Skip to main content

Full text of "Recent advances in physical and inorganic chemistry"

See other formats



ISTRY. With an Introduction by J. NORMAN 
COLLIE, Ph.D.,'LL.D., F.R.S. Fourth Edition. 

STEREOCHEMISTRY. With 58 Illustrations. 
Crown 8vo. Second Edition. 

[Text- Books of Physical Chemistry. 




FK;. 23. The Periodic Surface. Front View. 

FIG. 24. The Periodic Surface. Side View. 

SS493 rec 











All rights reserved. 


IN view of the fact that the last edition of this book was 
exhausted within little more than a year of its publication, it 
might be supposed that a fourth edition could have been prepared 
simply by reprinting the third one; but on account of the 
important researches which have been carried out within the 
last twelve months, it was felt that it would be better to rewrite 
part of the volume rather than to let it go out incomplete. As 
a result, this edition differs considerably from its predecessor. 

Additions have been made to the chapter on absorption 
spectra ; the position of the rare earth elements in the Periodic 
Table has been more fully examined ; the newer work on mass- 
spectra has been described ; the problem of transmutation has 
been given increased space to include the work of Eutherford 
on nitrogen ; and the chapter on the isotopes has been slightly 

In view of the recrudescence of active interest in the 
Periodic Arrangement of the elements, brought about by recent 
developments in radiochemistry and positive ray analysis, a 
new chapter has been written on the Periodic Law ; and it is 
hoped that this may help to direct attention to certain points 
in that field. Included in this chapter will be found a descrip- 
tion of a Periodic Surface, which appears to have certain 
advantages over the usual methods of representing the periodi- 
city of elemental properties; and directions for constructing 
the surface are given in an Appendix. Two modifications of 
the Atomic Volume Curve are given, each of which seems to 
present peculiar advantages over the ordinary graph. 

These changes have necessitated the recasting of material 
elsewhere in the book. The chapters on the pseudo-acids and on 
the inert gases have been deleted, as the subjects are now old 
and the space was required for other matter. 


I take this opportunity of explaining some of the objects 
which I have kept before me in the preparation both of this and 
of previous editions. I have endeavoured to write for the 
chemist and not for the mathematician. Unless a subject can 
be treated in such a way that all chemists can understand it 
whether they be mathematically inclined or not I have omitted 
it from this volume. Thus the applications of the quantum 
theory, Nernst's heat theorem, and kindred subjects find no place 
here. It may be objected that to set up such a criterion is to 
make the work one-sided, and this may be frankly admitted. 
I believe, however, that from the point of view of the plain 
chemist, what is lost in this way is gained in others ; and I doubt 
if the inclusion of a few elementary exercises in the calculus 
would recompense the general chemical reader for the corre- 
sponding necessary elimination of some material of more practical 
interest. Another reason for the omission of such subjects is the 
very doubtful value of many of them from the chemical point of 
view. For example, what application to chemistry has Bohr's 
view of the atom shown as yet ? It has not even satisfied the 
purely physical requirements of an atomic hypothesis ; and, so 
far as I am aware, its chemical services have been nil. For this 
reason I have omitted detailed consideration of it, though the 
mathematics required for its comprehension is extremely 

In conclusion, I wish to thank Professor Collie, F.E.S., 
Professor Smiles, F.R.S., O.B.E., and Dr. Wright for assistance 
in the preparation of this edition. My thanks are due also to 
those reviewers whose criticisms have enabled me to improve 
the text in certain places. 


August, 1920. 


IN the arrangement of the material it seemed best to discard 
the sharp divisions between inorganic, physical, and radio- 
chemistry, and to deal with the various subjects in an order 
which obviates as much as possible the assumption that the 
reader is acquainted with matters described later in the volume. 
Thus, for example, X-ray spectra are dealt with earlier than 
the rare earth elements, because Moseley's results define the 
number of the elements which are still missing and thus enable 
the position of the rare earth group in the Periodic System to 
be examined. 

No excuse need be offered for the extension of the part of 
the book devoted to radioactivity, as the recent advances in 
this region are such as to revolutionise our whole outlook on 
the fundamental problems of chemistry. 

As in previous editions, figures indicate references to the 
literature, whilst explanatory footnotes are marked by asterisks. 
In this way the reader can avoid unnecessary reference to the 
foot of the page. The abbreviations employed are mainly those 
adopted by the Chemical Society in its Journal. 

Some readers may detect in the following pages a certain 
distaste for Physical Chemistry of the kind " Made in Germany." 
Lest this should be ascribed to a revision of values consequent 
upon the war, it may be pointed out that exactly similar views, 
not too obscurely worded, are to be found in previous editions. 
I have seen no reason to revise my opinions on this subject 
.since 1909, when the first edition was published. 

October, 1918. 



INTRODUCTION ........... xv 



1. The Types of Electric Furnaces ..... 1 

2. The Essentials of Electric Furnace Construction . . 3 

3. Graphite ..... 

4. Calcium Carbide ........ 6 

5. Silicon and Silicon Carbide ..... 

6. Silicides ...... . . 10 

7. Alundum ........ 11 

8. Phosphorus ......... 12 

9. Carbon Disulphide ........ 12 

10. Alkali Manufacture . . . . . . . 13 


1. The Circulation of Nitrogen in Nature .... 16 

2. The Production of Nitrogen Oxides, Nitrous and Nitric 

Acids .......... 18 

3. The Manufacture of Ammonia ...... 24 

4. The Synthesis of Cyanides ...... 28 

5. The Production of Nitrides ...... 31 

III. THE PERMUTITES ......... 