 and is merely, so far as valency 
is concerned, an amplification of Lewis's views. This 
theory postulates that the two-electron bond or " co- 
valency " is the invariable bond of chemical unit valency, 
and that the outer surface of every atom in combination, 
except hydrogen, consists of an invariable " octet " of elec- 
trons ananged with cubic symmetiy. As has already been 
indicated this type of theory, based on two-electron junc- 
tions per bond and eight elections per atomic gioup, com- 
pletely fails in application to the whole of the elements 
having valency greater than four, and moreover to the 
whole of the elements which Werner has proved to be 
associated with six co-ordinated groups 01 atoms. This 
restricts the application of Langmuii's theory almost 
exclusively to hydrogen and the seven elements fiom 
lithium to fluorine, and, though the theoiy is claimed to be 
applicable to elements generally, it is noteworthy that 
nearly all the examples cited in the theoiy aie compounds 
of carbon, nitiogen, and oxygen. It may further be 
remarked that no theory founded on a bond of two elec- 
trons can explain the existence of iii-atomic hydrogen, the 
ions of tri- and di- atomic hydrogen, the boron hydrides, 
nitric oxide, caibon compounds with triple bonds, and the 
non-existence of the tetia-, penta, and hexa- fluoioethaneb, 
compounds of six of the eight elements to which Lang- 
muir's theory must be expected to be in complete accord. 
In addition to its bearing on valency, Langmuir's 
theoiy presents an inteipretation of the periodic 
classification based on a scheme of atomic structiue. 

1 y Amer Cbem Soc , 1919, 41, 868, 1543 ; 1920, 42, 274. 

Valency and Sub-atomic Chemistry 125 

This poition of his theoiy will be dealt with in 
Chaptei XIII. 

Lewis has recently extended his theory of 1916, on which 
Langmuii's theoiy is founded, and, while retaining the 
co valency postulate of the two-election bond, has rejected 
Langmuir's theory in so far as it postulates static electrons 
and an invaiiable atomic outer stiucture of eight electrons. 1 
"Lewis, howevei, has been no moie successful than Lang- 
muir in his explanation of many of the compounds to which 
Langmuir's theory is inapplicable, and his new theory 
necessitates so gieat elasticity in the outer groups of atomic 
structure that one is diiven to conclude that this elasticity 
is necessitated not by expenmental facts, but solely by his 
hypothesis of the invariable two-electron bond, and repre- 
sents the last lesouice of a dying theoiy The carbon 
compounds alone aie so numerous and so vaiied in types 
and propei ties that it is inconceivable that any adequate 
explanation of them could be furnished by a theory 
involving merely paiis of equivalent electrons in the caibon 
bonds with other atoms. A sufficient body of evidence 
has now accumulated to indicate that the reactive electrons 
in carbon atoms in combination consist of at least two and 
probably three diffeient types, and that minute changes 
in the structure of compounds suffice to convert electrons 
of one type into elections of another type. It appeals 
certain that theories of chemical valency must begin at a 
lower level than the suiface of an atom, that the solution of 
the problem of valency must follow not piecede the solution 
of the more geneial problem of atomic stiuctuie, and that, 
theietoie, the problem of the numbei of elections in any 
chemical bond will be the last solved of the pioblems of 

and. tbc ^tincture. <>j Atom* and Mjkcules, Ntn, York, 1923 



IN 1896, in the course of a search for substances spon- 
taneously emitting Rontgen or X-rays, Becquerel disco veied 
that the fluorescent double sulphate of uranium and potas- 
sium was capable of affecting a photogiaphic plate even in 
complete darkness, 1 and later pioved that this propeity was 
independent of the uranium salt used and of its previous 
exposure to daylight, thus demonstrating that the radiation 
pioduced was an intrinsic property of the element uranium. 
Two years later, Schmidt pioved that the element thoiium 
had radiative properties similar to those of uianium 2 

Becquerel' s further discoveiy in 1896 that uranium 
radiation had the property of discharging electufied bodies, 
that is, of making the surrounding air a conductoi, rendered 
possible the accurate measuiement of the intensity of even 
extremely feeble radiation, and his gold-leaf electroscope 
method is still in common use for the deteimmation of 
ladioactivity. The ease and rapidity of this method of 
measurement of radioactivity led Madame Maiya Cuiie to 
the examination of a large number of minerals, 3 and, in 
conjunction with her husband. Fieri e Cuiie, to the dis- 
covery of polonium, an intensely radioactive substance 
having properties allied to those of the element tellunum. 4 
In the same year, they discovered the now well-known 
element radium, 5 which was proved to have properties 
very similai to those of banum. It is, however, piobable 
that radium is moie closely allied to stiontium (both give 
intensely red flame colorations), just as thorium is most 
closely allied to ziiconium, and uianmm most closely allied 
to molybdenum. It is, in fact, a geneial obseivation that 
alternate members of a valency group in the peiiodic table 
show the gieatest chemical resemblance, foi example, 

1 Compt. rend , 1896, 122, 420. 

2 Ibid , 1898,127, 1264 

3 Doctonal Thesis, Faculte des Sciences, Pans, 1898 
1 Compt rend, 1898, 126, 175. 

Ibid, 1898, 127, 1215 

Radioactivity and Sub-Atomic Chemistry 127 

iodine and chloiine ; biomine and fluoiine , bismuth, 
aisenic and nitrogen ; and antimony and phosphorus. 
The properties of the elements of the last period appear 
to indicate that this peiiod resembles the 2nd long period 
lather than the 3rd, and it is piobable, if the whole of the 
elements concen able existed, that there would be eighteen 
membeis in the last period as in the 2nd long period, rather 
than about thnty membeis as m the 31 d long period 

The atomic weight of ladium was found by Madame 
Cune in 1903, by the analysis of ladium chloride to be 
approximately 225, and in 1907 she obtained the more 
accurate value, 226-2, by the analysis of a lelatively large 
quantity (o 4 giam) of pure radium biomide. 1 The most 
lecent detcimmations by other woikeis indicate that the 
figure, 2260, icpresents the atomic weight of i admin with 
gieat accuiacy. Radium is the only element of the thirty- 
nine ladioactive elements other than uranium and thorium, 
for which the atomic weight has been experimentally 
determined It was isolated in the free state in 1910 by 
Madame Curie and Debierne, and found to resemble 
metallic baiium. The fiee metal was found to have pie- 
cibcly the same ladioactivity as the same weight of the 
metal in the foim of its salts, thus demonstrating con- 
clusively that radioactivity is an intnnsic propeity of an 
atom and independent of the operation of valency. 

In 1899, Debierne discovered a new radioactive substance, 
actinium, allied in piopeities to the trivalent rare-earth 
elements. He was unable to sepaiate it from admixed 
lanthanum owing to the minuteness of the quantity obtain- 
able. 2 In the same year the Cunes 3 showed that ladium 
had the property of communicating radioactivity to bodies 
in its vicinity, and Rutheiford independently showed that 
tlionum had the same property. 4 In 1904, Debiemc 
showed that actinium icsembled radium and thorium in 

1 Compt lend , 1907, 145, 422 

z Ibid , 1899, 129, 593 , 1900, 130, 206. 

3 Ibid, 1899, 129, 714 

1 PA//. Mag, 1900, [6], 49, i 

128 Chemistry and Atomic Stntctuie 

its capacity to make surrounding objects radioactive 1 
Rutherford's expeiiments indicated that this piopeity of 
thorium was due to the liberation of a gas or emanation 
(thoron) with radioactive properties, and Dorn in 1900 
pioved that radium hbeiated a similar emanation. 2 

In 1902, Rutherford and Soddy 3 came to the conclusion 
that the ladium emanation had the propeities of an ineit 
gas of the same family as helium and aigon, and suggested 
that the helium always found in ladioactive minerals might 
be causally not casually connected theiewith Mai tin in 
the same year suggested that the iddioactive elements wcic 
undergoing decomposition, 4 and Rutherford and Soddy 5 
shortly afterwards put forward a hypothesis of radioactive 
disintegration and tiansfoimation, by which it was assumed 
that a definite proportion of the atoms of a ladioactive 
element are unstable and disintegrate by the emission of 
one or moie of the thiee known types of radiation associated 
with radioactive elements, the residues of the ongimil 
disintegrated atoms being atoms of new elements, of which 
another but definite piopoition are unstable and fuitlier 

The three types of radiation fiom radioactive substances 
are a-rays, consisting of heavy positively charged pai tides 
not materially deflected by a magnet ; p-iays, consisting of 
very light negatively charged particles (electrons) easily 
deflected by a magnet ; and y-iays not aftected by magnetic 
or electric fields and identical with X-iays, except in being 
usually more penetrative, thus being in fact oidmaiy light 
pulses of the shortest known wave-length. 

In 1903, Ramsay and Soddy pioved conclusively that 
radium emanation disintegrates by the foimation of the 
inert gas helium, and that therefore the <y-iays, detected by 

1 Compt rend , 1904, 138, 411 

3 Abb. Naturforscb Ges Halle, 1900 

3 Phil Mag, 1902, [6], 4, 580 

4 Chetn News, 1902, 85, 205 

8 Pbil Mag, 1902, [6], 5, 576 

u Proc Roy Soc , 1903, 72, 204 , 1904, 16, 346 

Radioactivity and Sub-Atomic Chemistry 129 

Rutherford and Soddy * as the only radiation from the 
emanation, consisted of helium. This may be regarded as 
the first direct experimental proof that atomic disintegra- 
tion and transformation are realities. In 1909, Rutherford 
and Royds proved by spectroscopic methods a that a -rays 
in all cases consisted of positively charged helium atoms, 
thus demonstrating the reality of disintegration and trans- 
formation for the whole of the radioactive elements giving 
rise to a-radiation. 

With the exceptions of radium and helium it was not 
certain that any of the radio-products were in fact chemical 
elements, until Ramsay and his co-workers had determined 
the chief physical constants of radium emanation, the density 
of which was found by Whytlaw-Gray and Ramsay to be 
111-5, it thus having the highest density and the highest 
molecular weight of any known gas. 3 It having been con- 
firmed that radium emanation is completely destitute of 
chemical properties, i.e. valency, it was placed by analogy 
in the same periodic group as the ineit gases. If, like them, 
it is a monatomic gas, its atomic weight must be double its 
density, indicating the atomic weight, 223, from Gray and 
Ramsay's density measurement, a value confirmed by 
Perkin's (1908) and Debierne's (1910) experiments on the 
rate of effusion of the gas. To mark his identification of 
radium emanation as a definite chemical element, Ramsay 
proposed for it the name, niton, though it is still often 
referred to as radium emanation or radon. 

In 1905, Debierne * showed that actinium, like radium, 
yields helium in disintegrating to form actinium emanation 
or actinon, discovered by him in 1903. In the same year, 
Soddy and Mackenzie 5 proved that radium is pioduced by 
the disintegration of uranium. Two years later, Boltwood 
showed that this disintegration involves the production of 

Mag, 1903, [6], 5, 445. 
z Ibid, [6], 17, 281 

3 Proc Roy Soc , 1911, A, 84, 536 

4 Compt rend, 1905, 141, 383. 

B Phil Mag 3 1907, [6], 14, 272 

130 Chemistiy and Atomic Stmclwe 

an intermediate element to which he gave later the name 
ionium, he having at first assumed that this element was 
actinium. 1 Boltwood further showed that ionium was 
chemically inseparable from thorium, and differed from it 
only in possessing different radioactive properties. In 
1907, McCoy and Ross 2 showedthat thorium and radio- 
thorium, one of its disintegration pioducts, were chemically 
inseparable, and differed only in ladioactive pioperties. 
Three years later Maickwald 3 showed that radium and 
mesothoiiunij were similarly chemically inseparable, yet 
differed in radioactive properties. 

In 1911, Soddy* independently came to the same con- 
clusion as Marckwald as to the non-identity but chemical 
inseparability of the two bodies radium and mesothormnij, 
and proposed the term isotopes for all elements chemically 
inseparable, i.e. having identical chemical properties and 
valency, but differing usually in atomic weight and always 
in radioactive pioperties. In 1913, Fleck 5 pointed out 
that all the then-known ladioactive elements, over twenty 
in number, corresponded with only ten positions in the 
periodic classification, and that five of these positions were 
occupied by the five elements, thallium, lead, bismuth, 
thorium, and uianium, all of which were known before 
radioactivity was discovered. It was, consequently, evident 
that many of the radio-elements were not only isotopes of 
existing elements but of one another, and that the dis- 
integration products of an element weie often its 

Many suggestions had been made piior to 1913 as to the 
rule or law underlying radioactive disintegrations and 
transformations, for it had become abundantly evident that 
particular transformations were associated with specific 
types of radiations and were accompanied by expulsion of 

J Sci., 1907, [4], 23, 93, 1908, [4], 25, 365 

2 y Amcr Chew Soc , 1907, 29, 1709 

3 Ber , 1910, 40, 3429 

4 J Ch<.w Soc, 1911, 99, 72 

B Ibid I9f3 103, 381 and 1053 

Radioactivity and Bub- Atomic Chemistry 131 

positive helium atoms or a-particles, and negative or 
p -particles or electrons, and emission of y-radiation or 
X-rays. A. S. Russell, 1 early in 1913, formulated a scheme 
for the particular case of the transfoimation of uranium to 
ionium, by which the valency decreased two on the loss of 
an a-particle and increased one on the loss of a p-particle, 
uranium of valency 6 thus passing into uranium Xj of 
valency 4 by the loss of an a-particle, this passing into 
uranium X 2 of valency 5 by loss of a p-paiticle, then into 
uramum 2 of valency 6 by loss of a second p -particle, and 
then into ionium of valency 4 by loss of a second a-paiticle. 
A similar scheme was put forward by Fajans. 2 In the 
same year Soddy formulated the complete law of radio- 
active change, known as the Law of Radioactive Group 
Displacement, 3 which states that an element, on expulsion 
of a doubly-charged helium atom or a-particle, passes into 
an element two less in valency, and, on expulsion of a 
p-particle or negative electron, passes into an element one 
greater in valency. 

Certain well-defined exceptions to the displacement law 
must be recognised. Bivalent radium emits both a- and 
p-particles but passes into non-valent niton as if only an 
a-paiticle had been emitted. Quadnvalent radioactimum 
similarly emits both sorts of particles, and y-rays as well, 
but passes into bivalent actinium X as if only an a-particle 
had been emitted. Quadiivalent radiothorium similarly 
emits both sorts of particles but passes into bivalent 
thorium X as if only an a-particle had been emitted. 
Neither bivalent mesothormni! nor trivalent actinium emit 
p -particles in changing to trivalent mesothorium 2 and 
quadiivalent radioactmium respectively. At the points of 
branching of both the radium series and the actinium senes, 
where both a- and p-pai tides should be emitted, only one 
type of particle is emitted, quinquevalent radium C passing 

1 Chem News, 1913, 107, 49 

2 Pbys Zeit , 1913, 14, 131 and 136, Eer , 1913,46, 4*2 

8 Cbem News, 1913, 107, 97, Jabrb Radtoakltvat, 1913, 10, 188 

132 Chemistry and Atomic Structure 

into trivalent radium C z without emitting a-particles, and 
quinquevalent actinium C passing into sexavalent actinium 
Q without emitting p -particles. 

The substantial accuracy of the law of group displace- 
ment may, however, be accepted, as the foregoing dis- 
crepancies are largely negative in character and the appro- 
priate radiation may yet be discovered. The positive 
exceptions consist in the emission of p -particles or electrons, 
and this emission has been attributed to the ionisation of 
the atoms by the loss of an outer electron due to some 
peculiarity in the method of a-particle emission. 

The law of radioactive group displacement involves that 
unit change in electric charge is accompanied by unit 
change in valency, the loss of the doubly-chaiged helium or 
a-particle accompanying decrease of two in valency, and the 
loss of one negative election 01 p-particle accompanying 
inaease of one in valency. It is evident that atoms must 
consist of a large number of positive helium particles and 
negative electrons, for, in the change from uranium to 
lead, eight a-particles and seven (3 -particles are lost in the 
fourteen valency transformations. It is equally evident 
that the electrons lost cannot be the valency electrons of 
uranium, which has six valency elections, for, not only are 
seven electrons in all lost, but the transformations result 
four times in sexavalent elements with six valency electrons. 
In fact after thirteen of these tiansformations involving the 
loss of seven electrons, the resulting element, polonium, has 
six valency electrons, being sexavalent like uranium. 

As the loss of a helium particle from an atom must be 
accompanied by the diminution of atomic weight by 4, the 
atomic weight of helium, and can be compensated chemi- 
cally by loss of two p-pai tides, it is evident that atomic 
weight is not chemically characteristic of atoms. This is 
illustrated by the fact that quadiivalent thorium after the 
loss of an a-particle changes to the element' mesothoriumj 
of atomic weight less by 4, which further loses two succes- 
sive electrons, without change in atomic weight, changing 

Radioactivity and Sub- Atomic Chemistry 133 

successively into trivalent mesothorium 2 and quadrivalent 
radiothoiium, the last an element chemically indistinguish- 
able from the original thoiium. Not only may chemically 
identical elements have the same or different atomic weight 
(isotopes), but chemically different elements may thus have 
identical atomic weights, e.g. mesothorium l5 mesothorium 2 , 
and ladiothorium. Different elements with identical 
atomic weight aie described as isobars, and numerous sets 
of triplet and doublet isobaric elements are known. 

The following table includes all the known radioactive 
elements the mutual relationships of which have been 
determined, valency (in roman numerals) and atomic weight, 
known or assumed (in figures), being shown in brackets. 


Uranium Series 
U (vi,238a) 


jfiv, 234) 

UX,(v, 234) 

P " 

Actinium Series 

Thorium Series 


(iv, 232-15) 



(v, 230) 


(u, 228) 




(111, 226) 

M Th, 

(ui, 228) 

Io (iv, 230) 

Ra (u, 226) 

Nt (0,222) 

RaA(vi, 218) 

RaB(iv, 214) 

RaC(v, 214) 


RaD(iv, 210) 

P (Actinium-lead) 

RaE (v, 210) 

Po (vi, 210) 


RaG (iv, 206) 

RdAc (iv, 226) 


(iv, 228) 



AcX (n, 222) 


(n, 224) 



An (0,218) 


(o, 220) 



Ac A (vi, 214) 


(vi, 2 1 6) 



AcB (iv, 210) 





AcC (v, 210) 


(V, 212) 

1 N 
71, 2 to) AcC, (ill, 206) T 



a P 



AcD (iv, 206) 


(iV, 20S) 


134 Chemistry and Atomic Structure 

In addition to the foregoing thirty-nine elements, two 
other radioactive elements, uranium Y and Z, have been 
detected. Both emit p-iays, but in neither case are their 
immediate generators known, though both are deiived from 
uranium. Uranium Y is usually regarded as immediately 
derived from either uranium or uranium 2 by an a-ray 
emission, and as being the progenitor of protactinium by a 
p-ray emission, thus forming the link between the uranium 
and actinium senes, the latter being a blanch seiies of the 
former. The main objection to this assumption appears 
to be that either uianium or uranium 2 must emit two 
different sorts of a-rays, and such should be distinguished 
by difference in range. The Geiger and Nuttall Law, 
however, indicates that the lange of the a-particles is a 
function of the " period of half -life " of the emitting 
element. As the " period of half-life " is a radioactive 
constant foi each element, difference in range is evidence 
of the existence of different elements. It may be inferred, 
therefore, that neithei uranium nor uranium 2 can be the 
immediate progenitor of either uianium Y or uianium Z. 

All the three known cases of blanching of a series are 
due to the emission of two different types of rays, the 
disintegiation following either an a-iay or p-ray emission. 
It would therefore appear by analogy that uranium X and 
uranium Y are not pioduced from an element giving rise 
to already known a-ray emission. If, however, an element 
can disintegrate in two different ways, emitting in both 
cases similar rays of the same lange, there appears to be no 
objection to either of these elements being the immediate 
pioduct of uranium or uramum 2 . 

One of the curious features of the three series of radio- 
active elements is that in all thiee the course of transforma- 
tion is identical for seven successive transformations, and 
that these seven transformations include similar types of 
branching of each seiies at conesponding points. Ionium, 
radioactinmm, and radiothonum, each after two a-ray 
emissions yield an emanation which is an inert gas, niton, 

Radioactivity and Sub-Atomic Chemistry 135 

actinon, and thoion respectively. Each of these, after two 
successive a-ray and one (3-ray emission yield an element, C, 
emitting both a and p rays, so that two products are 
obtained, Cj and C 2 , yielding by single different emissions 
the same product. Radium C thus yields both radium C^ 
and C 2 , both of which pass to radium D ; actinium C 
yields both actinium C x and C 2 , both of which pass 
to actinium D ; and thorium C yields both thorium Cj 
and C 2 , both of which pass to thorium D. These three 
series of seven transformations thus give isotopic elements 
at each stage, the final stage yielding the three D isotopes 
of lead In the uranium series the dismtegiation passes 
through thiee moie stages yielding another isotope of lead, 
radium G or radium-lead. It seems probable that the 
two D isotopic products of actinium and thorium aie not 
end members of their series, but pass, by analogy with 
radium D, through three conespondmg stages to foim two 
elements also G isotopes of lead. 

On page 132 it was indicated, in the fourteen transforma- 
tions from sexavalent uranium to quadrivalent radium G 
(radium-lead), that seven electrons aie expelled as p-rays. 
This, however, cannot represent the whole of the elections 
lost, because the valency changes are not taken into account. 
In an a-ray change the valency diminishes by two, and two 
electrons must be expelled fiom the valency group of the 
atom. In a (3-ray change the valency inci eases by one, and 
one electron must be gained by the valency group of the 
atom If the expelled election leaves the neutral atom 
completely, an electron must be captured by the valency 
group of the now positively charged atom to neutralise it 
and increase its valency by one If the expelled electron 
does not leave the atom, it must remain in the valency 
group of the neutral atom as a valency electron. In either 
event the p-ray change does not result in an atomic net 
loss or gain of electrons. It consequently follows, in the 
eight a-ray and seven p-ray changes from uranium to 
uranium-lead, that eight sets of two or sixteen electrons 

136 Che mistt y and Atomic Structure 

must be lost completely in the radioactive processes. As 
lead is quadrivalent and thus possesses at least foui valency- 
electrons, it is evident that the original uranium atom must 
have contained sixteen plus four or twenty electrons, and 
as the element has a valency of six, it follows that the 
radioactive processes disclose in uranium fourteen electrons 
other than those in the valency group. 

Fuither, since the radioactive changes from uranium to 
niton result in the reduction of the valency from six to 
zero, and as the removal of two electrons from niton results 
in radium A with six valency electrons, it must be inferred 
that niton has a group structure of eight electrons which is 
readily degraded to a group of valency electrons, varying 
from six for the isotopes of polonium to three for the 
isotope of thallium. It may likewise be inferred that the 
niton group of eight electrons is present in the isotopes of 
the elements uranium, protactinium, thorium, actinium, 
and radium, in addition to a group of electrons characteris- 
ing their valency. Not only must an eight electron group 
be present in all the isotopes of these five elements, but the 
absolute identity m the properties of the three inert gases, 
niton, actinon, and thoron, definitely shows that the 
arrangement of these eight electrons is identical in these 
three gases, and also, therefore, in the isotopic elements 
from which they are derived by successive removals of 
valency elections. 

The foregoing atomic structures for the thirty-nine 
radioactive elements have been deduced solely from experi- 
mental evidence, and are independent of any theory or 
hypothesis whatever as to atomic structure, and it must 
be regarded as certain that the whole of these elements 
have atomic structures of similar type, i e. that the struc- 
tures of the atoms of smaller atomic weight repiesent stages 
in the structural arrangements of the atoms of higher 
atomic weight, even when the higher atomic weight atoms 
are not, those that yield the lower by disintegration. 

In view of the remarkable similarity in the properties of 

Radioactivity and Sub-Atomic Chemistry 137 

the elements, radioactive and non-radioactive, isotopic 
and non-isotopic, of the same periodic group, as chemical 
ions or free atoms it may be inferred that the whole 
of the known elements consist of atoms constituted on 
the uniform plan to which the radioactive elements con- 
form. This plan in the case of radioactive elements has 
been shown to consist of groups of eight electrons in the 
inert gases and the elements of atomic weight up to 
uranium, and of the same group, though incomplete, in the 
case of radioactive elements of atomic weight less than the 
radioactive inert gases. By analogy the atoms of the non- 
radioactive inert gases also con Lain a group of eight elec- 
trons, and the same applies to the elements immediately 
following each ineit gas. This is borne out by the fact 
that the elements preceding and following each inert gas 
have valency increasing by one unit from two to six in all 
cases, and from one to seven in one case. This includes 
more than half the known elements. The elements to 
which the analogy cannot be applied are those series in 
which more than seven elements occur without an inter- 
vening inert gas. These are the elements of the middle of 
the Mendeleeffian long periods of the periodic table, known 
as the transition groups or series Except that these 
elements must contain the eight-electron groups of inert 
gases of lower atomic weight, the electronic structures of 
these elements cannot be determined by reference to the 
structures of the radioactive atoms, as these do not include 
a sufficient number of members of a transition series. 

As the element helium is not preceded by a group of 
multivalent elements, it is not permissible to assume, on 
the radio-element analogy, that it contains a group of 
eight electrons. The facts that hydrogen with only uni- 
valency is the only element preceding helium, and that 
helium is capable of yielding the doubly-charged a-particle 
by loss of two electrons, involve that helium, unlike all the 
other inert gases, possesses a group of only two electrons, 
and that successive electrons are added m subsequent 

138 Chemistry and Atomic Structure 

elements until the next inert gas is reached after the addi- 
tion of eight electrons. The structures of the known inert 
gases can therefore be determined, at least in part, helium 
consisting of one group of 2 electrons ; neon of two groups, 
2 and 8 ; argon of three groups, 2, 8, and 8 ; krypton of 
fom groups, 2, 8, 8, and 8 ; xenon of five groups, 2, 8, 8, 8, 
and 8 ; and niton of six groups, 2, 8, 8, 8, 8, and 8. 

The law of radioactive group displacement indicates 
that every unit increase m valency is accompanied by unit 
loss of negative electiic charge, or, in other words, that in 
passing from one element to the element next higher in 
periodic group number, the number of valency electrons 
increases by one while the total number of electrons in the 
atom is actually unchanged. Extending the law for 
valency increases to non-radioactive elements, involves that 
in passing from one valency gioup to the next higher, the 
total number of electrons is increased by one, foi no 
electron is lost by p-iay emission, while the valency elec- 
trons increase by one. It therefoie follows that every 
increase in the number of the periodic group of an element 
is accompanied by an equal inciease in numbei of elections. 

In passing from argon to krypton, eighteen successive 
unit increases in number of the peiiodic gioup occur, and 
consequently krypton must contain 18 more electrons than 
argon. As argon has aheady been shown to have 8 of these 
in one group (the ineit gas group completed in pas&ing 
from univalent copper via septavalent bromine) the group 
of eighteen electrons can be dissected into a group of 10 
and a group of 8, kiypton thus consisting of 2, 8, 8, 10, and 
8 electrons or 36 in all. 

A similar argument applies to xenon, which, taking 
account of the unknown eka-manganese, must consist of 
2, 8, 8, 10, 8, 10, and 8 electrons, or 54 in all. 

From xenon to niton, 29 elements are known, and 2 others 
are missing from the periodic group 7, making 31 possible 
valency changes, involving the addition of 31 electrons, 8 oi 
which are accounted for in the final inert gas group, niton 

Radioactivity and. Sub- Atomic Chemist-ty 139 

thus having the structure 2, 8, 8, 10, 8, 10, 8, 23, 8, or 85 
electrons in all, and uranium a similar structure with the 
addition of 6 valency electrons, making 91 electrons. 

As atoms are neutral structures, a positive charge must 
exist for every electron in the stiucture of an atom. Further 
as the weight of electrons is negligible in comparison with 
the weight of atoms, each election being about 1845 times 
lighter than the atom of the lightest atom hydrogen, the 
whole of the mass of atoms must be located in their positive 
charges, two of which are known to exist in the radio- 
active product, the a-particle. It has been indicated 
that the helium atom contains at least two outer electrons, 
removable to form the o -particle, and it is certain that 
hydrogen consists of one easily removable electron and one 
positive charge or nucleus responsible for its entire atomic 
weight. It is inferrable that the helium positive particle is 
the nucleus of a system of two outer electrons, and that 
every element is built up of a positive nucleus or nuclei and 
an independent system of electrons. 

The general law suggested by ladioactivity transforma- 
tions has already been indicated, that the atomic structures 
of all elements follow the same constitutional plan. This 
law applied to helium indicates that it must follow the 
plan of hydrogen in which the electron is accompanied by 
a single positive nucleus, helium thus consisting of four 
electrons and four positive particles, to account for its 
atomic weight of 4. It has already been shown that the 
evidence points to a single group of two electrons in helium 
removable to form the a-particle of radioactive changes. 
The remaining two non-removable electrons of neutral 
helium must therefore be firmly associated with the four 
positive charges. This complex of four positive particles 
and two negative electrons carrying a net positive charge 
of two can be identified definitely with the a-particle, thus 
being describable as the helium nucleus. Lithium with 
umvalency and atomic weight 7 must be derived from 
helium by the addition of three positive particles and three 

140 Chemistry and Atomic Structure 

negative particles or electrons. The law of atomic plai 
indicates that the helium outer group of two electron 
must persist in lithium, which being univalent must there 
fore consist of three outer electrons The remaining fou 
electrons must be associated with the seven positive particle 
as a nucleus, which is derived from that of helium by tin 
addition of three positive pai tides and two electrons 
Similarly beryllium (atomic weight 9) consists of 2 valenc] 
electrons, and a helium-group of two electrons, associatec 
with a nucleus of 9 positive particles and 5 electrons. Neor 
(atomic weight 20) consists of an inert-gas gioup of { 
electrons, and a helium-group of 2 elections, associatec 
with a nucleus of 20 positive particles and 10 electrons 
and xenon (atomic weight 130) consists of successive group 
from the surface inwards of 8, 10, 8, 10, 8, 8, 2, or 54 outei 
electrons, associated with 130 positive and 76 electron: 
forming a nucleus 

In general if the atomic weight of an element be W anc 
the number of outer electrons be N, the number of positive 
particles in the nucleus is equal to W, the number o 
nuclear electrons is equal to W N, and the net positive 
charge on the nucleus is equal to N. 

The foregoing considerations do not enable any piecisc 
position in the atom to be allocated to the nucleus, nor dc 
they indicate that the nucleus may not consist of sub 
sidiary nuclei scattered about the surface or inteiior of ar 
atom. The fact that the emission of helium nuclei fiorr 
radioactive elements is not accompanied by free or attachec 
electrons appears to indicate that the helium nuclei arc 
either extraordinarily minute as compaied with electron' 
and thus do not collide with electrons, or that the helium 
nuclei aie located on the atomic surface. The lattei 
suggestion is negatived by the fact that isolated helmir 
nuclei on or near the surface of an atom would be in close 
juxtaposition with many of the electrons of the valency anc 
outer electron groups, and would thus leadily acquire 
electrons and form neutral helium atoms, the atom thu< 

Radioactivity and Sub-Atomic Chemistry 141 

rapidly completely disintegrating by the loss of neutral 
helium atoms. Radioactive changes, however, invariably 
produce doubly-positively charged helium nuclei, and the 
disintegration process comes to a definite end in all cases 
with the formation of lead, an element of very high atomic 

The only conclusion to be drawn is that the helium nuclei 
are not located on or near the surface of any atom, and that 
the helium nuclei are extraordinarily small even compared 
with the electron. The simplest structure, conforming 
with the whole of the evidence from radioactive and non- 
radioactive elements, is that of a very small central positive 
nucleus, containing the whole mass of the atom, surrounded 
by successive groups of electrons, the order of the groups 
being that indicated already for niton and uranium on 
pages 138 and 139. 

The system of atomic structure above outlined is based 
entirely on the interpretation of experimental facts by 
means of a suggested law of uniform atomic plan inferred 
fiom radioactivity phenomena, and is independent of 
theories of atomic structure based on the dynamics of 
electrically charged particles. It will be shown in later 
chapters that the latter lead to an essentially similar system 
of atomic structure, and that a combination of the two 
systems results in atomic structures capable of offering a 
harmonious explanation of most of the known chemical 
and physical phenomena attributed to atoms 



THE existence of two opposite kinds of electrification had 
been recognised as long ago as the early part of the 
eighteenth century, and Franklin in 1750 put forward the 
" one-fluid " theory in which all non-electrified matter was 
assumed to possess a normal amount of the single electric 
fluid, positive electrification being an excess and negative 
electrification a deficiency of the electnc fluid. Fianklm 
further assumed that the electric fluid was material and 
consisted of extremely small particles, capable of moving 
without resistance through ordinary matter. Except that 
his electric particles were positive particles, Franklin may 
be said to have anticipated in a remarkable fashion the 
modern conception of the negatively charged material 
electron. Franklin's one-fluid hypothesis was however 
shown to be incapable of explaining a laige number of 
electiical phenomena without further equally fai -reaching 
assumptions, and it was abandoned in favour of the two- 
fluid hypothesis, in which electncity was assumed to be of 
two opposite types, positive and negative, and was not 
regarded as necessarily granular in structuie or even matenal 
in nature. 

About a century later, Weber 1 expanded Franklin's 
hypothesis of electiic particles into a theory of the stiucture 
of matter, which incorporated the two-fluid theory of 
electricity. Weber postulated two contrasting types of 
electnc particles, positive particles devoid of weight and 
negative pai tides responsible for the whole of the weight 
of matter. An atom was represented as a number of 
positive particles revolving round a massive central negative 
particle. Except for the interchange of electrical signs, 
this theory very closely resembles the modern theoiy of the 
atom containing rotating negative elections and a massive 
central positive nucleus. Weber further regarded an 
electnc current as a flow of weightless positive particles 

1 See Ch. IV, p. 50. 

Atomic Physics 

through the matter of an electric circuit, and applied the 
theory to an explanation of a great number of electrical 
and magnetic phenomena, and postulated the variation in 
mass of electric pai tides in viitue of their velocity. 

_Weber's theory of electric particles with mass varying 
with velocity was rejected by contemporary physicists, 
though Clerk Maxwell admitted in 1873 that Faraday's 
laws of electrolysis could be explained only on the ground 
of a constant irreducible unit of electric quantity. In 
developing his " Electromagnetic Theory of Light," how- 
ever, Cleik Maxwell definitely rejected the conception of 
discontinuous electrical quantities, and made use of the 
conception of electrical quantities susceptible of infinitesi- 
mal variation. This theory of the continuous nature of 
eneigy and electrical changes was greatly developed by 
Kelvin, Heaviside, and Sir J. J. Thomson, and became 
known as the " classical theory," and postulated the exist- 
ence of a continuous ether in and by which the radiations 
of heat, light, and electricity were propagated. The 
classical theory of electricity and energy chiefly concerned 
matter in bulk, but was largely divorced from matter and 
applied most successfully to the radiation of energy in free 
space. In recent years it has proved to require very con- 
siderable amendment in its application to the particles of 
mattei, in which radiation of all kinds has its ultimate 
origin. Had Clerk Maxwell accepted rather than rejected 
Webei's theory of electric particles, it is more than possible 
that the discontinuous character of the energy changes 
resulting in radiation would have been elucidated long 
prior to the establishment of Planck's " quantum theory " 
of 1900. 

Clerk Maxwell's deduction in 1873 of the unit or atom 
of electricity in electiolysis, was followed in 1874 by 
Johnstone Stoney's postulate of the natural unit of electric 
charge, the electrine (now electron), and by his identificaton 
of it in the rupture of each chemical bond. In 1881 
Helmholtz suggested the identity of electricity and chemical 

144 Chemistry and Atomic Structure 

affinity, unit affinity or valency being measureable by the 
number of units of electrical charge passing in electrolysis, 
and in the same year Sir J. J. Thomson showed that an 
electric charge must possess inertia or mass in virtue of its 
velocity, and calculated the rate of variation of such mass 
with velocity. 1 

In 1880, Crookes, as the result of his investigations of 
the nature of the cathode rays formed on the passage of 
electricity at high voltages through rarefied gases, came to 
the conclusion that these rays were matenal and carried 
negative electric charges, yet did not consist of solid, liquid, 
or gaseous particles, and suggested that they constituted a 
fourth form of matter, for which he proposed the term 
" radiant matter." 2 The material nature of these cathode 
rays, their power of penetration of matter, and their 
similarity whatever the gas operated with and whatever 
the material of which the cathode was made, suggested 
that the material pai tides of which the rays are composed 
are a constituent of all matter, and in 1891, Johnstone 
Stoney applied to these particles the name electron, and 
regarded their charge as identical with the fundamental 
unit charge proposed by him in i874. 3 He further 
attempted to show how, in the classical theory of light 
radiation, electrons vibiating within an atom could give 
rise to the light radiation causing the bright lines in spectra. 

In 1895, Lorentz 4 showed that many optical and 
electrical phenomena were susceptible of explanation on 
the assumption that oscillating or vibrating negatively 
charged particles existed in matter, and predicted several 
spectral phenomena which were later experimentally 

Two years later Sir J. J. Thomson announced that the 
cathode rays consisted of corpuscles, each having mass veiy 

i Phil. Mag, 1 88 1, [5], 11, 229 
a Proc Roy Soc , 1880, 30, 469 
s Scient Proc Roy. Dublin Soc , 1891, 583 

* An Attempt at a Theory of Electrical and Optical Phenomena in Moving 
Bodies, Leiden, 1895. 

Atomic Physics 145 

much less than that of the hydiogen atom, and carrying a 
negative electrical chaige of the same ordei as the election 
predicted by Johnstone Stoney in I874- 1 ^ n ^ ie same year 
Kauf maim 2 and Wiechert 3 obtained similar results, and 
the mass of the election is to-day known with an accuracy 
little less than that of the atom, and is about 1845 times 
less than that of the hydrogen atom. 

In 1900, Planck 4 put foiward the hypothesis that radia- 
tion is emitted or absoibed by oscillating or vibrating 
electric particles in a discontinuous manner, such that 
energy lost or gained by a vibiatmg particle is proportional 
to the frequency of vibration of the paiticle. He furthei 
assumed that the eneigy emitted or absorbed by the 
vibrating particle could not be emitted or absorbed in a 
continuously varying manner, but only in integial multiples 
of a minimum quantity, a " quantum " of energy, and that 
the piopoitionahty between eneigy and fiequency was 
conveited into equality by multiplying by a constant. This 
constant, known as " Planck's Constant of Action," was 
assumed to be the same for all vibiatmg pai tides in every 
sort of mattei. A vibrating particle having a fiequency of 
vibration v could therefore emit or absorb eneigy only in 
amounts of 7;v, h being Planck's constant. This theoiy 
onginally applied to the explanation of heat-radiation, and 
extended from oscillating to lotatmg electric particles, 
forms the basis of Bohr's theory of atomic structure and the 
explanation of optical and X-iay spectra elaborated by 
Bohr, Sommeifeld, and otheis. 

In 1903, Lenard 5 came to the conclusion that the elec- 
trons of the cathode lays, led out of a cathode tube thiough 
a thin metallic " window," weie absoibed by the an only 
after traversing such distances as indicated that only an 
extremely small proportion of the space m atoms is impene- 

Mag, 1897, [5], 44, 293 

2 Wied Ann , 1897, 62, 589 

3 Jf'ied Ann Supp , 1897, 21, 44.3 
* Ann P/yn, 1901, [4], 4, 553 

Ibid, 1903, [4], 12, 714 

146 Chemist* y and Atomic Structure 

tiable by electrons. He assumed that the impenetrable 
parts of atoms, called " dynamids," consisted of one 
positive and one negative electron, and calculated that the 
impenetrable portions of an atom only amounted to a 
thousand-millionth of the volume of the atom, and regarded 
the number of dynamids per atom as proportional to the 
atomic weight. 

In 1904, Sir J. J. Thomson, elaborating a suggestion of 
Kelvin, put forward a theory of atomic structure x in which 
an atom was assumed to consist of a sphere of positive 
electricity, having embedded in it a number of electrons 
or coipuscles equal to the positive charge, the electrons 
ananging themselves symmetrically in rings in accordance 
with the electrostatic forces. 

Between 1906 and 1911, the investigations of Rutherford 
and of Geiger, and their co-workers, on the scattering of 
a-particlcs from radioactive sources in passing through 
matter, led Rutheiford in 1911 to propose the theory of 
the nuclear atom, known as the " Rutherford atom," 2 in 
which the whole mass of the atom is regarded as concen- 
trated on a minute positively charged central nucleus, 
surrounded by a number of negative electrons equal in 
total charge to the charge on the nucleus. Subsequent 
investigations have amply confiimed the reality of this type 
of structural atom, which forms the basis of all existing 
theories, dynamic and static, of atomic structure. 

The work of Geiger having indicated that the charge or 
the atomic nucleus is approximately one-half of the atomic 
weight, van den Broek 3 put forward the hypothesis thai 
the charge on the atomic nucleus is equal to the numbe: 
of the element in the sequence of elements arrangec 
according to atomic weights, or the nuclear charge is equa 
to the " atomic number." 

In 1913, Bohr, then working with Rutherford in Man 

1 Phil Mag , 1904, [6], 7, 237 , 8, 548 , 1905, [6], 10 3 695 , 1906, [6], 11, 769 

2 Ibid, 1911, [6], 21, 669 

3 Nature, 1911, 87, 78 , 1913, 92, 373 and 476, Pbys Zeit , 1911, 12, 490 

Atomic Physics 147 

cliestei, put forward 1 a theory of the structure of the 
hydrogen and ionised helium atoms which incorporated 
Rutherford's theory of 1911 (the small nuclear atom with 
surrounding elections), van den Broek's hypothesis of the 
same year (the equality of the nuclear charge, the number of 
atomic electrons, and the atomic number), Planck's quan- 
tum theory of 1900 (the discontinuous nature of the energy 
radiated by a vibrating electric particle in an atom), and 
Balmer's law of the hydrogen optical spectrum of 1885 2 
as generalised by Rydberg in i890 3 and Ritz in 1908.* 
Bohr assumed that an atomic electron, circulating round 
the nucleus in vntue of the attractive force between the 
opposite electrostatic charges on nucleus and electron, could 
tiavel in circular orbits at fixed distances from the nucleus, 
proportional to the squares of successive natuial numbeis, 
without the expenditure of energy in radiation. In the 
classical theory of electro-dynamics based on Newton's 
and Keplei's laws of motion, an electric charge could 
revolve round an attracting charge only in continually 
decreasing orbits until finally the charges coalesced, the 
energy of the circularly " falling " particle being con- 
tinuously dissipated by radiation out of the system as 
electromagnetic light waves. The assumption of fixed 
orbits in Bohr's theory was, therefore, equivalent to the 
identification of uniform circular motion (in the classical 
theory involving centripedal acceleration) with rectilinear 
uniform or non-accelerated motion. These fixed orbits, 
in which electrons circulate without emission or addition 
of radiant energy, were called by Bohr " stationary states," 
but are strictly states of motion of the electrons they are 
only stationary states in the sense that the orbits in which 
motion takes place are fixed. Bohr further assumed that 
the difference in the energies of an electron in any two of 
the stationary states or fixed orbits represented twice the 

1 Phil Mag , 1913, [6], 26, i and 476. 

2 Ann Phys, 1885, [3], 25, So 

3 Compt rend, 1890, 110, 394 

4 Collected Works, Swiss Physical Society, Pans, 1911 

148 Chemistry and Atomic 

eneigy radiated or absoibcd by the electron in being tians- 
f erred from one oibit to another, and that the eneigy 
difference between two consecutive oibits was equal to twice 
the product of the frequency of the emitted ladiation and 
Planck's univeisal constant. Numbering the oibits fiom 
the centie outwards fiom I onwaids, the difteience between 
any two oibit numbeis lepiesented the number of multiples 
of the unit 01 quantum oi " moment of momentum " (not of 
eneigy] peculiar to each atom, and these numbeis were 
accordingly called quantum numbers of the election oibits 
in an atom. Combining Balmer's law, Keplei's and New- 
ton's laws, and Planck's quantum relation, Bohi deduced 
that the lines of the liydiogen spectium wcie given by the 

in which x and v are the wave-length and frequency 
respectively of the emitted radiation, e the election charge, 
m the electron mass, c the velocity of light, /; Planck's con- 
stant, and ! and n 2 the quantum numbeis of the final and 
initial oibits o the election. He also showed that the lines 
of the similai spectium of ionised helium (i.e. with one 
election) were given by the same foimula when the calcu- 
lated fiequencies were multiplied by foui, the square of the 
charge, 2, on the helium nucleus. 

In the same year, Moseley x began an investigation of the 
characteristic X-rays emitted by elements bombarded by 
cathode rays with the object of decidrng whether the 
atomic weight or the atomic number of an element was the 
deter mining factor in the frequency of the X-radiation of 
elements, and for this purpose analysed the emitted X-rays 
into a spectium foi photography by a diffraction grating 
consisting of the regularly arranged atoms in a crystal, 
X-rays being of too short wave-length to be analysed by 
ordinary ruled difli action giatings. 

Two groups of spectral lines, known as the K and L 

1 Phi. Mag , 1913, [6], 26, 1024 , 1914, [6], 27, 703. 

Atomic Physics 149 

seiies, were examined by Moseley. In Boln's theory the 
K lines aie assumed to be due to the '" falling " movement 
of an election to the inneimost oibit with the quantum 
numbei I fiom outer oibits, and the L lines to the " falling " 
movement of an election to the second innermost orbit 
with quantum number 2 fiom outer oibits, these " falling " 
movements of an electron being consequent upon the ejec- 
tion of an election of quantum number I or 2, lespectively, 
from the atom by the bombarding X-rays. 

The simplified Bohr formula connecting the fiequency of 
the ladiation vuth the atomic numbei is, 

= , = CN , R 

or = 

wheie N is the nucleai charge, R is Rydberg's spectroscopic 
constant, c is the velocity of light and \ and v aie the wave- 
length and fiequency respectively of the emitted ladiation, 
this formula applying for an election moving fiom a 
2 quanta to a I quantum oibit. Moseley, however, found 
in the case of the principal 01 a -line of the K senes of the 
elements examined that the equation was true only if N 
weie reduced by approximately one unit, and that such 
amended equation gave a sti aight-line graph, thus proving 
that the atomic numbei of an clement not the atomic 
weight is the characteristic factoi in detei mining the 
electric charge on the atomic nucleus 

Moseley also showed that sti aight-line graphs aie given 
on plotting frequencies of the L radiation against atomic 
numbei s, though Bohi's simple equation was satisfied only 
by i educing N in the equation by a constant, appioximately 
7-4, foi the L oc-line. 

A calculation of the values of these two constants from 
the most lecent deteimmations of the X-iay spectia of the 
elements indicates that " Moseley's constants " aie any- 
thing but constant, and in the case of the K ex-line, the 
" constant " vaiics fiom about plus 0-8 to minus 2-3, and 
fluctuates in a highly nregulai manner fiom element to 

150 Chemistry and Atomic Structure 

element. It approximates to I for elements only between 
atomic numbers II and 35, mostly examined by Moseley, 
and is practically zero for the elements of atomic number 
49, 50, 51, 56, and 57, i.e. indium, tin, antimony, barium, 
and lanthanum. The values for " Moseley's constant " for 
the bulk of the elements for the a, p, and y lines of the 
K series are shown in Diagram X. The value of the 
" constant " for the lines of the L series varies from nearly 
plus 14 to minus 4, but only from 6-8 to 7-4 for the L a-line. 

The discrepancy between the theoretical and calculated 
value of the nuclear charge disclosed by Moseley's work is 
presumed to be due to the screening effect of non-radiating 
electrons in the neighbourhood of the nucleus, the screening 
effect varying with the orbits of both non-radiating and 
radiating electrons. The pecularity of the irregular values 
of the " screening-constant," in passing from element to 
element, and the close coirespondence in the irregulaiities 
in the various lines of the K series do not appear to have 
been observed or to be susceptible of explanation in the 
foregoing way. 

Though Moseley's method cannot suffice to determine 
the precise value of the nuclear charge, it indicates quite 
definitely that the nuclear charge increases by one unit in 
passing from element to element in the classification of the 
elements according to atomic weight and chemical proper- 
ties, and consequently enables a decision to be made as to 
the precise order of the elements in the chemical classifica- 
tion in the cases of the three pairs of elements argon- 
potassium, cobalt-nickel, and tellurium-iodine, in which the 
atomic weights necessitate the oider potassium-argon, 
nickel-cobalt, and iodine-tellurium, as against the foregoing 
indicated by their chemical properties. Moseley's values 
for the nuclear charges show clearly that the order is that 
of the chemical properties, i.e. that of the periodic classifica- 
tion. The nuclear charge, further, enables a decision to 
be made as to the number of missing elements in the 
periodic classification, and indicates that elements of the 

Atomic Physics 




\ 20 


Moselsy's K series Constant 
X-ray Spectra 

ft fa Lin 9 



Mosefeys Constant 

152 Chemistry and. Atomic Sttuctute 

atomic numbeis 43, 61, 75, 84, 85, 86, 87, 89, and 91, are 
alone missing. Numbeis 84, 86, 89, and 91, have been 
identified with the ladioactive substances, polonium, niton, 
actinium, and protactinium, and their isotopes, but none 
of these, howevei, has been the subject of an atomic weight 
determination, and only one, niton, has been definitely 
chaiacteiised as a chemical element. The tact that the 
atomic number 61 alone is missing between lanthanum, 57, 
and lutecium, 71, indicates conclusively that the "laie- 
earth " tiansition period (see Ch. VI, p. 79) consists of 15 
possible elements, and that, theiefore, the 3rd long penod 
of the periodic classification consists of 32 possible elements 
(see Ch. X, p. 138). 

In 1886, Croolces 1 suggested that atomic weights might 
be aveiages of whole numbei atomic weights, elements thus 
being mixtures of atoms identical except in atomic weight 
This suggestion was confiimed by the discoveiy moie than 
twenty yeais latei of isotopes among the ladioactive 
elements, Soddy 2 in 1911 leviving Ciookes' suggestion 
that elements geneially weie mixtuies of chemically non- 
sepaiable atoms in a constant pioportion. In 1912, A S. 
Russell and Rossi 3 showed that isotopes weie not only 
identical in chemical piopeities but weie identical even in 
optical spectia, such atoms consequently differing only in 
mass and ladioactivity. 

It may here be remaiked, if the atomic weights assigned 
to the actinium series are conect, that isotopes may exist 
with identical atomic weights as well as identical chemical 
pioperties, eg. ladmm C x and actinium A, both with 
atomic w T eight 214, radium D and actinium B, ladium E 
and actinium C, polonium and actinium C l5 all six vuth 
atomic weight 210. Such pans of elements drftei in no 
respect except ladioactivity, then atoms being identical m 
mass, nucleai chaige, numbei of elections, valency, and 

1 Bnl Assoc Rep , 1886, 558 

-J Cbetii Soc , 1911, 99, 73 

3 Pioc Roy Soc , 1912, A, 87, 478 

Atomic Physics 153 

geneial physical and chemical properties. The extent of 
this identity throws extreme doubt on the accuiacy of the 
atomic weights assigned to the membeis of the actinium 
senes, not one of which has had its atomic weight detei- 
mined experimentally, nor even been definitely character- 
ised as a chemical element. In order that no isotopes 
should have identical atomic weight it would be necessaiy 
to reduce the actinium series by six units in atomic weight. 

In 1911, Sir J. J. Thomson 1 showed that the positive 
rays obtained in cathode tubes were material pai tides of 
atomic and molecular dimensions, and determined the 
absolute values of the masses of the atoms of many elements. 
The method of generating the positive rays was similar to 
that employed by theii discoverer Goldstein in iS86, 2 
positively chaiged atoms and molecules, created by iomsa- 
tion-loss of electrons in front of the negatively charged 
cathode, being allowed to " fall " thiough a hole or channel 
pieiced in the cathode. These " canal rays " weie analysed 
into homogeneous lays of similar type by means of super- 
imposed electric and magnetic fields, Sir J J. Thomson thus 
being enabled to obtain photographs of impinging paiticles 
in the foim of paiaboks on the plates, each parabola being 
due to paiticles having the same ratio of charge to mass. 
In an examination of neon, paiabolas were detected corie- 
spondmg to atomic weights 20 and 22, the piesence of 
doubly-chaiged atoms of calcium (atomic weight 40) and 
molecules of carbon dioxide (molecular weight 44) being 
luled out. The only explanation possible was that neon 
(atomic weight 20-2) was a mixtuie of atoms of atomic 
weight 20 and 22, the former piepondeiating, as was con- 
firmed by the compaiative faintness of the 22 paiabola. 
This constituted the fiist detection of isotopic atoms among 
the non-i adioactive elements. 

The existence was also pioved of tiiatomic hydrogen, and 
the oidmaiy atom of hydrogen was obtained with one 

1 Phil Mag, IQTT, [6], 21, 21^ , 19 r 2, [6], 2*, 209 
- h\rl Btr , iSSO, 39, 691 

154 Chemistry and Atomic Structure 

positive charge, but not more than one, thus confirming 
that the atom of neutral hydrogen possesses only one 
negative electron. The molecule of hydrogen was also 
detected with one positive charge, but never two positive 
charges, proving that chemical combination, if due to 
electrons, need be due to no more than one electron for a 
chemical bond This was further confirmed by the fact 
that triatomic hydrogen was never detected with more 
than one electron missing (or one positive charge), i e. with 
less than two present of the three electrons, for a molecule 
of three linked atoms must contain at least two chemical 
bonds. It was also proved that the inert gases are always 
positively charged, never negatively, thus indicating that 
they can lose but not gain electrons. In general it was 
found that the electronegative elements alone could give 
negatively unit-charged atoms, whereas all elements could 
give positively charged atoms, the acquirement of electrons 
thus being parallel with the chemical property of electro- 
negativity, and the loss of electrons being the common 
property of all elements The maximum number of elec- 
trons lost by any atom was proved to be eight, a significant 
fact taken in conjunction with the chemical fact that the 
maximum valency for any element is eight. 

Sir J J. Thomson's researches on positive ray analysis 
have yielded results of extreme chemical importance, 
altogether apart from the question of isotopic elements. 
His results may be regarded as an experimental demonstra- 
tion that chemical combination, measured numerically in 
valency, is due to the reactions of electrons, valency being 
numencally identical with electrons in the case of iomsable 
compounds , that electronegativity is identical with gam 
of electrons and elect ropositivity with loss of elections , 
that one election may suffice for the chemical bond in the 
case of non-ionising compounds ; that the structuies of the 
atoms of the inert gases, in no circumstances, accommo- 
date any more electrons , that the atoms of hydrogen and 
helium cannot yield stiuctuies containing more than two 

Atomic Physics 155 

electrons per atom ; that the maximum number of electrons 
that an atom can lose is eight ; and that the maximum 
number of electrons that an atom can gain is one. 

Sir J. J. Thomson's method of positive-ray analysis was 
fuither improved by Aston in 1919 l by an adjustment of 
successive electric and magnetic fields so that the particles 
having the same ratio of charge to mass are focussed on to 
the photographic plate as lines, which are the images of 
the slit through which the rays pass. This arrangement, 
known as the mass-spectrograph, disperses the various types 
of rays over a wide range while giving sharp definition to 
each type of ray, and it has been possible to measure the 
masses of the atoms or molecules forming the rays with an 
accuracy of I in 1000. The most important result obtained 
has been the proof that the majority of the elements are 
mixtures of isotopic atoms having atomic weights which 
are integral multiples of one-sixteenth of the atomic 
weight of oxygen, or, in other words, that atomic weights 
are whole numbers. The only outstanding exception is 
hydrogen, with an atomic weight of 1-008. This has been 
" explained " by the supposition that part of the mass of 
all other atoms has disappeared in the condensation of the 
atomic nucleus from electrons and hydrogen nuclei. If all 
mass is merely the inertia of an electric charge, two opposite 
electric charges in very close juxtaposition will partially 
obliterate one another, with resulting diminution in inertia 
due to electric charge, i.e. loss of mass. In the case of 
helium, the nucleus of which is supposed to be made up 
of four hydrogen nuclei, called protons, and two electrons, 
the loss of mass is the difference between 4 x 1-008 and 4, 
that is 0-032, the difference between the atomic weight of 
four free atoms of hydrogen and the observed atomic 
weight of helium. 

It may be remarked that this " explanation " of the 
deviation of hydrogen from the " whole number rule " has 

1 Phil Mag, 1919, [6], 38, 707, 1920, [6], 39, 611, 40, 628, 1921, [6], 
42, 140 and 436 , Isotopes, London, 1922 

156 Chemist? y and Atomic Sttuctute 

no evidence to suppoit it, and that some evidence exists 
which thiows doubt upon it. Accoidmg to the most 
recent work of Ellis 1 on the [3- and y-iays of ladioactive 
elements, the electrons in the nucleus exist in quantum 
orbits similar to those of the outer elections. It must be 
supposed that the elections in the mtra-nucleai oibits 
possess much greatei orbital velocity (owing to the 
extiemely minute size of the nucleus as compaied with 
the atom) than the extia-nuclear electrons. As the K 
electrons of uranium have a velocity (125,000 miles per 
second or 67 pei cent of light speed) nearly sufficient to 
add peiceptibly to the mass of the elections, the velocity 
of the intra-nucleai electrons, even in the case of helium, 
must be sufficient to make an appreciable addition to the 
mass of these electrons. Helium should consequently have 
a greatei mass than foui atoms of hydiogen, which has no 
intia-nuclear electrons, a deduction which is not borne 
out in fact. This does not necessarily indicate the non- 
existence of intia-nuclear quantum oibits with electrons 
whose speed is not a negligible fi action of the velocity of 
light. It may meiely indicate that the nuclei of all 
elements, hydrogen not excluded, contain a number, 
possibly a very laige numbei of electrons, and that the 
mass of all nuclei is due partly or entirely to the mass of 
elections as a consequence of their velocity. It may yet be 
necessary to assume that the positive chaige of electuary 
is totally divoiced fiom mass, and that mass is a function 
of negative electucity. It is lemarkable that oui present 
theories of mass and positive electucity aie founded less 
on any positive knowledge of positive electricity than on the 
negations known of negative electricity 

A difteicnt method of generating positive lays, due to 
Gehicke and Reichenheim z has been elaborated by Demp- 
ster, 3 in which a metallic salt is electrically heated on a 

1 PIOC Camb Phil Soc , 1922,2!, 121. 

2 Veih Dent Phys Geiell , 1906, 8, 559 , 1907, 9, 76 

J Pbys. Rev , 1908, 11, 316 , 1921, 18, 415 , 192?, 20, 631 

Atomu Physics 157 

hollow platinum anode and simultaneously bombaided by 
electrons fiom a cathode, and the libeiated positive particles 
acceleiatecl by a small electnc field and focussed by deflec- 
tion lound a semicncle by a magnetic field. The advantage 
of the method is that it can be used foi non-volatile com- 
pounds, wheieas the cathode "canal ray" method is 
applicable only to gases which can be ionised in the space 
in front of the cathode. The lesults obtained by both 
Dempster and Aston, using Dempster's method, have 
largely and supplemented the facts elucidated by 
Aston, using Sir J. J. Thomson's modified method, that 
the majonty of the elements are mixtuies of isotopes con- 
foiming generally to the " whole number lule " of atomic 
weights. In the case of selenium the atomic weight of one 
of the isotopes, 82, is identical with the atomic weight of 
one of the isotopes of laypton, and these non-radioactive 
elements therefore contain isobanc atoms, many cases of 
which aie known among the radioactive elements. 

The veiy gieat eneigy of the massive helium nuclei 
(<y~pai tides) ejected fiom the atoms of the radioactive 
atoms, due to the enormous velocities of the ejected 
a-pai tides (about 12,000 miles per second for the a-particles 
from ladium C), led Ramsay in 1907 to bombaid other 
elements with them in the hope of disintegrating then 
atoms. Though Ramsay announced seveial cases of dis- 
integiation and transformation, using this method, his 
lesults have not been confirmed. Nevertheless this method 
has within very lecent yeais in the hands of Rutheiford, 1 
and Rutherfoid and Chadwick, 3 led to what appears to be 
leal disintegration of one or more of the lighter atoms. 
When such eneigetic bombardment of hydiogen atoms 
takes place, some of the hydiogen atoms aie piopelled 
foiwardb with a velocity that gives them a range of travel 
of about 28 cms. in air, the a-particles themselves, being 

iPW Mag, 19^9, M, 37, 537, >9^, [6], , 37 5 *' %' Soc "9> 
A, 97, 374, Nature, 1922, 109, 614 
Pbtl. Mag., 1911, [6], 42, 8oq , 1921, [6], 44, +17 

158 Chemist? y and Atomic Situctute 

helium atoms four times as heavy, having a range in air 
of only a quarter, i.e. about 7 cms. When, however, the 
gas bomb aided was nitrogen, particles identified as hydro- 
gen atoms, having a range of 40 cms , weie obtained, and, 
after ruling out the possibility of impurities, Rutheifoid 
assumed that the hydiogen particles arose from the dib- 
mtegiation of atoms of nitrogen. Hydrogen particles of 
range greater than 28 cms were also obtained on bombaid- 
ing compounds containing atoms of boion, fluorine, sodium, 
aluminium and phosphorus, having the atomic numbers 5, 
9, II, 13, and 15. In all of these cases hydrogen particles 
having a range greater than 28 cms. were also found to be 
ejected in the backward dnection by the bombardment, 
though in the case of nitrogen, of atomic numbei 7, the 
backwaid range was only 1 8 cms. No paiticles having a 
range exceeding 28 cms. were obtained from the bombard- 
ment of lithium, beryllium, carbon, oxygen, neon, mag- 
nesium, silicon, sulphur, chlorine, argon, and potassium, 
having the atomic numbers 3, 4, 6, 8, 10, 12, 14, 16, 17, 
1 8, and 19. The general conclusions are that long lange 
hydrogen particles are obtained only from elements of odd 
atomic number, but not from any elements of atomic 
number less than 5 or greater than 15. Rutherfoid regards 
the result as conclusive evidence that the atoms giving rise 
to the long range hydrogen particles actually contain 
hydrogen nuclei (protons) m their nuclei, and that a real 
disintegration has been obtained. He has also announced 
that mtiogen and oxygen yield particles of range 9 cms. 
having an atomic weight of 3 and a double positive charge, 
and has suggested that these atomic nuclei contain particles 
of atomic weight thiee, i e consisting of three protons and 
one electron. 

It is possible, however, that the hydrogen particles of 
mass I, and the short-range particles of mass 3 do not come 
from the atoms bombarded, but from the bombarding 
helium nuclei of mass 4, and that the results are due to the 
disruption of the helium nucleus into two parts of masses 

Atomic Physics 159 

i and 3, each part with one electron, the particle of mass I 
having its electron " brushed " off in its long travel before 
producing the scintillation by which it is detected, and the 
particle of mass 3 falling to pieces after a short travel to 
give rise to two hydiogen nuclei and one hydrogen atom, 
or one hydrogen nucleus and one ionised hydrogen mole- 
cule, just as ionised triatomic hydrogen falls to pieces on 
further lonisation. The unsymmetrical disintegration of 
the helium nucleus may be attributed to the unsymmetrical 
structure of the bombarded atoms of odd atomic number. 
Atoms of even atomic number may, on the other hand, 
disrupt the helium nucleus symmetrically into two ionised 
molecules of hydrogen, which would have a short range 
owing to their mass and thus allow the disintegration to be 
overlooked. Rutherford's results are truly evidence of 
disintegration of elements, but perhaps only of the radio- 
actively-produced a -particle or helium nucleus. Until this 
possibility has been definitely excluded, judgment may 
properly be withheld as to the precise nature of the 
disintegration actually effected. 

162 Chemistry and Atomic Structure 

ether shell, during the process of entry and ejection of 
ether pai tides, is constant over the whole suiface of the 
shell no mattei what its radius may be, i.e. no matter how 
far the wave has travelled in space. This transverse- 
compiession wave meeting an electron will subject the 
electron to the pressure of the shell, and it then becomes a 
question as to whether the wave will continue by movement 
of ether particles or will move the electron. If the electron 
be moved and its volume is equal to that of the ether 
particles being transferred fiom shell to shell, the com- 
pression in the shell fiom which the electron moves must 
disappear and the whole sphencal wave throughout space 
suddenly cease to exist, even though the electron occupy 
only a very small area of the whole spherical shell. The 
electron thus acquires the energy of the light wave. If 
the election were originally in motion its kinetic eneigy 
would be inci eased by the foregoing amount, and so long 
as the electron travelled in space wheie successive dis- 
placements did not alter in the state of compiession of 
ether particles the electron would continue to move with 
this fixed amount of kinetic eneigy. This could occur in 
space far from the nuclei of atoms or on the surface of 
shells of ether particles about a nucleus where the extent 
of compression was uniform An electron could therefore 
tiavel with undiminished energy in the spherical shell 
surrounding an atomic nucleus. This spherical shell con- 
tains all equivalent Bohr orbits. If the electron weie 
originally moving in such a shell it could find a place for 
uniform circular motion (without loss of kinetic eneigy) 
in any other shell or Bohr orbit. 

The foregoing mechanism provides only for the exchange 
of energy between electrons and light waves and for the 
cnculation of electrons in fixed orbits without loss of 
energy, and makes no provision for the exact positions of 
the vaiious possible Bohr orbits, nor for any relation 
between the energy radiated and the wave length of light 
radiated in transfers of electrons from one orbit to another, 

The Dynamic Atom 163 

nor for the " quantising " of the successive orbits in equal 
amounts of moment of momentum or angular momentum. 
It is merely a qualitative picture and has no claims to 
reality. A simple extension of the mechanism, however, 
provides for the reflection of light by electrons but not by 
nuclei, and for the known fact that all reflected light is 
plane polarised, and indicates that the plane of polarisation 
is that of the electron orbit. 

Bohr's theory of quantised circular orbits was originally 
applied to atoms, hydrogen and ionised helium, containing 
only one electron outside the nucleus, and was successful 
in indicating the precise positions of the various lines in 
the optical spectra of the two elements. It also predicted 
the existence of spectral lines later discovered, and assigned 
the correct value to the dimensions of the atom, and the 
exact value of Rydberg's spectral constant. It must there- 
fore be admitted as a valid qualitative and quantitative 
mechanism for Ruthei ford's atom, and it is the only exist- 
ing theory of the atom winch is in conformity with the 
known facts of atomic structure and spectrum analysis. 
It must consequently be accepted that the atom of the 
physicist and the chemist is a dynamic atom, and theories 
based on static electrons must give place to it, no matter 
how difficult the conception of the dynamic atom may be 
for the mechanism of chemical combination. 

Bohr's theory of simple circular orbits was considerably 
expanded by Sommerfeld's theory that elliptic orbits as 
well as circular orbits must be possible for electrons circu- 
lating round a nucleus, as in the case of astronomical 
bodies. 1 Sommerfeld's postulate was that an electron 
moves in an ellipse at one focus of which the nucleus is 
situated, and he utilised a mathematical expression for 
elliptic orbits, developed by Wilson 2 and almost simul- 
taneously by himself 3 in 1915, characterised by two 

1 Atombau und Spektralhmeji, Munich, 1919, Atomic Structure and Spectral 
Lines, London, 1923 

- Pbd Mag, 191 6, [6], 31, 161 

3 Mun Eer , 1915, 459 , Ann. Pbys , 1916, 50, 125. 

164 Chemistry and Atomic Structure 

quantum numbers, a radial and an azimuthal quantum 
number, elliptical motion being a pioblem of two degrees 
of freedom of the moving particle, the position of which 
is determined by the radial distance from the focus and by 
the angle (azimuthal angle) between the variable ladms 
and the major axis of the ellipse Sommeifeld extended 
this mathematical expression to the equation of Bohr's 
form of Balmer's law. He showed that the energy of the 
elliptic orbit was identical with that of the circular oibit, 
the energy of the elliptic orbit being a function of the sum 
of the azimuthal and radial quantum numbers, whereas 
that of the circular oibit is a function of the azimuthal 
quantum number only, the radial quantum number 
becoming zero owing to the constancy of the radius. He 
further showed that the number of possible types of oibits 
was numeiically equal to the numbei of ways of obtaining 
the quantum sum of the ladial and azimuthal quantum 
numbers, being one for circulai orbits wholly made up of 
azimuthal quantum number, and -i foi elliptic orbits 
having both sorts of quantum number equal to n in sum. 
The elliptic orbits for the sum numbei 2 are one, two for 
the sum number 3, thiee for the sum number 4, and so on. 
The sum of the quantum numbers or total quantum number 
thus repiesents one, two, three, four, etc., types of orbits, 
including both ciicular and elliptic, when the total quantum 
number is I, 2, 3, 4, etc., or in general the total quantum 
number n lepresents types of orbit. Bohr's form of the 
Simple B aimer Law, 

thus became 


in which a^ and r l are the azimuthal and radial quantum 
numbers of the final orbit, and a z and i z are the conespond- 
ing numbers for the initial orbit of the electron giving rise 
to the radiation. 

The Dynamic Atom 165 

Fuither the fact that the total quantum number 2 could 
represent two orbits and the quantum number 3 three 
orbits indicated that the B aimer series in which an electron 
moves from a 3 to a 2 quanta orbit, necessitated that an 
electron in each of the three 3 orbits could move to each of 
the two 2 orbits, each simple line of the Balmer series 
of spectral lines thus consisting of six lines m two sets of 
triplets all close together, if the energies of the electron in 
each orbit were slightly different. Sommerfeld calculated 
that the energy of the various orbits should not be quite 
identical for the same total quantum number, and that 
the more elliptic orbits should have slightly greater energy 
in proportion to their eccentricity, owing to the slight 
increase in mass of the electron in virtue of its greater 
velocity when moving near the nucleus (at the ellipse 

Sommerfeld's prediction of this "fine structure " of each 
of the Balmer lines has been realised by the magnification 
and separation of the lines by more accurate spectroscopic 
methods, and the calculated extent of the separation of the 
lines agrees precisely with the observed sepaiation. This 
may be accepted as very conclusive evidence for the reality 
of Bohr's theory as amended by Sommerfeld and for the 
variation of the electron mass with velocity (see Appendix). 

The Bohr-Sommerfeld theory has been extended from 
the optical spectra of hydrogen and ionised helium atoms 
to the X-ray spectra of elements in geneial, and with slight 
modifications has proved capable of adequately accounting 
for the manifold phenomena, both qualitative and quantita- 
tive, of the numerous known X-ray spectral lines, and all 
the known lines have now been piecisely allocated to the 
movements of electrons from outer to innei oibits conse- 
quent on the ejection of inner electrons on X-ray bombard- 
ment. The related series of X-ray spectial absorption 
bands have similarly been precisely allocated to the ejection 
of elections from inner to outei unoccupied orbits or to 
outside the atom The inteipietation of the absorption 

1 66 Chemistry and, Atomic Structure 

spectra in terms of the theory has led to very exact know- 
ledge of the various types of electron orbits, which have 
been found to be moie numeious than the simple form of 
quantised elliptic orbits indicated. It is now known that 
the quantum number 2 represents not two but three types 
of orbit, only one of which is circular, that the quantum 
number 3 represents not three but five types of orbit, only 
one of which is circular, and that the quantum number 4 
represents not four but seven types of orbit, only one of 
which is circular. 

In the simple theory of elliptic orbits, any particular 
orbit is represented by the notation n k) where n is the sum 
of the radial and azimuthal quantum numbers and k is the 
azimuthal quantum numbei, the single I quantum or K 
orbit being I 15 the two 2 quanta or L orbits, being 2^ and 
2 2 , the three 3 quanta or M orbits being 3^ 3 2 , and 3 3 , and 
so on. The new oibits pioved to exist by the absorption 
spectra are characterised by the notation, K seiies ija, 
L series 2 1 b, 2^, 2 2 a, the M series 3 1 b, 3^, 3 2 a, 3 a b, 3 3 b } 
or by the still newer notation, i u , 2 115 2 12 , 2 32 , 3n, 3i 2 , 

322, 323, 3sa, 4n> 4i2, 422 5 423, 4s3 3 4s4, 444, thus including 
the one K orbit, the thiee L orbits, the five M orbits, 
and the seven N orbits. A similar scheme of notation has 
been applied to the O and P orbits known to consist of 
five and three types, 5 11} 5 12 , 522, 523? Sss, an ^ 611, 6 12 , 
and 6 22 . 

In ciicular orbits the enetgy of the circulating electron is 
inversely proportional to the diametet, and in elliptic orbits 
to the majoi axis, the diameter of circular orbits being 
identical with the major axis of elliptic orbits having the 
same total quantum number. The diamete-is of the circular 
and the major axes of the elliptic orbits are dwectly pro- 
poitional to the squares of the total quantum numbei s, while 
the lateia recta (shoitest chords through the nuclear focus) 
are directly pioportional to the squares of the azimuthal 
quantum numbets. Fiom the geometrical propeiues of the 
ellipse, it follows that the minor axes of elliptic oibits aie 

The Dynamic Atom 167 

directly propoitioiial to the poduct of tlie total and 
azimutbal quantum numbers, for the minor axis of an 
ellipse is equal to the square root of the product of the 
major axis and latus rectum. Similarly it follows that the 
distance between the foci of elliptic orbits is equal to 
wVV __ 2 } w here n and k are the total and azrmuthal 
quantum numbers. This expression reduces to zero foi 
ciicular orbits, the foci being the single centre. For elliptic 

n( \ 

orbits the perihelion ladius is equal to -\n- Vn z k*9 


n z 
this expression reducing to for circular orbits, which is 


equal to the radius of the circle. Fiom these expressions 
the forms of the vaiious circular and elliptic oibits of any 
total and azimuthal quantum numbers can be readily con- 
structed. All the possible types of orbit of total quantum 
numbers I, 2, 3, and 4 are shown in Diagrams XI, XII, 
XIII, and XIV. For the simple circular and elliptic 


of f/ectron Orbits 

T/iree Orbital Types of 
QuanLum Number 2 

oibits of the Bohr-Sommeifeld types, the orbits are i ll3 

211, 2 22> 3ii> 322, 333, 4n> 422 3 433, 444- The evidence of 
absorption spectra indicates one additional 2 quanta orbit, 


Chemistty and Atomic Structure 

Planar B-'p-esentabton 
of E/ectron Orbits 

Orbital Types of Quantum Number 3 
Unsymmetncal 3 /2 and 3^ 3 Orbits 

P/anar Representation of E/ectron 

J>W? Orbital Types of Quantum Number 4- 
Unsymmetr/ca,' 4-,- <&, ancf 4- Orbits 

' 3 J4 

The Dynamic Atom 


two additional 3 quanta orbits, and three additional 4 
quanta orbits. The Bohr-Sommerfeld theory has not, 
however, prescribed any forms for these additional orbits. 
From their azimuthal quantum numbers it may be inferred 
that they are closely related to the simple ellipses of a single 
azimuthal quantum number. Two possibilities are obvious, 
first that the orbits are unsymmetrical about the minor 
axis, and secondly that they are symmetrical about both 

P/anar Qepresentat/on of Electron Orbits 



axes. The first type (see Diagrams XI, XII, and XIII) 
coi responds to an orbit of one azimuthal quantum near the 
nucleus and of the next higher azimuthal quantum number 
distant fiom the nucleus, the orbit thus having two different 
latera recta. The second type (see Diagram XIV) corre- 
sponds to an 01 bit having an azimuthal quantum number 
the mean of two consecutive azimuthal quantum numbers. 
It seems probable that the unsymmetrical type of orbit 
corresponds better to the evidence, and its formation may 

170 Che mist? y and Atomic Structure 

be attributed to the unsymmetiical type of electiic field 
caused by an electron approaching the nucleus in an elliptic 
orbit. This corresponds with the known fact that orbits 
of two diffeient azimuthal quantum numbers are not 
evident m the optical spectra of atoms having a single 
valency electron, and that such orbits are always evident 
in X-ray spectra, for in these cases atoms have pre-existing 
orbits of the simple elliptic type. 

The unsymmetrical 2 12 orbit is shown in Diagram XI, 
the two unsymmetrical 3 12 and 3 23 orbits in Diagram XII, 
and the three unsymmetrical orbits 4 12 , 4 23 , and 4 34 in 
Diagram XIII, and the whole of these are shown in 
Diagram XV, which indicates how the elliptic oibits pene- 
trate not only into othei elliptic and ciicular orbits, but 
even into the inner parts of similar oibits. These diagrams 
are no more than an indication of the complexity of inter- 
penetrating orbits, and in a real atom the complexity is 
vastly increased owing paitly to the many electrons in each 
class of orbit and to the fact that the orbits do not exist in 
a plane, but are symmetrically disposed about the nucleus 
in all the three dimensions of space. Only by an extreme 
licence in the use of words can each class of orbit be regaided 
as constituting an energy level or shell in an atom, for the 
orbital interpenetration makes a precise conception of levels 
impossible except in the case of truly ciicular orbits. 

Sommeif eld's deduction that the energy of an election 
circulating in an elliptic orbit is slightly gi eater than that 
of an electron in a corresponding circular orbit, due to the 
increase of electronic mass with increase in velocity, necessi- 
tated that the mass should increase as the electron 
approached the nucleus, and consequently that the 
electrostatic foice of attraction should not be quite sufficient 
to retain the heavier electron in the path of the theoretical 
orbit, the electron thus travelling slightly outside the 
theoretical orbit. This continuous falling away of the 
electron in approaching the nucleus causes the real 
approaching path to lie iurther and further away fiom 

The Dynamic Atom 






c- Q 

^ -C3 


95 ^ 
" c^ 

K Qj 





<r> I 




172 Chemistry and Atomic Structure 

the theoretical orbit, and the electron to pass close to the 
nucleus at a perihelion position further round the nucleus 
in the same clockwise direction as the curved path of 
approach. On receding from the nucleus the velocity 
diminishes and the mass now deci eases, the electron thus 
describing a path inside the theoretical orbit, and returning 
to an aphelion in a new position. As the circulation of the 
electron recurs, the actual path similarly diverges from the 
new theoretical path (see Appendix), the election thus 
describing a path which is not closed and resembles an 
eccentric spiral. This type of path can be regaided as 
compounded of a circulation of the electron in a true 
ellipse the major axis of which is turning slowly about the 
focus, so that the penhehon of the ellipse moves in the 
diiection of the outward poition of the ellipse and the 
aphelion in the direction of the inward portion of the ellipse. 
This perihelion motion is commonly described as the 
lelativity effect (see Appendix), and is analogous to the 
motion of the penhehon of the rapidly moving planet 
Mercury in describing its elliptic orbit about the sun. It 
is impossible to be content with asciibmg this penhchal 
motion to a mathematical piopeity of space, and it is 
natuial to legard so-called lelativity effects as due to a real 
material structure of space. The ether mechanism sug- 
gested at the beginning of this chapter provides a more 01 
less adequate explanation of this penhehal motion. If the 
particles in the ether shells are increasingly compressed the 
nearer to the nucleus, an electron approaching the nucleus 
travels fiom a shell of less to a shell of slightly greatei 
compression or density and is accordingly slightly letaided 
in its inward motion to the nucleus, and tends to travel in 
the circular path in a shell of uniform compression, i.e. m 
a true Bohr circular orbit. Put m another way, an approach- 
ing electron tends to be deflected so as to travel on an oibil 
of uniform moment of momentum, and the lesultant path 
is a compromise between the initial and tending directions 
This path is at eveiy point a compromise between an 

The Dynamic Atom 173 

ellipse and a circle, and is thus a deformed ellipse the 
perihelion of which rotates in the same clock-wise direction 
as the electron circulates in the ellipse. 

If the ether shells are of continuously decreasing density 
(the ether particles may be regaided as true spheres in fiee 
space and as deformed spheres approaching cubes when in 
compression) the further fiom an atomic nucleus, the ether 
round any collection of atoms, forming a mass of matter 
of any size, must deciease in density the further from the 
mass of matter, and perihehal motion of all planets in 
elliptical orbits should therefore occur, but will be observ- 
able only when the orbital velocity is sufficiently great to 
cause an appreciable inciease in mass. A similar ether 
mechanism may be advanced to explain the bending of 
light rays, correctly predicted by the relativity theoiy, in 
passing close to the sun. The observed effect has been 
suggested by Sir Oliver Lodge as due to refi action of light 
waves possibly caused by the increased density of ether 
near laige masses of matter. 

Sommeif eld's relativity effect for electrons in elliptic 
orbits was originally applied to atoms containing only one 
circulating electron, The effect must, however, exist to 
an enhanced extent in atoms containing many elections in 
elliptic orbits, for the velocities of the electrons increase in 
propoition to the atomic numbers of the atoms. In the 
case of the one-electron atom the relativity theory predicts 
that the plane of rotation of the perihelion will be the 
plane of the orbit. This conception of orbital precession 
becomes impossible in the case of elliptical orbits where 
the number of electrons in such orbits are more than two 
or three, owing to the spatial distribution of the electron 
01 bits. The conception of the planar motion of the orbital 
perihelion must therefore definitely be abandoned ; yet the 
increase of mass with mci easing velocity must exist, for 
the separation of X-iay spectral lines clearly indicates that 
the electrons in the various orbits of the same total quantum 
number have not the same energies. This addition of 

174 Chemistty and Atomic Structure 

energy to the non-circulai orbits must manifest itself in 
some form of orbit, though the numbers of the electrons 
and the extent of interpenetration make it impossible for 
the orbit to be an ellipse rotating in a plane containing the 
elliptic orbit. The positions of the vaiious X-iay spectral 
lines with relation to one another, further, prove that the 
energies of the elliptical orbits are the same as if the ellipse 
were rotating in a plane, and it is thus evident that the 
ellipse is substantially pieserved together with the extent 
of its precession. The extent of the precession effect may 
be regarded as a function of the diameter of the circle 
swept out by the rotating latus rectum, which is dnectly 
proportional to the square of the azimuthal quantum 
number of the elliptic orbit. In Sommerfeld's relativity 
effect this circle is desciibed in the plane of the orbit. This 
may be ascribed simply to the absence of forces caused by 
othei electrons. In the atom of many electrons, however, 
the mutual repulsions of the electrons with orbits spatially 
disposed about the nucleus will cause a precessing electron 
to remain within a small domain in space, though the 
attraction of the nucleus will maintain an effective elliptic 
orbit, This interaction of the nucleus and the surround- 
ing electrons with a precessing electron will thus cause the 
precession effect to take place in a plane at right angles to 
the axis of the ellipse, thus fixing the perihelion in one 
place in space and consequently also the aphelion and the 
major axis. The resultant motion of the electron is the 
compounding of an elliptic motion with a circular motion 
at right angles, the path described by the electron being 
on the surface of the solid obtained by rotating the ellipse 
about its fixed major axis. The precession motion for 
elliptic orbits is thus confined to spiral paths described on 
the ellipsoid of revolution about the major axis, the election 
twisting rapidly round a narrow spiral neai the nucleus and 
a wider spiral neai aphelion. A representation of this con- 
tinuous spiral orbit of an electron having the quantum 
number 43 is shown in Diagram XVI. 

The Dynamic Atom 175 

This fixation of the major axis of elliptic orbits, while 
allowing for precession and increased energy with inci eased 
velocity, removes the chief obstacle to the utilisation of 
dynamic atoms in the chemical combinations between 
atoms foi which the valency directions are definitely known 
in chemistry to be fixed in space round the atoms. This 
utilisation of dynamic atoms is obviously impossible so long 
as the aphelion of an elliptic orbit of a valency electron is 

Spat/3/ Representation of Electron Orb/td/ Domain 

Precession Paths of a 4.3 Electron moving on the Surface of 'an 
Imaginary Sol/d Ellipsoid of Revolution wth Nucleus at the focus 

regarded as necessarily lotating round the periphery of 
an atom. 

The Bohi-Sommerfeld theory of atomic structuie, being 
deduced fiom considerations of optical and X-ray spectra, 
specifies no more than the possible orbits in which electrons 
can rotate round an atomic nucleus, and in its present form 
can furnish no evidence as to the actual numbers of elec- 

176 Chemistry and Atomic Structure 

trons rotating in the various types of orbit. For the 
purpose of determining the total numbers of electrons in 
the orbits of any particular total quantum number, Bohr 
has had to enlist the aid of the chemical evidence, particu- 
larly that of valency and the periodic classification of 
elements according to atomic weight and chemical pro- 
perties. This evidence, however, is of too general a nature 
to determine accurately the distribution of the electrons in 
the various subgroups corresponding to the various azi- 
muthal quantum numbers of electrons with any particular 
total quantum number. This distribution can be ascer- 
tained only from the consideration of the minute details of 
the idiosyncrasies of the chemical properties of each 
element separately and in relation to other allied and con- 
trasting elements, and this is peculiarly the knowledge of 
the chemist not the physicist It may therefore be expected 
that future advances in the knowledge of the intimate 
structure of atoms will be made largely in chemistry, 
though it will be shown in Chapter XIV that the vast 
chemical knowledge already accumulated is now competent 
to decide the structural details foi many of the elements. 




RYDBERG'S 1 investigations of the numerical relations 
between the atomic weights of the elements in the periodic 
classification led him to conclude that the pioperties of the 
elements are peiiodic functions of a number, a common 
factor in atomic weights, and to propose a series of oidmal 
numbers for the elements, helium being number 4, by 
assuming two elements between hydrogen and helium. 
These ordinal numbers, reduced by 2, were identical with 
the present atomic numbers up to uranium, he thus being 
successful in detei mining the numbei of missing elements 
even before van den Broek had suggested or Moseley had 
established the precise values of the atomic numbei s. 

Rydberg was enabled to infer the number of missing 
elements by his discovery that the elements in the periodic 
classification fell into quadiatic groups each consisting of 
4p 2 elements, p being an integer increasing from I by unity 
for the various quadratic groups The first group con- 
sisted of 4 X I 2 = 4 elements from hydiogen to helium, 
the second gioup (two periodic groups) of 4 x 2 2 = 16 
elements, the third group (also two periodic groups) of 
4 x 3 a = 36 elements, and the fouith group of 4 x 4 2 = 64 
elements. Excluding the first group, the second and 
third peiiodic groups each thus consist of 2 X 2 2 = 8 
elements, the first two long periods each of 2 x 3 2 = 18 
elements, and the last period of 2 x 4 2 = 32 elements. 
The various periodic groups m pairs were, consequently, 
equal to twice the squares of the natural numbers 1,2, 3, 
and 4. These quadratic groups aie now termed Rydbeig's 
series and are expressed by the formula 2 (i 2 , 2 2 , 2 2 , 3 2 , 3 2 , 
4 2 , etc.), or 2, 8, 8, 18, 18, 32, etc. Aftei Moseley's deter- 
minations of the atomic numbers, Rydberg' s series weie 
seen to express accurately the number of elements in the 
successive periodic groups 

1 Zett anorg Chem , 1897, 14, 66, Lituds Umv Arsslnjt, 1913, 9, 18 

178 Chemistiy and Atomic Structure 

In 1919, Langmuir 1 adopted van den Brock's hypothesis 
as to atomic numbers and suggested that Rydberg's series 
expressed the number of elections in various successive 
shells about the nuclei of atoms, and on this basis elaboiated 
a comprehensive theory of atomic structure. The hydro- 
gen atom was assigned a structuie of a nucleus and one 
election, helium a shell of two electrons, the elements from 
lithium to fluorine a partly completed additional shell of 
eight electrons, neon the completed additional shell of 
eight electrons, the elements fiom sodium to chlorine a 
partly completed further additional shell of eight electrons, 
argon the completed additional shell of eight, the elements 
from potassium to bromine an additional partly completed 
shell of eighteen electrons, completed in krypton, the 
elements from rubidium to iodine a further partly com- 
pleted shell of eighteen electrons, completed in xenon, the 
elements from caesium to bismuth a furthei partly com- 
pleted shell of thuty-two electrons, completed in niton, 
and the elements from radium to uranium a further partly 
completed shell of thiity-two electrons. 

Langmuir postulated that the radii of the shells were 
proportional to the natural numbers I, 2, 3, and 4, and the 
areas of the shells to the squared natural numbers as in 
Rydberg's series. This involved the two outer shells of 
argon having the same radius and 16 elections, which were 
arranged in 8 pairs as a cubic structure with two electrons 
at each corner. The two outer shells of xenon similarly 
had the same radius and 36 electrons in 1 8 pairs. 

It was furthei postulated that the electrons in incomplete 
shells were valency electrons, by loss of or addition to 
which chemical combination occurred, with another atom, 
acquired electrons completing a shell. As the maximum 
known valency is eight, Langmuir was driven to assume a 
natural tendency of atoms to form only shells of eight 
electrons called octets. This further involved assigning to 
many atoms special forms of assumed inertness like the 

1 J Amer Chem. Soc , 1919, 41, 868 and 1543 

Atomic Structure and the Periodic Classification 179 

inert gases, though no such inert foims aie known. The 
chemically active element hon, for example, had to be 
assumed as inert in order to account for the bivalency of 
nickel, which was also assigned an inert form to account 
for the bivalency of zinc. If all valencies among elements 
had to be accounted for in a similar way it would be neces- 
sary to assume inert forms foi more than half the known 
elements, and some very active elements, potassium for 
example, would require to possess inert foims corresponding 
to valencies of four or more different elements. 

Langmuir adopted Lewis's postulate of 1916 that 
chemical combination between atoms, giving lise to non- 
lonising compounds, is due to the sharing of two electrons 
per chemical bond so that each atom completed a shell of 
eight electrons by the sharing process It has been indi- 
cated in Chapter IX, that the postulate of two electrons 
per bond is not compatible with an octet stiucture for 
elements having valency gi eater than four nor for elements 
giving rise to compounds with a co-ordination number 
gi eater than four, and that this applies to all but a veiy 
few of the known elements. It was also indicated that the 
two-election bond is incapable of explaining many types 
of compounds containing hydrogen, boion, caibon, nitro- 
gen, oxygen, fluoiine, and chloiine. It may further be 
icmarked thai the two-electron bond, called by Langmuir 
a covalency, is incapable of explaining the non-ionising and 
ineit hexafluorides of sulphur, selenium, and tellurium, 
which contain on the covalency basis four more electrons 
than the next higher inert gases, and the same aigument 
applies to many of the compounds of the elements of 
groups V, VI, and VII, preceding the inert gases in the 
periodic classification. It may be stated that the postulate 
of covalency is inapplicable to the majority of the com- 
pounds dealt with in inorganic chemistiy, with the out- 
standing exceptions of some compounds of carbon, nitrogen, 
and oxygen, and that many of the compounds even of these 
elements possess in all probability not more than one 

180 Chemist) y and Atomic Sttucturs 

electron per chemical bond. It cannot even be stated 
unequivocally that the compounds of organic chemistry 
usually contain covalency bonds, and much more informa- 
tion must be accumulated before methods can be devised 
for determining with certainty the number of electrons in 
the bonds in organic compounds. 

The numerous conflicting theories of election chemical 
combination to-day alone indicate that the bases of all 
these theones is insecuie. There is Dr. Fluisheim's theory 
postulating the transmission of affinity or electron demand 
over a chain of atoms, and denying the existence of alternate 
polarity and the directing effect of polarity of atoms in 
substitution processes. There are Piofessor Fiy's and 
Vorlander's theories of alternating polarity. There are 
Professor Lapworth's and Professor Robinson's theories of 
induced alternate polarities and key-atoms in conjugated 
systems of double bonds, based also on the covalency and 
octet theories or modifications of them. There is Professoi 
Lewis's new theory of valency and covalency based on 
Bohi's dynamic atom. There is Professor Lowiy's theoiy 
of alternate polarity, ciossed polarity, and intramolecular 
lonisation or hali-polansed double bonds, founded on 
covalency, a modified octet theory, and partly on Bohr's 
dynamic atom. There is Dr. Sidgwick's non-polar link 
theory based on covalency and an extension of Bohr's 
dynamic theory to the combination of atoms. There is 
Sir J. J. Thomson's theory that a chemical bond may 
consist of one, two, or three shared electrons, and a host of 
minor theories based on covalency, the octet theory, and 
static or dynamic atoms. Professois Thorpe and Morgan 
both agree that the time is not yet ripe for the application 
of general electronic theones to organic chemistry. 

The failure among chemical theorists to come to agiee- 
ment about electron combination is to be attributed to a 
wide variety of causes, e.g., the incorrect assumption in 
Lewis's theory of invariable covalency combination, the 
inadequacy of Langmuir's octet theory, and the incorrect- 

Atomic Structure and the Periodic Classification 181 

ness of Bohi's theory of the detailed structures of electron 
subgroups. Befoie any generalised theoiy of chemical 
combination by electrons becomes possible, some general 
agieement must be ai rived at among chemists as to the 
type of atom, dynamic or static, to be utilised in chemistiy, 
as to the piecise stiucture of the electron subgroups of this 
atom, and as to the precise meaning to be attached to the 
terms polaiity, bond, and valence, each of which is used 
with completely different meanings by different theorists. 

In 1921, an amendment of Langmuir's theory of atomic 
structure was proposed by Bury, 1 who suggested that some 
of the difficulties in the theoiy weie due to too complete 
acceptance of Rydberg's seiies and could be removed by 
the further postulates that the maximum number of 
electrons that the outermost shell of an atom can accommo- 
date is eight, that more than eight elections can be present 
in an inner shell only when theie is an accumulation of 
electrons in the outeimost shell, that groups of eight 01 
eighteen can exist stably in a shell even when the shell 
can accommodate more than eight or eighteen, and that, 
where the number of electrons in a shell is increasing from 
eight to eighteen or from eighteen to thiity-two, the 
elements concerned can have variable structure and 
valency, i.e. are the transition elements of the periodic 
classification. By these postulates Bury was enabled to 
assign electronic structures to the heavier atoms in closer 
accordance with their chemical behaviour, particularly as 
regarded fixed and variable valency. He assumed that the 
first transition series began with titanium and ended with 
copper, the second at ruthenium ending with pa ladium, 
that the rare earth series of elements began with cerium 
and ended with lutecium (the then-unknown element 
celtium or hafnium of atomic number 72 being thus quad- 
rivalent unlike the rare eaith elements), and that the third 
transition series began with osmium and ended with platinum. 
The evidence on which Bury based his transition series 

1 J Amer Chem Soc ; 1921, 43, 1602 

182 Chemistry and Atomic Structure 

was however very incomplete ; lie assumed^that zirconium, 
columbium, molybdenum, and silver, were excluded from 
the second tiansition series, and tantalum, tungsten, and 
gold excluded fiom the third transition series, and thorium 
and uranium were not placed in a transition series, on the 
incorrect assumption that all these elements have only one 
structure and valency All of these elements, and the 
fourteen raie earth elements, are properly transition series 
elements, and, excluding some of the rare earth elements, 
all of them have variable valency and therefore variable 

In 1921, Sn J. J. Thomson 1 put forward a new theory 
of atomic structuie and chemical combination, largely 
based on his earlier theory of I9I4- 2 He endowed the 
election groups with stability, impossible in the ordinary 
classical theory, by postulating a law of force in atoms which 
changed from attraction to repulsion with approach to the 
nucleus By a method of trial and error, he obtained the 

Ne *( C \ 

~ U J 
7-2 \ r / 

in which F is the force in the atom between an electron 
charge e and a nucleus charge Ne at a distance r, and C is a 
constant being the distance at which the force changes 
from attraction to repulsion, and is of the order io~ 8 cm. 
(a hundred millionth of a centimetre). This law of force 
allows eight electrons at the corners of a hypothetical cube 
to be stably arranged round a nucleus of equal charge. It 
further allows of the addition of successive shells of eight 
electrons as the nuclear force increases, thus obtaining 
structures in which the outside shell never contains more 
than eight electrons, in conformity with the chemical 
limitation of valency to eight. In atoms of many shells 
it was assumed that inner shells could hold more than eight 

l Pbtl Mag, 1921, [6], 41, 510, The Electron in Chemistry. Philadelphia 
1923 J v ' 

8 pbtl Ma , '9H, [6], 27, 757 

Atomic Structure and the Periodic Classification 183 

electrons, so that different atoms of high atomic numbei 
might have the same valency but a different number of 
electrons in innei shells, thus accounting for transition 
series of elements with variable valency. 

The theory was applied also to chemical combination by 
loss or gam of electrons or by sharing of electrons so that 
the external shells of the reacting atoms all became groups 
of eight electrons. The evidence of positive ray analysis 
having proved that a bond of one electron was possible, 
Sir J. J. Thomson did not feel compelled to adopt the 
covalency postulate, and in his theory chemical bonds con- 
tain numbers of elections varying fiom one to four electrons 

Sir J. J. Thomson's theory is an admnable attempt to 
find a solution for the pioblem of atomic structure without 
having to sui render the classical theoiy of electricity and 
energy, but it fails to account in any detail for the varying 
lengths of the periodic gioups without special assumptions 
at each stage, and, being a theoiy of static electrons, like 
Langmuir's, it fails to account adequately for the facts of 
optical and X-ray spectroscopy, for a multitude of general 
and particular facts in chemistry, and even foi some of the 
facts elucidated by his own positive ray analysis ; he quotes 
the fact that the maximum chaige found is eight as con- 
firming his octet structuies and the known fact of maximum 
chemical octavalency, but it is not shown why this maximum 
charge should occur on the positive particle of mercury, 
known to have a maximum valency of two, whereas octa- 
valent osmium is the element that ought to have given this 
charged particle. 

In the same year, 1921, Bohr 1 extended his theory of 
the dynamic atom fiom the atoms of hydiogen and ionised 
helium to the elements generally, by allocating to his orbits 
of total quantum number the numbers of electrons dis- 
cernible in the various groups of the periodic system. The 
allocation closely follows the numbers in the shells of 

1 Fysisk Trdssknft, 1921, 19, 153 , The Theory of Spectra and Atomic Cons tit it- 
ttotij Cambridge, 1922 , Nature, 1923, 112, 29 

184 Chennshy and Atomic Structure 

Langmuii's atoms as amended by Bury. In this scheme 
the elementary atoms have orbits of quantum numbers I, 
2, 3, 4, 5, 6, and 7, as in the simple hydrogen atom, the 
maximum numbers in the successive oibits being 2, 8, 18, 
32, 18, and 8, up to the orbits of quantum number 6 To 
account for the inert gases argon, krypton, xenon, and niton, 
Bohr assumed that an outer gioup of eight electrons was a 
stable structure as in Buiy's modification of the Langmuir 
theoiy The following aie the structuies assigned to the 
inert gases : 


Quantum number -12 3 4 5 6 

Helium (2) - - 2 
Neon (10) - - 2 8 
Aigon (18) - - 2 8 8 
Krypton (36) - 2 8 18 8 

Xenon (54) - - 2 8 18 18 8 
Niton (86) - 2 8 18 32 18 8 

Fiom the foregoing it is seen that the 3 quanta gioup, 
absent in neon, inci eases to 8 electrons in passing to argon 
and then changes to 1 8 electrons, while the 4 quanta gioup, 
absent in aigon, increases to 8 electrons in krypton, and 
that a similar change in the 4 and 5 quanta groups occurs 
in passing from argon through krypton to xenon, whereas 
the 4 quanta group fuither increases from 18 to 32 electrons 
followed by increase in the 5 quanta gioup to 18 electrons 
and the creation of the 6 quanta gioup with 8 electrons. 
These changes have been precisely confirmed by the inter- 
pretation of the optical and X-ray spectra of the elements 
immediately following each ineit gas, and the reality of 
these groups and the appioximate positions at which 
change occuis cannot be held to be in doubt. According 
to Bohr the change of the 3 quanta gioup commences 
immediately aftei calcium (20), in which the energy of the 
4 quanta valency electrons is only very slightly greatei than 
that of the inner 3 quanta electrons, and that scandium (21) 
has the same valency, the additional election going to 

Atomic Sttuctute and the Periodic Classification 185 

increase the 3 quanta gioup fiom 8 to 9 electrons, and that 
successive elements continue to add electrons to the inner 
group till it has the full quota of 18 electrons as in the 
cuprous ion, the transition seiies thus consisting of 9 
elements from scandium to copper. 

Similarly the second transition series extends from 
yttrium (39) to silver (47). 

After barium (56) Bohr assumes that the next electron 
goes to the 5 instead of the 6 quanta Dibits, as in 
lanthanum (57), but that the next electron goes to the 4 
quanta orbits in cerium (58), and that the next 13 electrons 
also go to the 4 quanta orbits, thus forming the rare earth 
group of 14 elements from cerium to lutecium (71), in 
which the 4 quanta group of 32 elections is complete. 
Further addition of electrons then occurs in the 5 quanta 
orbits, from celtium (72) to gold (79), in which the 5 quanta 
oibit is complete with 18 electrons. Further addition of 
electrons then occurs in the 6 quanta gioup till 8 elections 
are added as in niton (86) Only three more chemical 
elements are known, ladium (88), thorium (90), and 
uranium (92), and Bohr assumes that these elements are in 
a periodic gioup of 32, which, like the previous group of 32, 
includes a long transition series commencing with the 
tnvalent element of atomic number 89 (actinium if admitted 
as a chemical element). 

The foregoing scheme gives a very complete explanation 
of the form of the periodic classification, and is supported 
at the critical points by veiy strong spectroscopic evidence, 
and in outline this scheme of atomic structure must be 
accepted as a reality. 

In order to account for the fact that the X-ray emission 
spectra provide foi two orbits of quantum number 2, three 
of quantum number 3, foui of quantum numbei 4, three of 
quantum number 5, and two of quantum number 6, Bohr 
divided the various quantum gioup electrons into sub- 
groups characterised by the azimuthal quantum numbers 
pioposed by Sommerfeld The question then arose as to 

1 86 Chemistry and Atomic Structure 

how many of the group electrons should be placed in each 
subgroup. Obviously the first quantum group from 
hydrogen (i) to helium (2), which consists of only one 
group of 2 electrons in helium, must supply the answer if 
any general law underlies the formation of subgroups. 
Bohr noticed that this group, with quantum number I, 
has only I gioup of 2 electrons or double the total quantum 
number. Bohr's guess consisted in the assumption that 
this constituted the subgroup law, particularly as it gave 
symmetrical arrangements to all the quantum groups, the 
i quantum group having I group of 2 x I = 2 electrons, 
the 2 quanta group 2 subgroups each of 2 x 2 = 4 elec- 
trons, the 3 quanta group 3 subgroups each of 2 x 3 = 6 
electrons, and the 4 quanta group 4 subgroups each of 
2x4 = 8 electrons. This law is that the maximum 
number of subgroups is equal to the total quantum number, 
and the maximum number of elections in a subgroup is 
equal to twice the total quantum number. 

It will be shown in the next chapter that Bohr's sub- 
groups do not fit chemical facts. It inevitably gives 
symmetrical arrangements because the total number of 
electrons in a group is twice the square of the total quantum 
number, i.e. Rydberg's series, the unsquared numbers of 
which are accidentally the same as Bohr's total quantum 
numbers. Tt is obvious that the square of a number is 
symmetrically divisible by the number. It is unfortunate 
that Bohr was misled by a numerical coincidence and the 
properties of numbers, because his subgroup scheme has 
been made the basis of theories of chemical combination 
which come near enough to agieeing with chemical facts 
to make them a valid expedient in chemistry, and have led 
only to a non-concordant interpretation of chemical piob- 
lems of atomic and molecular structure and the 
mechanism of electronic combination. 

In 1923, Sidgwick 1 put forward a theoiy of chemical 

1 J. Cbem. Soc , 1923, 123, 725 , Bnt Assoc' Rep, 1923 , Chem Ind. t 1923, 
42, 901 

Atomic Stiuctuie and the Periodic Classification 187 

combination based on Bohi's subgroup scheme and the 
covalency theory of chemical bonds in non-ionising com- 
pounds. He assumed that every chemical bond in every 
type of compound consists of two shared electrons, which 
go to complete in each atom two of the subgroups of 
Bohr's scheme, though no reason is advanced as to why 
only two completed subgroups confer stability on atomic 
structures, when there are three possible subgroups for 
3 quanta orbits, four subgroups for 4 quanta orbits, and 
only one subgroup for I quantum orbits. The criticism 
already advanced in earlier chapters, as to the geneial 
inapplicability of the covalency of two electrons per 
chemical bond applies with even greater force to theories 
based on Bohr's subgroup scheme than to those based on 
an octet scheme, for the outermost subgroups throughout 
the scheme never contain more than six electrons. Sidg- 
wick's theory is consequently frequently inapplicable to 
the whole of the elements pieceding the inert gases in 
which the non-ionising valency or the co-ordination number 
is greater than three. This includes almost the whole of 
the known elements. The aluminium atom in cryolite, 
AlF 6 Na 3 , foi example, must have four more electrons than 
the next higher inert gas. The covalency postulate, in 
fact, can be maintained only by excluding the inert gases 
from the atomic structures of maximum inertness and 
stability. A similai argument applies to all hydrogen com- 
pounds assigned a covalency of two, such compounds having 
in consequence two more electrons than the nearest inert 
gas helium, which, in common with the other inert gases, 
Sir J. J. Thomson's positive ray analysis has proved, is 
incapable of acquiring any additional electrons whatever. 

Though Sidgwick's theory is based on the incorrect 
subgroups of Bohr's scheme, and on the incorrect assump- 
tion of invariable covalency bonds, he has supplied the 
elucidating key of a simple rule for determining the probable 
outer structures of the majority of complex and co-ordin- 
ated compounds. This rule states that the maximum 

i88 Chemtttiy and Atomic Stiuctnre 

covalency 01 co-ordination number is equal to the maxi- 
mum possible number of elections in a valency subgroup, 
next to the largest completed group in the atom This 
rule, divorced from covalency, is a particulai case of the 
general law of uniform atomic plan, which was deduced in 
Chapter X from the phenomena of radioactivity and the 
relations between the periodic groups. This law may be 
stated in more precise terms, including the atomic pheno- 
mena summarised in the Bohi-Sommerfeld dynamic atom, 
that all atoms are constructed on the same uniform atomic 
plan, and that for every total quantum number in an atom, 
there is the same maximum number of electronic orbits of 
the same shape and distribution m space, each containing 
the same maximum number of electrons, shared or unshared 
with other atoms. This law has been the basis, in one or 
more of its particular applications, of the whole of the 
theories of atomic structure, and was dimly apprehended 
by Mendeleeff, more clearly apprehended by Rydberg, 
still more clearly by Rutherford, Sii J. J. Thomson and 
Langmuir, almost grasped by Bohr, and is the key to 
Sidgwick's rule of electionic chemical combination. 



BOHR'S theory of atomic structme is strictly a theory 
relating to single atoms, neutral or ionised, fai removed 
from the influence of other atoms. The fact that it is an 
interpretation of the periodic classification of the elements, 
largely based on the properties of atoms in combination, 
indicates that it must be valid foi atoms in combination, 
at least so fai as the broad outlines of the theory are con- 
cerned. The theoiy ceitainly necessitates that the general 
type of the structure of an atom is preserved even when 
the atom loses some of its elections, and thus acquires 
positive charges. This is confirmed by the spectroscopic 
law of periodic group displacement, that the optical 
spectrum of an element that has lost one election lesembles 
closely the spectrum of the element one less in atomic 
number, and that the spectrum of an element that has lost 
two electrons lesembles closely the spectrum of the element 
two less in atomic number. This is fuither confirmed by 
Moseley's spectioscopic law connecting the wave-length of 
X-radiation with atomic number and with quantum con- 
ditions of atoms, the various quantum conditions being 
unchanged in atoms in combination. It may therefore be 
accepted that the structure of an atom is not influenced in 
qualitative aspects and only slightly in quantitative aspects 
by chemical combination In general it may be assumed 
that chemical combination between atoms is powerless to 
alter moie than the external portions of the structure of 
atoms, and that this alteration occurs in or near the elec- 
trons of the structure which are known as valency electrons. 
One of the outstanding periodic propeities of atoms in 
general is their tendency to yield compounds which are 
alkaline, acid, or amphoterically neutral. An examination 
of the periodic system indicates that the property of 
alkalinity is intense for the univalent alkali metals and is 
less intense for the bivalent alkaline earth metals, and 

19 Chemistry and Atomic Structure 

suddenly disappear in the trivalent elements boron, 
aluminium, gallium, etc., and does not appear elsewheie 
in the periodic classification, increasing acidity being 
observed as the periods are traversed. 

It is further a noticeable property of the nivalent 
elements of group III to yield compounds containing two 
valences, co-ordinated or non-ionising, and one unco- 
oidmated or ionising valence, and that the piopeity of 
two co-oidmated valences is common to the whole of the 
quadrivalent elements of group IV, and persists to a con- 
sideiable extent in the elements of groups V, VI, and VII, 
and that this property is only observable when the elements 
are acting electiopositively, i.e. when oxidised or paitly 
deprived of elections. This may be interpreted as evidence 
that all elements containing more than two valency elec- 
trons have two elections more fiimly attached than other 
valency electrons, which, fuithci interpreted in terms of 
01 bits, indicates that two electrons in the outer structuie 
of atoms aie in quantum orbits the energies of which are 
difFeient from that of other outer electrons. This is 
diicctly confirmed by the well-known spectioscopic fact 
that the outer orbits of the aluminium atom with three 
valency electrons arc of two different energy types, and 
that the spectra of tiivalent elements generally bear con- 
siderable resemblance to the spectia of umvalent elements. 

Corresponding observations on the property of acidity 
of elements indicate that it is a maximum in group VII 
and diminishes only gradually in groups VI and V, and 
suddenly changes to perceptible basicity 01 at least ampho- 
tericity in gioup IV. It is further notable that nearly 
all of the elements of these gioups have a marked tendency 
to yield compounds having the co-ordination numbei 4, 
independently of the number of ions with which the 
co-ordination complex may be associated. This may be 
interpreted as evidence that the valency electrons in 
excess of four aie all equally feebly attached to the atoms, 
which, further interpreted in terms of orbits, indicates 

Atomic Sttucturg and Chemical Properties of Elements 191 

that the valency elections in excess of four are all in similar 
quantum orbits the energy of which is less than that of 
the first four electrons. 

The detailed chemical evidence, by which it can be 
shown that the first two valency electrons are differentiated 
in energy 01 firmness of binding from the third and fourth 
electrons, and that both of these pairs aie differentiated 
from the remaining valency electrons, is so vast that no more 
than a tithe of this evidence need be cited to prove the point. 

Cupric salts yield characteristically insoluble and stable 
di-pyridmo-compounds, gold the dyad symmetrical di- 
alkyl auric salts and thallium the similar di-alkyl thallic 
salts. The reactivity of the magnesium Grignard reagents 
is attributable to the unsymmetncal valencies. 

Dyad symmetry is also evident in the chain structure of 
the valencies of such atoms as boron in the hydrides, carbon 
in oiganic compounds, nitrogen in polyazo-compounds, 
silicon in the silico-hydrocarbons, sulphur in polythionates 
and polysulphides, selenium m the seleno-dithionates 1 
and, fuither, in such compounds as carbon monoxide, 
ketones, aldehydes, and other carbonyl derivatives, also m 
gm-dialkyl groups attached to carbon, nitrogen, phosphorus, 
arsenic, antimony, tin, and tellurium atoms. Boron 
acetylacetone fluoride has the simple formula, BF 2 Ac, not 
the complex formula [BFJ [BAc 2 ], 2 and only one of the 
three chlorine atoms in aluminium chloride is displaced in 
the Fnedel-Craft's reactions. 

Chlorates, bromates, and iodates, are readily foimed and 
furnish perchlorates, perbiomates, and penodates only with 
difficulty. Nitrogen and arsenic yield no pentachlondes, and 
phosphorus and arsenic pentafluorides readily yield free fluo- 
rine on heating. Selenic, arseni, perbroniic, and bismuthic 
compounds are not readily formed by ordinary oxidations. 

The foiegoing relate entirely to the " odd " periodic 
series. The tnvalent elements of the " even " series, 

1 Morgan and Drew, J Cbem Soc., 1921, 117, 1456, Morgan and Mam 
Smith, y Chem Soc , 1921, 119, 1066, 

Morgan and Tunstall, y Cbem Soc , 1924, 125, 1963 

192 Chemist) y and Atomic Sttuctute 

scandium, yttrium, lanthanum, and the " rare earths " 
have all three valency electrons equivalent, due to the fact 
that the third electron is m a group of lower quantum 
number and possesses the same energy. This equal film- 
ness of binding for different quantum orbits has been 
directly deduced from X-ray spectra. 

It may therefore be regarded as certain that the first 
three subgroups of any total quantum number consist of 
2, 2, and 4 electrons. 

If the types of quantum orbits persist in atoms with 
increase in atomic number, it must be concluded that the 
valency subgroups discernible in the atoms from lithium 
to neon, persist in the succeeding 3 quantum group also 
of 8 electrons, and that this grouping of 2, 2, 4 elections is 
pieserved in the inner groups of the atoms from sodium to 
argon, which certainly possess this grouping in their outer 
valency electrons. 

By similar arguments the succeeding elements also possess 
these groups of 2, 2, 4 electrons, and it is known from 
Millikan's x work on the spectia of multiply-ionised atoms 
that the helium structure of 2 electrons is preserved in the 
elements from lithium to sodium, and that the neon 
structure is preserved in the elements from sodium to 
argon. This has 1 recently been fully confiimed by Dauvil- 
lier's z work on the widths of the X-ray spectral absorption 
bands of heavy metals for the L 01 2 quanta spectral senes. 
Dauvillier found that the relative widths of the bands, 
which are accepted to be due to the number of energy- 
absorbing electrons in the atom, are in the ratio of I : I : 2, 
and as there are 8 L elections, the numbers in each sub- 
gioup must be 2 2 : 4, as indicated by the foiegoing 
chemical evidence. 

This evidence shows conclusively that Bohi's subgroup 
scheme, of two subgroups of 4 electrons for the 2 quanta 
group, cannot be maintained. His postulate of three sub- 

1 Faraday Lecture, J Chem Sue , 1924, 125, 1405 
~ Compt rend, 1924, 178, 476 

Atomic Structure and. Chemical Properties of Elements 193 

groups for the 3 quanta electrons also cannot be maintained 
for the absorption spectral evidence indicates that there 
are five subgroups in the quantum group. The same 
evidence also indicates that the postulate of four subgroups 
for the 4 quanta electrons can not be maintained, for 
seven subgroups are discernible in the spectra. Bohr's 
rules that the number of subgroups is equal to the quantum 
number and that the number of electrons in a subgroup is 
equal to twice the quantum number are therefore both 
incorrect. It is remarkable, when Bohr devised his sub- 
group scheme and characterised the subgroups by total and 
azimuthal quantum numbers, that he did not discern 
something of the meaning of the variable factor in the 
quantum scheme, i.e. the increase in the azimuthal quantum 
number in the subgroups, and that he failed to realise the 
importance of the evidence of absoiption spectia in his 
scheme of atomic structure. 

The complete scheme of notation for the X-ray spectra 
is ija, 2jb, 2^, 2 2 a ; 3 x b, 3^1, 3 2 a, 3 2 b, 3 3 b , ^b, 4^, 4 2 a, 
4 2 b, 4 3 b, 4 3 a, and 4 4 a. The notation used in Bohi's atomic 
structures is ^ ; 2 13 2 2 ; 3^ 3 2 , 3s 5 4i 4a> 4s> and 4 4 . It 
will be obseived that the former scheme is derived from 
the latter by dividing all the orbits coi responding to ellipses 
into two types a and b, or, in other words, in subdividing 
into two all orbits having radial quantum number, the 
circular orbits, in which the total and azimuthal quantum 
numbers are equal, not being affected. It will also be 
observed that the resulting number of orbital types or sub- 
groups, for the quantum numbers I, 2, 3, and 4, is I, 3, 5* 
and 7, or one less than twice the total quantum number, 
i.e. the number of subgroups = 2n I. 

The foregoing chemical evidence indicates that the first 
quantum group is a single subgroup of 2 electrons, and 
that the second quantum group consists of 3 subgroups 
containing 2, 2, and 4 elections. It may now be observed 
that the number of subgroups is in fact equal to one less 
than twice the group number, in confoimity with the 

194 Chemisti y and Atomic Structure 

spectral evidence. The numbers 2, 2, 4 are further identi- 
fiable as twice the azimuthal quantum number, so that the 
rule for the number of electrons in a subgroup is 2k, where 
k is the azimuthal quantum number. This indicates that 
the electrons in the 3 quanta subgroups are, in order, 2, 2, 
4, 4, and 6, being five subgroups, and that the numbers 
in the 4 quanta orbits are 2, 2, 4, 4, 6, 6, and 8, or seven 
subgroups. The law of uniform atomic plan may there- 
fore be regarded as established, and has thus been deduced 
from the evidence of radioactivity, the evidence of the 
periodic classification, the evidence of general and specific 
chemical properties including valency and co-ordination, 
the evidence of the widths of X-ray absorption bands, the 
evidence of the wave-lengths of optical spectra, the evidence 
of the wave-lengths of X-ray emission spectra, and the 
evidence of the wave-lengths corresponding to the heads 
of the bands of X-ray absorption spectra. It may further 
be remarked that the evidence of the total intensity of the 
lines of X-ray emission spectra is to the same effect ; in 
the L series the total intensity of the lines due to circular 
orbits is 24, and the average intensity of the two senes of 
lines due to elliptic orbits is about II, a ratio of approxi- 
mately 4 to 2 ; the total intensity of the lines due to 
circular orbits in the M series is 10, and the average intensity 
of the lines due to elliptic orbits is 4, a ratio of 2f to I, 
whereas the ratio of the electrons in circular to elliptic 
orbits is actually 2 to I. 

The establishment of the law of uniform atomic plan on 
the foregoing consideiations renders it unnecessary to cite 
the further chemical evidence by which it can be shown 
that the electrons in the 3 quanta group actually are dis- 
posed in subgroups of 2, 2, 4, 4, and 6, but it may be re- 
marked that the law was deduced from this evidence before 
the rule of doubling Sommerfeld's azimuthal quantum num- 
ber had been appreciated, and indicates how complete is the 
chemical evidence as to atomic structure, despite the 
usual dictum that chemistry has concern only with 

Atomic Structure and Chemical Properties of Elements 195 

the peripheral properties of atoms, i.e. with their valency 

The chemical evidence, as to the precise points in the 
periodic classification at which added electrons go into 
interior groups and into what subgroups, and as to the 
numbers of electrons entering the various subgroups, is 
not quite so precise, but it indicates that the first transition 
series commences with scandium as in Bohr's scheme ; 
that the second transition group commences at yttrium ; 
and the third at lanthanum. It further indicates that the 
fourth 3 quanta subgroup commences to be filled at man- 
ganese ; that the fourth 4 quanta group commences to be 
filled at ruthenium, and the sixth at europium ; that 
the fifth and fourth 5 quanta groups commence to be filled 
at celtium and osmium respectively, and that the fourth 6 
quanta group commences to be filled at actinium. 

The structures of the vaiious atoms at the several critical 
points in the periodic and quantic classifications aie given 
in the subjoined tables, and these may be regarded as 
expressing with considerable accuracy the atomic structuies 
of the whole of the known elements as deduced laigely 
from the idiosyncratic properties of atoms ascertained in 


Chemistry and Atomic Structure 



Total quantum num- 
ber = 



number k= 




























and Sc 3 (21) - 




Cu 1 








and Y 3 (39) - 





Ag 1 









and La 3 (57) - 






Lu 3 








Au 1 


























Total quantum num- 

= i 


Azimuthal quantum 


k= i 


112 23 



23 23 

23 23 

Valency of 





4 5 

6 7 

Sc (21) - 





Tl (22) - 







v (23) - 






01 OO 

Cr (24) - 








Mn (25) - 




r 4 




Fe (26) - 




2 4 



Co (27) - 







Ni (28) - 






Cu (29) - 



224 46 



Atomic Stnuture and Chemical Properties of Elements 197 



Total quantum 

number = i 2 3 

Azimuthal quan- 
tum num- 
ber k= i ii2 11223 

Valency of ion 

23 23 23 23 23 23 23 23 

Y (39) - - 2 

224 22446 224 


Zr (40) - - 2 

224 22446 224 

02 01 00 

Cb (41) - - 2 

224 22446 224 

03 02 01 


Mo (42) - - 2 

224 22446 224 

04 03 02 

01 00 

(A i\ 



Ru (44) - - 2 

224 22446 224 

24 23 22 

20 10 oo 

Rh (45) - - 2 

224 22446 224 

25 24 23 


Pd (46) - - 2 

224 22446 224 

26 24 

Ag (47) - - 2 

224 22446 224 

46 45 




Total quantum num- 

ber = 

I 2 3 



Azimuthal quantum 

number k= 

I 112 II223 

11223 34 

34 34 


Valency of ion 


3 4 

La (57) - - 

2 224 22446 




Ce (58) - - 

2 224 22446 




Pr (59) " - 

2 224 22446 


O2 01 


Nd (60) - 

2 224 22 j-46 




- (61) - - - 


Si (62) - 

2 224 22446 

22446 06 



Eu (63) - - 

2 224 22446 

22446 i 6 



Gd (64) - 

2 224 22446 




Tb (65) - - 

2 224 22446 


17 16 


Dy (66) - 

2 224 22446 




Ho (67) - 

2 224 22446 




Er (68) - 

2 224 22446 


38 - 


Tm (69) - 

2 224 22446 


48 - 


Yb (70) - - 

2 224 22446 


58 - 


Lu (71) - - 

2 224 22446 





Chemistry and Atomic Structure 


Total quantum 

number = 123 4 

number k== i 112 11223 II22 334 

112 23 23 23 23 23 23 23 23 
1234. <; 6 7 8 



- 2 








j - / 



- 2 











- 2 












- 2 








20 oo 



- 2 











- 2 






2 4 




- 2 





4 6- 




BOTH the classical theory and the theory of relativity postulate that 
continuous increase m mass accompanies continuous increase in the 
velocity with which the mass moves. Any portion of matter, 
therefore, decreasing its distance from an attracting force and 
consequently increasing its velocity, must undergo increase of mass. 
If the attracting force is due to electric charges of fixed quantity, 
the force causing motion will not depend upon the mass to be 
moved. If this mass has increased the unchanging charges cannot 
move the larger mass with the same velocity, the velocity of the 
new mass consequently falls off, and the movement of the mass is 
retarded. If the mass were originally travelling in an orbit subject 
to a constant central force, increase of mass must be accompanied 
by increase in the radius of the orbit. Conversely, decrease in mass 
must be accompanied by decrease in the radius of the orbit. The 
velocity of a body in an elliptic orbit, with the centre of attraction 
at one focus, must increase as the body moves from aphelion further 
away to perihelion nearer the nucleus, and in consequence the mass 
of the body must increase with approach to perihelion and decrease 
with approach to aphelion. The increase of mass must involve 
increase of orbital radius and the decrease of mass involve decrease 
of orbital radius. The real path of a massive body in an initially 
elliptic orbit thus lies further away than a true ellipse in approach- 
ing, and nearer in receding from the focus. In Diagram XVII, the 


elliptic orbit of a rnassless particle is shown dotted, the full line 
representing the actual path of a massive particle. It will be seen 
that the original major axis, PQ, of the orbit is turned through the 
angle, SFQ, in one revolution, and that the rotation of the axis is 

2 o Cbemistiy and Atomic Sttuctute 

in the same clock-wise direction as the circulation in the orbit 
Diagram XVIII shows successive positions of the axis with successive 
revolutions in the orbit, the aphelion describing a circle of radius 
equal to the aphelion distance from the focus, and the perihelion a 


smaller ciicle of radius equal to the perihelion distance from the 
focus. The actual orbit described by the particle is that due to 
the confocal superposition of a circular motion on an elliptic motion 
in the same plane. 

Sommerfeld's postulate that an electron circulating about a 
nucelus can describe elliptic as well as circular orbits, involved the 
extension of velocity-mass-effects to electrons in atoms, and the 
foregoing diagrams may be regarded as expressing, on an exaggerated 
scale, the effect in elliptic Dibits, ie orbits having both radial and 
azimuthal quantum numbers He predicted that the more eccen- 
tric orbits (having smallest azimuthal quantum numbers) would 
exhibit the greatest mass increase, and thus possess greatest energy, 
and that, consequently, every possible elliptic orbit would give rise 
to a separate line m spectra, both optical and X-ray. This pre- 
diction has been abundantly confirmed m the " fine-structure " of 
spectral lines, and electrons with mass varying with velocity must 
therefore be accepted as realities in atomic structure. 

The circular motion of the whole elliptic orbit about the focus 
is described as a precession motion in the plane of the orbit, and is 

Appendix 20 1 

Sommerfeld's " relativity effect." It is obvious that the foregoing 
considerations as to planar precession can apply only to an electron 
completely unperturbed in its motion by other forces. The intro- 
duction of another electron into the system must perturb the pre- 
cession motion profoundly. Several electrons about the focus must 
still further perturb the precession, and if all the electrons are 
similar in type and undergo similar precession, the mutual inter- 
penetration will render normal plane precession impossible ; yet 
precession must occur, must occur symmetrically and to the same 
extent for all similar electron orbits, and must not be materially 
altered by the addition of further electrons in other precessing 
orbits of different type This can be achieved, as indicated on p. 174, 
Chap XII, if the plane, in which precession occurs, rotates about 
the fixed major axis of the elliptic orbit, the electron thus describing 
spiral paths between a fixed perihelion and aphelion on the surface 
of an ellipsoid of revolution, each electron in an atom of whatever 
total or radial quantum number thus appropriating a small ellip- 
soidal domain around a radius of a sphere with the nucleus as centre. 

The spatial " relativity effect " above described consists of the 
superposition of a circular motion on an elliptic motion not in the 
same plane. It differs from Sommerfeld's relativity effect only in 
the plane of superposition. It extends the plane conception of 
simple atoms to the spatial conception of complex atoms, ]ust as 
modern chemistry is a spatial interpretation of the old chemistry in 
a plane 

Electron orbits in free atoms are " borgne," " one-eyed," have a 
nucleus only at one focus of the orbit. Atoms, however, which enter 
into chemical combination and remain indissolubly attached to one 
another, as in non-ionisable compounds and co-ordination com- 
pounds generally, are in another category their electron orbits are 
" two-eyed," and, according to the most recent views, some of the 
electrons, valency electrons, circulate round both atomic nuclei in 
orbits common to both atoms. Obviously as such orbits have two 
foci of attracting nuclear charges, electrons cannot describe elliptic 
01 bits nor yet circular orbits, both of which are characteristic of 
orbits with a single central force Professor G. T. Morgan came 
to the conclusion some years ago that the closed curves, known to 
mathematicians as " ovals of Cassini," which are " two-eyed," might 
suffice to represent the orbits of shared electrons The Cassinian 
system of curves is such that the product of the distances from the 

2O2 Chemistty and Atomic Shucture 

foci of a point on a curve is constant. Clerk Maxwell, in his Elec- 
tricity and Magnetism, 1873, showed that the sections of equi- 
potential surfaces about two equal similar charges were Cassinian 
ovals It follows, therefore, that a Cassinian oval is an orbit of 
constant potential, and, consequently, a possible Bohr orbit for an 
electron. Some of the Cassinian ovals are shown in Diagram XX, 
page 204, which represents a series of Bohr orbits of shared 
electrons. The outer ovals are nearly elliptic, the median curve, 
figure-of-eight, is known as the lemniscate of Bernouilli, and the 
inner separated curves approach circles as the foci are approached 

The figure-of-eight curve has been independently suggested by 
Professor Lowry, as representing cases of shared-electron orbits. In 
Pofessor Morgan's view the figure-of-eight or lemniscate curve is 
the limiting case of such electron-sharing, and represents an atom 
on the verge of becoming a free ion by loss or gain of the shared 
electron, the smallest alteration in force sufficing to send the electron 
into one or other of the " inner ovals " This view has been one 
of the mainsprings and guiding principles of his direction of the 
Birmingham research school in the last three or four years, and has 
been applied with success to the explanation of all easily-hydrolysable 
organic and inorganic compounds. The chief objection to the 
Cassinian system of orbits is that they are incompatible with relati- 
vistic change of mass with velocity. 

It was suggested in the Faraday Society's general discussion on 
the electronic theory of valency (1923), that the figure-of-eight curve 
allows for the " relativity effect," on the ground that " the outward 
loop cuts the preceding inward loop instead of being superposed upon it. 
Ibis is precisely what happens in the figure-of-eight" This state- 
ment is correct but misleading Many orbits have similar loop- 
cutting properties, but they are not in consequence able to accomniO' 
date a relativistic change of mass with change in velocity. 

The figure-of-eight curve is merely the lemniscate and a special 
case of Cassinian oval. The Cassinian system is derived from the 
superposition of two co-planar non-concentric circular motions, 
and thus is fundamentally different from Sommerfeld's relativity 
effect of a superposition of a circular motion on a confocal co-planar 
elliptic motion. The " relativity effect " necessitates the rotation 
of the orbital axis in the orbit plane about the focal centre. The 
Cassinian system cannot be rotated in the orbit plane about the 
focal centre because theie are two foci Rotation about either 



focus must rotate the whole other atom or the Cassinian system will 
be ruptured The actual path of an electron exhibiting relativity 
effect is indicated in Diagram XIX, the orbit being initially the 
figure-of-eight. After one revolution the electron leaves the 


Election Precession Otbit changing continuously from Figure-of-Eight to 
Rectilinear Vibration 



figure-of-eight and enters an outer oval, and after a few revolutions 
in a deforming and expanding figure-of-eight, finally travels tan- 
gentially to the median line MN, and continues to vibrate to and 
fro on this line indefinitely In general, it may be stated that an 
orbit allowing for relativistic change of mass with change of velocity 
cannot exist in the orbital plane, where the force causing motion is 
directed from more than one point This applies to two electrons 
about one nucleus, equally as to one electron about two nuclei. 

The difficulty is overcome by allowing precession to take place 
in a plane other than the orbital plane, as with a single atom with 
several electrons (see p. 174) The whole Cassinian system on 
revolution about the hoiizontal axis describes a series of confocal 
surfaces which are Cassinian ovals of revolution, the surfaces being 
in fact Clerk Maxwell's equipotential surfaces Any Cassinian oval 
of revolution must comprise all possible Bohr orbits, for the potential 


Chemistry and Atomie Structure 

is constant at all points. All such orbits, described spirally between 
the two fixed perihelion points, allow for " relativity effect." 

It may be pointed out that, strictly, the question of " relativity 
effect " is not applicable to Cassinian orbits, for they are derived 
from the superposition of two circular orbits, i.e. relate to the 
conjoint orbit of two circular orbits, to neither of which relativistic 


Oval Orbit about Two Atomic Nuclei at tie Fact of a Cassinian System 

change of mass applies The system, to which elliptic orbits con- 
form on superposition, is not the true Cassinian system, but the 
harmonic curves resulting from the superposition of two elliptic 
orbits. This system, which may be termed the elliptic Cassinian 
system, does not appear to have been investigated in detail by 
mathematicians. It is however the system of orbits common to 
two atoms in which electrons lotate that would have elliptic orbits 
in either atom separately. Spatial precession may be readily applied 
to such orbits, the motion of the electron being the superposition 
of a coaxial circular motion on two superposed non-confocal co- 
planar coaxial elliptic motions. Obviously a similar spatial circular 
precession can be applied to a circular orbit superposed on a non- 
confocal co-planar elliptic orbit, thus representing the orbit of an 

Appendix 205 

electron which in one atom separately would be circular and in the 
other elliptic. 

An electron travelling from an inner to an outer orbit of super- 
posed simple harmonic motions of the foregoing types, will give rise 
to light radiation in which these superposed motions will be com- 
ponents in the light vibration, and will give rise to spectral lines 
according to the nature of the composite initial and final types of 
orbital motion, thus giving a clue to the meaning of Bohr's " corre- 
spondence principle," in which the various types of emitted radiation 
are likened to the overtones and harmonics of the musical scale, the 
frequency of every spectral line, even those produced by external 
forces as in the Stark and Zeeman effects, being related to a corre- 
sponding vibration frequency in orbital motion of electrons. 


THE following is a list of the more important works and original 
papers in English, published in recent years, on atomic structure, 
chemical co-ordination, and valency direction. 


ANDRADE, E. N. DA C. The Structure of the Atom. 1923. (Bell). 

ASTON, F. W. Isotopes. 1922. (Arnold). 

BOHR, N. The Theory of Spectra and Atomic Constitution. 1922. 

(Cambridge University Press). 

BORN, M. The Constitution of Matter. 1923. (Methuen). 
BRAGG, SIR W. H., and W. L. X -Rays and Crystal Structure. 1923. 


CRANSTON, J. A. The Structure of Matter. 1924 (Blackie). 
FAJANS, K. Radioactivity. 1923 (Methuen). 
GRAETZ, L. Recent Developments in Atomic Theory. 1923. 

KRAMERS, H. A., and HOLST, H. The Atom and the Bohr Theory of 

its Structure 1923. (Gyldendal). 
LEWIS, G. N. Valence and the Structure of Atoms and Molecules. 

1923. (Chem. Catalog Co ) 
LODGE, SIR OLIVER J. Atoms and Rays. 1924. (Benn) Electrons. 

1906. (Bell). 

LORING, F. H. Atomic Theories. 1923. (Methuen) 
MIALL, S. The Structure of the Atom. 1922. (Benn). 
RUSSELL, B. The A B C of Atoms. 1923. (Kegan Paul). 
RUTHERFORD, SIR E Radioactive Substances and their Radiations. 

1913. (Cambridge Univ Press). 

SODDY, F. The Chemistry of the Radio-elements. 1914 (Long- 
SOMMERFELD, A. Atomic Structure and Spectral Lines 1923. 


STOCK, A. The Structure of Atoms 1923. (Methuen). 
SULLIVAN, J. W. N Atoms and Electrons. 1923, (Hodder and 

THOMSON, SIR J. J. The Electron in Chemistry. 1913. (Franklin 

Inst. Press). Rays of Positive Electricity 1921. (Longmans). 
WYCKOFF, R. W. G. The Structure of Crystals. 1923 (Chem. 

Catalog Co). 

(Continued on page 208) 

208 Chemistry and Atomic Structure 

BIBLIOGRAPHY (continued} 



BARKER, T. V. J Clem. Soc., 1914/1924. 

BRIGGS, S. H. C. J. Cbem Soc., and Phil. Mag , 1908/1921. 

DUFF, J. C. J. Chem. Soc , 1920/1924 

KING, H. J S. J. Chem. Soc., 1924, 125, 1329. 

LOWRY, T. M., and LOWRY AND BURGESS. Chem. Ind , J. Chem. 
Soc , Phil. Mag., and Trans. Faraday Soc , 1923/1924. 

J. Soc. Dyers and Colounsts, J. Soc. Chem Ind., and Proc Roy 
Soc,, 1906/1924. 

MORGAN, H. H. J. Chem. Soc , 1923, 123, 2901 
PICKARD, R. H. and KENYON, J. J Chem. Soc., 1906/1907 
PRICE, T. S., BRAZIER, and DUFF. J. Chem. Soc , 1913/1920. 

and SIIIBATA (K.). J C0M Set., Tokyo, 1916/1920. 
SIDGWICK, N V. Chem. Ind., J Chem. Soc., and Trans. Faraday 

Soc., 1923/1924. 
SMITH, J. D MAIN. Chem. Ind., Chem News, and J Chem. Soc., 

THOMAS, W. Complex Salts 1924 (Blackie) J. Chem. Soc, 


WARK, I. W. J. Chem. Soc., 1923, 123, 1815, 1826 
WERNER, ALFRED. New Ideas on Inorganic Chemistry. 1911 


(Continued on page 209) 

Bibhogi apby 209 

BIBLIOGRAPHY (continued} 
Valency Direction 

ARMSTRONG, SIR H E , and RODD, E. H Ptoc Roy Soc , 1912, A 
87, 204 

BARLOW, W J Chem. Soc., 1906/1916 

BRADY, O L , and DUNN, F P J Chem Soc , 1916/1924 

KENNER, J. J. Chem Soc, 1923/1924 

MILLS, W. H. J Chem. Soc , 1910/1924. 

MOIR, J J Chem Soc , 1922/1924. 

POPE, SIR W. J J Chem Soc., 1899/1916 

THORPE, J. F., and INGOLD, C K J Chem. Soc, 1915/1924 

Chemistty in the Twentieth Century. Committee of Scientific 

Societies 1924 (Benn). 
General Discussion on the Electronic Theoiy of Valency 1923 

(Faraday Society). 


OC-RAYS, see Helium nuclei 

ABEGG, 122 


192, 193 

ACETYLACETONES, 95, 115, 191 
ACIDITY of bases, 31 
ACIDS, see basicity 
ACKERMAN, 1 , 208 
ACTINIUM, a-iaysj 129 

emanation, 129, 136 

radioactivity, 127 

series atomic weights., 152 
ACTINON, see Actinium eman.ilion 
AFFINITY, see Chemical affinity, 
and Residual affinity 

elective, 42 

transmission, see Flursheim 
ALPHA-RAYS, see Helium nuclei 
ALUMINIUM, disintegration, 158 

fluoiide, 81, 187 

spectrum, 190 
AMMINES, Blomstrand, Co 

Gibbs and Genth, 60 

Hofmann, 59 

Jorgensen, 60 

Werner, 58, 86, 90, 92 
AMMONIUM theory, 58, 63, 90, 



ANDRADE, E. N. da C , 207 
ANION, 50 
ANODE, 50 
ARMSTRONG, Sir H E , 209 
ARRHENIUS, S , 49, 53, 57 
ARSENIC, optically active, 46, 98 
ARTIFICIAL disintegration, see 

ASADO, J , 208 
ASTBURY, W. T., 97, 208 
ASTON, F. W, 155, 157, 207 
ASYMMETRIC atoms, 102 
ATOM, chemical, chapters i and 2 

see Cubic 

electrical, 50, 53, 119, 143 

indivisible, 15, 27, 28 

see Isotopes, Octahedral, Rutherford, 

ATOMIC, asymmetry, 102 
combination, 34, 189 
heat, 25 
mattei, 15 
nucleus, sec Atomic structuie, 


number, 70, 146, 148, 177 
orbits, see Orbits 
plan, see Law of Uniform 
shape, see Kekule", Le Bel, Pl,ao, 

and Hoff 
structuie, see Bohr, Langmuir, 

Rutherfoid, Thomson (Sir J J ) 

inert gases, 136, 138, 139, 178, 

184, 196 
intranuclear, 139, 140, 155, 156, 

^8, iS9 
law, see Law of Uniform Atomic 

nucleus, 139, 140, 141, 146, 155, 

^ iS7> 158, 159 
ladioactive elements, 136, 139 
valency, 122, 135, 154, 178, 179, 

182, 191, 192 
volumes, 26 
weight, 28 et seq , and see Isobars, 

Isotopes, and Moseley 
ATOMICITY, atoms, 29, 41, 42, 

Gerhardt's, 41, 80, 118 
Kekuld's, 42, 57, 60, 80 
molecules, 29 
radicals, 41, 42 
AVOGADRO, electrochemical theory 


hypothesis, 24, 26, 42 

oxygemcity theory, 48 

AZIMUTHAL, angle, 164 

quantum number, 164, 166, 169, 
193, 196, 200 

(3-RAYS, see Electron 
BACON, Francis, 16 
BARKER, T. V , 208 
BARLOW, W , 209 
BASICITY, acids, 30, 38, 43 

Gerhardt, 41 

Liebig, 38 

radicals, 41 

Wurtz, 41 


Chemist) y and Atomic Stnicture 

BALMER, 147, 164 
BASIC salts, 89, 113 
BAYLEY, 73, 74, 77 
BEL, see Le Bel 
BENZOYL radical, 36 
BERZELIUS, atomic weights, 26 

dualistic theory, 48, 66, 80 

electrolysis, 47 

isomensm, 35, 57 

molecular weights, 38 

nomenclature, 21 

Proust's law, 19 

Richter's law, 18 

substitution theory, 37 
BETA-RAYS, see Electron 
BINUCLEAR complexes, 99, in 

orbits, see Orbits 
BOHR, Niels, see Orbits 

chemical bonds, 121 

correspondence principle, 205 

helium atom, 147, 163, 183 

hydrogen atom, 147, 163, 183 

law of atomic plan, 186 

Mendele'effian transition periods, 
72, 185, 195 

periodic table, 77 

Planck's quantum theory, 145 147 

structure of atoms, 163, 165, 175, 
183, 189 

subgroup scheme, 181, 186, 187, 
192, 193 

X-ray spectra, 149, 165, 170, 175, 

185, 193, 194 

BOILING-POINT elevation, 31, 53 
BOLTWOOD, B B, 129, 130 
BONDS, see Valency electron 

Frankland's, 43, 57 

valency, 44, 45, 52, 119 et seq 
BORN, M, 207 
BORON hydrides, 124 
BOWEN, A R , 208 
BOYLE, Robert, 16 
BRADY, O. L , 209 
BRAGG, Sir W. H., 207, 208 

BRAGG, W L., 207 
BRAZIER, S. A , 208 
BRIGGS, S H. C , 208 
BROEK, van den, 70, 146, 147 
BURGESS, H , 208 
BURY, C R., 181, 184 
BUTLEROFF, 43, 60 

CANAL rays, see Positive rays 
CANNIZZARO, atomicity (valency), 
42, 57, 80 

atomic weights, 27, 118 

Avogadro's hypothesis, 27, 41, 118 

electrochemistry, 42 

indivisible atoms, 27 
CARBON, tctrahedial atom, 98 
CARTESIAN philosophy, 16 
CASSINIAN ovals, 201 et seq, 
CASTELL, R A S , 208 

rays, 144, 145, 147 
CAVENDISH, 18, 19, tig 
CHADWICK, J., 157 
CHANCOURTOIS, de, 69, 75 
CHELATE groups, 91, 98, 103 115 
CHEMICAL affinity, 82 

classification, 76 
CIS-ISOMERS, 102, 104, 106, no 
CLASSICAL theory, 143, 144, 147, 

160, 161, 183 

CLASSIFICATION of elemento, see 

analytical, 68 

Bayley, 73 

Berzehus, 66 

Chancourtois, 69 

chemical, 76 

Cooke, 68 

Doberemer, 67 

Dumas, 67 

Faraday, 68 

Gmelin, 68 

Lavoisier, 66 

Mendele'eff, 70 

Meyer, Lothar, 73 

Newlands, 69 

Odhng, 68, 70 



Classification of Elements con 
Pettenkofer, 68 
Stoney, Johnstone, 75 
Thomsen, 76 
CLAUSIUS, 51, 119 

CLERK, MAXWELL see Maxwell 
COBALT, optically active, 107, 114 
COMBINING capacity, see Valency 

weights, see Equivalent weights 
COMMITTEE of Scientific Societies, 


CONJUNCT, see Copulation 
CONSTITUTIONAL formula?, 43 
CONTINUOUS matter theory, 15, 


COOKE, 68 

CO-ORDINATION, see Cubic, Octa- 
hedral, Tetrahedral, Wernei 
evidence, 112, 114, 115, 117 
number, Si, 83, 87, 95 et seq 
types, 88, 98, 103, 115 
COPPER ammo-acetate, 92 
cadmibromide, 114 
chloride, basic, 92 
COPULATION, Berzehus, 37 
Frankland, 40 
Gerhardt, 37 
Gibbs and Genth, 60 
COREY, F J, 208 
CORPUSCLE, see Electron 
CORRESPONDENCE principle, see 

COUPER, atomicity (valency), 42, 


atom linking, 43 
elective affinity, 42 
COVALENCY, 124, 125, 179 et seq , 


CRANSTON, J A , 207 
CROOKES, 144, 1 52 
CRUM BROWN, see Brown 
CUBIC atoms, 16, 56 

co-ordination, 115, 116 
CURIE, Marya, 126, 127 

Pierre, 126, 127 
CYANOGEN radical, 34 
CYCLIC, sec Chelate 
formula, benzene, 61 
structures, carbon, 63 
co-ordinated, 91 

DALTON, atomic theory, 19 et seq , 

23, 118 

Avogadro's hypothesis, 24 
butylene discovered, 34 
Gay-Lussac's law, 24 
law of simple multiple proportions 

by weight, 19, 20, 27, 118 
DAVY, 38, 47, 48, 49 
DEBIERNE, A, 127, 129 

DEMOCRITUS, 15, 1 6 
DEMPSTER, A J , 156 
DIAZONIUM ibomensm, see Iso- 

mensm, do\ible-bonds 
DIKETONES, 91, 9$, 115 
DISINTEGRATION, artificial, 67, 

157 et seq 
helium nuclei, 158 
light atoms, 157 
radioactive, 128, 129 
DISPLACEMENT, law of periodic 

radioactive, see Law of radioactive 

spectroscopic, see Law of spectral 

DISSOCIATION, ionic, tee lonisa- 


DORN, E, 128 

DOUBLE bonds, 45, see Isomensm 
decompositions, 51 
valencies, electrical, 58 
DREW, H D. K, 95, 191, 208 
DUALISTIC theory, see Bcrzelius 

and Lavoisier 
DUFF, J C , 208 

DUMAS, Avogadro's hypothesis, 25 
classification of elements, 67 
Doberemer's triads, 67 
Prout's hypothesis, 66 

ethenn, 35 
radicals, 36 

substitutions, 36, 49, 57 
types, 37 


Cbemistty and Atomic Structure 

DUNN, F. P , 209 
DYNAMIC atom, 160 et seq 

in chemistry, 163, 175 

ELECTIVE affinity, see Couper 
ELECTRICAL double valencies, see 

Double valencies 
ELECTRINE, 52, 119, 143 
rhenms, Avogadro, Berzebus, 
Ciamician, Clausius, Davy, 
Faraday, Grotthus, Helmholtz, 
Maxwell, Ritter, Stoney, and 

ELECTROLYSIS, laws of, 49 

light, 143, 161 

ELECTRON, see Cathode rays, Elec- 
trine, Orbits, Positive rays, 
Subgroup, Valency (electron), 

p-rays, 128, 135 
combination, see Covalency and 

energy, 166 
Franklin, 142 
groups, see Quantum 
Lodge, 120 
light absorbing, 161 
mass, 145, 156, 165, 170, 173 
named, see Electnne 
radioactivity, 128, 132, 13:;, 138 
sharing, 120 et seq. 
weight, see Electron mass 
Werner's theory, 85 
ELLIPTIC orbits, see Orbits 

precession, see Orbits 
ELLIS, C. D., 156 
ELVINS, O C , 208 

EQUIPOTENTIAL surfaces, 202 
EQUIVALENT proportions, law of, 

see Richter 

weights, 21, 23, 30, 31, 42, 118 
ETHER, classical theory, 143, 160 
compression, 161, 162, 172, 173 
relativity effect, 172, 173 
structure, 161, 172, 173 


112, 113 
ETHYLENE isomensm, see Isomer- 

ism, double-bond 

FAJANS, K, 131, 207 
FARADAY, classification of element 1 !, 

discovery of benzene, 34 

dualism of Eerzelms, 49 

electrolysis, 48, 49, 119, 143 

isomensm, 34 

FARADAY Society, 202, 209 
FIGURE-of-eight orbit, see Oibits 
FINE structure, spectral lines, see 


FLECK, A, 130 
FLUORINE, artificial disintegration, 

+, IS8 
FLURSHEIM, B. J , 180 

FRANKLAND, E., atomicity, 42, 57 

bonds, 43, 57 

copulation, 40 

latent bonds, 57 

organo-metallic compounds, 39, 40 

valency, 40, 41, 42, 57, 58, 80 
FREEZING-POINT depression, 31, 


FRIEND, J. N , 57 
FRY, H. S , 1 80 
FUKAGAWA, K., 208 
FUMARIC acid, 102 

y (GAMMA) RAYS, 128, 155 


GAY-LUSSAC, composition of water, 


cyanogen radical, 34 
isomensm, 34, 57 
law of simple multiple proportions 

by volume, 24 
racemic acid, 35 
GEHRCKE, ,156 
GEIGER, H., 134, 146 
GENTH, 60 
GEOMETRICAL isomensm, see 

GERHARDT, atomicity (valency), 

41, 80, 118 
basicity, 38, 40 



Gerhard L con 

conjugation and copulation, 37, 38 
molecular constitution, 43 
theory of residues, 36, 40 
simple types, 41 

GIBBS, 60 

GMELIN, 21, 40, 68 


GRAETZ, L , 207 


GRAY, see Whytlaw-Gray 

GREEK atomism, 15, 16 


GROUP displacement law, see Law, 
radioactive, and Law, spectral 

HEAT radiation, 145 
HELIUM, atomic charge, 154 

nuclei, 128, 129, 131, 132, 134, 139, 
141, 148, 155, 157, 158 

structure, 137, 139, I5 4, 178, 184, 
186, 187, 196 

velocity of a-particle, 157 
HELMHOLTZ, 52, 119, 143 
HIGGINS, Bryan, 17 

William, 17, 19 
HINDU atomism, 15 
HOFF, van'tj asymmetric atoms, 102 

atomic shape, 56, 61, 62 

geometrical isomerism, 62 

molecular weights, 53 

osmotic pressure, 53 

solution theory, 53 

tetrahedral carbon atom, 56, 61, 95 

valency, 62 
HOFMANN, 44, 39 
HOLMES, E , 208 
HYDROGEN, see Spectra 

atomic weight, see Isotopes 

atoms, number in i gr , 30 

charge on molecule, 154, 187 

nucleus, 153, 154, 158 

proton, 139, 155, 158 

tnatomic, 124, 153, 154 
HYDROLYSIS, 87, 123 
HYDRAZONE isomerism, see Iso- 
meribm, double-bond 

INACTIVE elements, 76 

INDIVISIBLE atoms, see Atoms 
INERT gases, discovered, 76 
positive charge on atoms, 154, see 

Helium nuclei 

periodic classification, 76, 77, 78, 79 
radioactive, 128, 129, 134, 136, 137 
structure, 136, 137, 138, 139, 140, 

178, 184, 187, 192, 196 
INERTIA, electric charge, 144 

ether, 67 

INGOLD, C K , 65, 209 
INTEGRAL valency, So, Si, 82 
INTERNAL compensation, see Me;>o 
ION, 50 

bound, 84 

IONISATION, see Arrhemus, Clau- 
sius, Faradav, Frankland 
(P F), Helmholtz, Mavwell, 
Stoney, Werner, Williamson 
Ciamician, 54 
hydrolysis, 87 
isomerism, no 
Lodge, 1 20 
valency, 62 
IONIUM, 130 
ISOBARS, isotopic, 152 
non-radioactive, 157 
radioactive, 133 

ISOMERIC atoms, see Isotopes 
ISOMERISM, tee Berzelius, Faraday, 
Werner (co-oidmation theory), 
cis-trans, 102, 115 
double-bond, 102 
geometrical, 61, 62, 84, 102, 115 
see lomsation, Optical and Slereo- 

syn-anti, 115 
ISOTOPES, atomic weights, 32, 67, 

130, 132, 155 
non-radioactive, 153 et seq 

ibobanc, 157 
radioactive, 130 et seq 

ibobanc, 152 
spectra, 152 
whole number rule, 155 

JONES, Mibs E , 208 

TORGENSEN, s M , 60 

JOULE, :> S 


Cbemistiy and Atomic Stmclwe 

K electrons, velocity, 156 

orbits, 149, 1 66 

series spectral, 148, 149, 150, 151 
KEKULE, arrangement of atoms, 43 

atomicity (valency), 42, 57, 60, 61, 

benzene formula, 61 

Canmzzaro, 42 

carbon affinity, 43 

methane type, 41 

mixed types,, 41 

tetrahedral carbon atom, 61 
KELVIN, 143 
KENNER, ] , 209 
KENYON, T , 208 , 
KIMURA, K , 208 
KING, H J S , 208 
KOLBE, 39, 41 
KOSSEL, W, 121 
KRAMERS, H A , 207 

L electrons, 149 
orbits, 149, 1 66 
series spectra, 148, 149, 150 
LANGMUIR, I, 124, 125, 178, 188 
LATENT valency, see Frankland (E ) 

and Friend 
LAURENT, 36, 57 
LAVOISIER, 19, 34, 66 
LAW, atomic heat, 25 
atomic plan, 137, 139, 141, 188, 194 
constant composition, see Proust 
Dalton's, see Dalton, law 
displacement, see Law, radioactive 
group, and Law, spectral 
Dulong and Petit's, see Law, 

atomic heat 

electrolysis, see Faraday, electro- 
equivalent proportions or ratios, 

*ee Richter 
gaseous volumes, see Gay-Lussac, 


Geiger and Nuttall, 134 
group displacement, see Law, radio- 
active , and Law, spectral 
isomorphism, see Mitscherhch 
Joule's, see Joule 
Mencleleeff's, see Law, periodic 

Law con 

Mitscherhch's, see Mitscherhch 
molecular heat, see Joule 
Moseley's, 148, 189 
Newlands', see Newlands 
octaves, see Newlands 
Newton and Kepler, 147 
periodic, 69, 70, 73, 75 

group displacement, see Law, 
radioactive, and Law, spectral 
proportionality, see Richter 
Proust's, see Proust 
radioactive group displacement, 

'Si, 132 

reciprocal proportions or ratios, see 

Richter's, see Richter 

simple multiple proportions by 
volume, see Gay-Lussac, law 

simple multiple proportions by 
weight, see Dalton, law 

Soddy's, see Law, radioactive 

spectral change, 189 

spectroscopic group displacement, 
see Law, spectral 

uniform atomic plan, see Law, 

atomic plan 

LE BEL, 56, 61, 62, 95 
LEDBURY, W, 208 
LEMNISCATE, see Orbits, 202 
LENARD, P , 145 

LEWIS, G N, 122, 125, 179, 180 
LIEBIG, 34, 36, 38, 57 
LIGHT, bending of, 173 

electromagnetic theory, 143 

energy, absorption by electrons, 161 
et seq 

polarisation, 163 

radiation, 144, 145, 147, 149, 160, 

reflection, 163 

relativity theory, 173 

wave mechanism, 161, 162 
LODGE, Sir Oliver, 120, 173, 207 
LORENTZ, H A , 144 
LORING, F. H., 207 
LOWRY, T M , 1 80, 202, 208 
LUMINIFEROUS ether, see Ether 
LUSSAC, see Gay-Lussac 
MAIN SMITH, see Smith 



MALETC acid, 101, 102 
MARION ^C, 66 
MARTIN, G, 128 
MARUKI, T , 208 
MASS, loss of, 155 

negative electricity, 156 

positive electricity, 156 

variation with velocit}, 32, 33, 50, 

143, 144, 156, 165, 170 
MATSUNO, K , 208 
MAXWELL, Clerk, 52, 119, 143, 202 
McCOY, H N , 130 
MENDELEEFF, 70 et seq , 122, 188 
MERCURY (planet), orbital motion, 


MEYER, Lothar, 67, 73 

Victor, 6 1 
MIALL, S , 207 


MILLIKAN, R A 3 192 
MILLS, W H , 98, 209 
MISSING elements, 79, 138, 150, 


MITSCHERLICH, 26, 35, 38 
MOLECULAR, compounds, So, 81 

heat, see Joule 

weights, 23, 24, 29, 31, 38, 53 
MOLECULES, 23 et seq 

of electricity, 52 
MOMENTUM, angular, 163 

moment of, 148, 163 
MORGAN, G. T, 82, 91, 95, 107, 

l8o, igi, 201, 202, 208 

H H , 208 
MOSELEY, H. G. J , 148 et seq , 

177, 189 
MOSS, H W , 208 

J E , 208 

MULTIPLE proportions, laws of, see 
Dalton and Gay-Lussac 

NEON, isotopes, 153 

structure, 138, 140, 178, 184, 192, 


NERNST, W, 119 
NEUTRAL affinities, 58 
NEWLANDS, 69, 70 

NEWTON, Isaac, 17 

NITON, 129, 134, 136, 138, 178, 184, 


NITRIC oxide, covalency, 124 
NITROGEN, artificial disintegration, 


optically active, 46, 63, 98 
valency, 92, 93 
NOBLE gases, 77, 78 / see Inert gases 

metals, 66, 77, 78 
NODDER, C R , 98 
NOMENCLATURE, see Berzehus, 

Brown, Odling, Werner 
Couper, 43 
Dalton, 43 
Frankland, 43 
Lavoisier, 34 

NUCLEAR atoms, see Atomic struc- 
ture (nucleus), Rutherford, 
Thomson (Sir J J ), Weber 
charge, 146, 147, 148, 149, 150, 155, 

158, 160 
co-ordination, 89, 98, 99, too, 107, 

113, 116 

NUCLEUS, theory see Laurent 
NUTTALL, J M , 134 

OCTAHEDRAL atoms, 16, 56, 59, 

95, 97, 103 et seq. 
co-ordination, 84, 95, 103 
OCTAVES, law of, see Newlands 
OCTET theory, 122, 123, 124, 125, 

178, 179, 181, 182, 183 
ODLING, classification of elements, 

68, 70 

nomenclature, 4.1, 44 
valency, 41, 44 

ONE-FLUID electrical theory, 142 
OPTICAL activity, 38, 39, 61, 63, 

84, 95, 98, 99, 102, 108, 114 
isomerism, 61, 62, 63, 85, 97, 98, 
99, 100, 102, 103 et !>eq , 114, 


spectra, see Spectra 

ORBITS, aphelion, 172, 175, 199 
axis, 166, 171; 
binuclear, 201 et seq. 
Bohr, 147 et seq, 160, 161, 162, 

163, 164, 166, 169, 171;, 184, 

186, 193 
Cabman, 201 et seq, 


Cheimstiy and Atomic Stnictme 

Orbits con 

circular, 147, 163, 166, 167, 170, 
172, 204 

ellipsoidal, 174, 175, 201 

elliptic, 163 et seq , 170 et seq , 199 

energy, 148, 162, 164, 165, 170, 200 

ether mechanism, 161 et seq, 172 
et seq 

figure-of-eight, 202 

intranuclear, 156, 161 

mterpenetration, 170, 171, 174, 201 

latus rectum, 166, 169, 174 

lemmscate, 202 

number of types, 148, 164, 165, 166, 
169, 175, 185, 186 

number of electrons, 175, 183, 184, 
185, 186, 192, 193, 194, 196, 
197, 198 

oval, 201 et seq, 

ovals of revolution, 203 

perihelion, 167, 172, 173, 174, 199 

precession, 173, 174, 200 

relativity, 172, 173, 174, 199 
ORIENTATION, 102, 115 
OSMOTIC pressure, 31, 53 
OVAL, Cassiman, see Cabman 

of revolution, see Orbit, ovals 
OXIME isomerism, see Isomcribin, 

PARSON, A. L , 121 
PASTEUR, 39, 57, 62 
PEACHEY, S. J , 46 
PERIHELION, see Orbits 
PERIODIC classification, 27, 32, 73, 
117, 126, 1763 177 et scq , 185, 

curves, 75 

groups, 72, 79, 177 

law, see Law, periodic 

properties, 189 et seq. 

tables, 71, 74, 79 
PERIPATETIC school, 16 
PERKINS, P. B, 129 
PETIT, 25 
PHOSPHORUS, artificial disintegra- 
tion, 158 

optically active, 46, 98 

pentahalides, So 

PICKARD, R H, 208 

PLANAR precession, see Oibits., 

PLANCK'S quantum theory, 143, 

145, 147, 1 60 

PLANETS, relativity effect, 172, 173 
PLATO, 16, 56 
POLYATOMIC radicals, 42 
POLYBASIC acids, theory of, 38, 57 
POLONIUM, 126, 132, 133 
POPE, Sir W. J , 46, 56, 209 
PORTER, C R, 208 
POSITIVE rays, 153 et seq, 183 
PRECESSION, see Orbits 
PRICE, T S , 208 
PROTON, see Hydrogen 
PROUT, 66 

QUANTUM, constant, 145, 148 

energy, 145, 147, 148 

groups, 148, 149, 164 

moment of momentum, 148, 163 

numbers, 148, 149, 164, 165, 166, 
167, 184, 1 86, 193 

subgroups, 176, 185, 186, 187, 188, 
192, 193, 194, 196, 197, 198 

subgroup rule, 186, 193, 194 

theory, see Planck 

RACEMIC acid, 35, 38, 100 

compounds, 85, 100, 112 
RADIAL quantum number, 164, 166, 

200, 201 

RADIANT matter, 144 
RADICALS, atomicity, 41 

basicity, 41 

Berzehus, 37 

Cannizzaro, 42 

Dumas and Licbig, 36 

Frankland, 39, 40 

Gerhardt, 41 

Kolbe, 39, 41 

Lavoisier, 34 

new theory of, 39 

older theory of, 36 

polyatomic, 42 

polybasic, 41 

Wurt?, 41 



RADIOACTIVE change, see Law, 

elements, 133 

group displacement, see Law, radio- 

RADIOACTIVITY, 126 et eq. 
RADIUM, 126, 127, 129 
RADON, 129 
RAMSAY, artificial disintegration, 

electron bonds, 121 

inert gases, 76 

radium emanation (niton), 128, 129 
RARE earths, 79, 152, 197 
REEVES, H. G., 208 
REGULAR solids, 16, 50", 84 
RELATIVITY, 160, 172, 173, 199 
RESIDUAL affinity, 58, 82, 84, 85, 


RESIDUES, theory of, 36, 38 

REY, 17 

RICHTER, 18, 19, 118 

RING structures, 63, 64, 6<;, 91, 101, 


RITZ, 147 
ROOD, E H., 209 
ROSS, W H , 130 
ROYDS, T., 129 
RUSSELL, A. S., 131, 152 

B., 207 
RYDBERG, 147, 163, 177, 188 

SATURATION capacity, see Basicity, 

and Valency 
SCHMIDT, G. C., 126 
SCIENTIFIC Societies Committee, 


SCREENING constant, 149, 150 
SELENIUM acetylacetones, 95 

optically active, 46, 98 
SHIBATA, K , 208 

Y, 208 
SIDGWICK, N. V, 180, 186, 187, 

188, 208 

SILICON, optically active, 46, 98 
SINGLE bonds, 45 

SMILES, S , 46, 56 

SMITH, J D. Mam, 107, 191, 208 

SODDY, F., 207, a-rays and helium, 

128 disintegration hypothesis, 128, 

129 isotopes, 130, 152 

law of radioactive change, see Law, 


radium emanation, 128, 129 
SODIUM, artificial disintegration, 

SOLUTION theory, 31, 47, 49s 53i 

S4> 55 
SOMMERFELD, A., elliptic orbits, 

163, 1705 200, 207 

quantum theory, 145 

relativity effect, 172, 173, 201 
SPATIAL precession, see Orbits, 


SPECTRA, see Absorption, Law, 
spectral, and X-ray 

fine structure, 165, 200 

helium, 148, 163 

hydrogen, 148, 163 

relativity effect, 165, 172, 173 
SPIRO-ATOMS, 92, 98 
STANLEY, H M., 208 
STARK, J , 121, 205 
STAS, 66, 67 

STATIONARY states, 147 
STEREO-CHEMISTRY, 56 et seq , 

84, 95 et seq 
STEREO-ISOMERISM, 61 et seq , 

84, 95 et seq. 
STOCK, A , 207 
STONEY, Johnstone, see Electnne 

classification of elements, 75 

electrolysis, 52 

electron, see Electnne 

natural units of measurement, 52 

periodic spiral curve, 75 

spectral mechanism, 144 
STRAIN theory, 63 
STRUCTURAL formula, 43 
STRUCTURE, see Atomic structure 

chemical, 43, 60 
SUBGROUP scheme, see Bohr, and 

191 to 198 

SUBSTITUTIONS, theory of, 36, 49 
SULLIVAN, J W. N a 207 


Cbemisfty and Atomic Structure 


SULPHUR, optically active, 46, 98 

SYN isomers, 115 

TARTARIC acid, 35, 38, 39, 100 

TAYLOR, C J A , 208 

TELLURIUM dimethyl hahdes, 95 

TETRAHEDRAL angle, 63, 64, 65 

atoms, 1 6, 46, $6, $7, ^, 6 3, 6$, 
84, 95} 97> 9 8 , 99> I00 > I0 3 

symmetry, 61, 63, 65, 84, 97, 98 

valences, 63 
THOMAS, W , 85, 106, 208 
THOMASON, R W , 208 

THOMSON, Sir J. J , 207, and see 
Positive rays 

atomic force law, 182 

plan law, iSS 

structure, 181 et seq 
valency, 182, 183 

chemical bonds, 180, 183 

classical theory, 143, 183 

electron identified, 119, 144, 145 

inertia of charge, 144 

positive sphere theory, 146, 160 
THOMSON, T, 37, 66 
THORON, 128, 136 
THORIUM emanation, 128 

radioactivity, 126 

transformations, 133 
THORPE, J F, 65, 180, 209 

T E, 80 

TIN, optically active, 46, 98 
TRANS isomers, 102, 104, 106, no 
TRANSITION series, 72, 73, 78, 79, 
117, 137, 181, 183, 185, 195, 
196, 197, 198 
TREBLE bonds, 45, 124 
TRIGONAL prism, Astbury's, 97 
TUNSTALL, R. B , 191, -208 
TYPES, mixed, 41 

older theory of, 37, 57 

theory of simple, 41, 57 

URANIUM, radioactivity, 126 et seq 
structure, 136, 139, 196 
transformations, 133 
velocity of K electrons, 156 

VALENCE, 44, 45 

VALENCY, see Cannizzaro, Coup 
Erlenmeyer, Frankland, Frier 
Gerhirdt, Hofmann, Kekn 
Nernst, Odlmg, Spiegel, W< 
ner, Wichelhaus 

see Atomicity, Atomic weigh 
Basicity, Contravalency, C 
valency, Equivalent weigh 
Double valencies, Latent v, 
ency, Polyatomicity, Rebiclu 

Arrhemus, 57 

carbon, 45, Sr, 125 

co-ordination, 54, 58, Si, 82, 8 


Dalton's law, 22, 40 
definition, 44, 82 
electron, 85, 120, 121, 122, 12 

I2 4, 125, 135, 154, 183, 18 

189, 190, 191 
Helmholtz, 52, 119 
integral, 80, 8t, 82 
manganese, 22 
maximum, 122 
nitrogen, 58, 59, 92, 93 
nomenclature, 44, 45, 82 
radioactivity, 131, 132, 135, 13; 

137, 138 

saturation capacity, 38, 57, 62 
variation, 41, 42, 57, 58, 62, 170 

181, 183 

VAN DEN BROEK, see Broek 
VAN DER WAALS, see Waals 
VAN'T HOFF, see Hoff 
VAPOUR density, 24, 31, 32, 53 

pressure, 31, 53 
VARIATION m charge with velocity 

see Velocity 
mass with velocity, see Mass vana 


valency, see Valency 
VELOCITY, variation m charge \vitl 


mass with, see Mass variation 
VERNON, R H , 95 
VIS tellunque, 69 

WAALS, van der, 54 
WARK, I W., 208 
WEBER, 50, 142 
WENZEL, 1 8 



WERNER, Alfied, ammines, 58, 59, 

89, 90 

ammonium theoiy, 158, 63, 90, 92, 93 
chemical affinity, 54, 81, 82 
co-ordination theory, 54, 58, 60, 62, 

8 1 et seq , 95 et seq. 
lomsation, 55, 58, 63, 83, 84, 85, 

89, 90 

metalammmes, see Werner, ammmes 
nomenclature, 89, 90 
optical activity, 39, 84, 106, 108, 

112, 113, 114, 115 
structure of atoms, 59, 60, 62, 63, 

95> 97, "7. 12 4 

molecules, see Werner, co-ordina- 
valency, 54, 58, Si, 82, 83, 84, 85, 

86, 92,95, 117 

WHOLE number rule, see Isotopes 

WILLIAMSON, atomic motion, 51 
double-decomposition, 5 1 
dynamic chemistry, 51 

Williamson con 

ethenncalion, 39 

lomsation, 51 

type theory, 39, 41, 57 
WILSON, C. T R , 163 
WOHLER, 35, 36 
WOLLASTON, atomic, combining, 
and equivalent weights, 21 

basicity, 38, 56 

Dalton's law, 21 

geometrical atomism, 56 

saturation capacity, 38, 56 

stereochemistry, 56 

tetrahedral structure, 56 
WURTZ, ammonia type, 39, 41 

atomic theory, 41 

basicity (valency), 41 
WYCKOFF, R W G , 207 

X-RAYS, 126, 148, 149, 165, 175, 
183, 185, 189, 192, 193, 194, 

200, 297 

YARSLEY, V. E , 208 
ZEEMAN, P, 205 









A I 

It i 





First large edition almost exhausted 


An Introduction to Modem Views on Atomic 
Structure and Radiation 


Demy 8vo. Illustrated. Price 2is. net. 

" One must pay homage to the masterly way Sir Oliver puts the essence of a 
difficult p! oblem befoi e the reader . He has a power of communicating the 

fascination of scientific research which no other autlto), perhaps, possesses to the 
same dcgiee " DR E. N DA C ANDRADE in the " OBSERVER " 


Edited by DR. E. F. ARMSTRONG, F R.S. 

A bnlliant exposition of all branches of modern Chemistry by 

leading scientists 

Crown 4to. 2nd impression Price 155. net 

" A most complete and striking success There has been nothing quite like it before, 
and it is improbable that theie willbe anything like it for a long time to come 
It is to be hoped that it will cuculale freely, for it is as stimulating as it is 


Covers Organic, Inorganic, Physical, Analytical, and Bio- 
Chemistry, including radioactivity and subatomics, thermo- 
dynamics, physiological and pharmacological chemistry, electro- 
chemistry, colloids. 

Crown 4to. Price i6s. Just Pubhshed. 

"Indispensable to all interested in chemistry " GLASGOW HERALD 
" Of immense assistance " CHEMICAL AGE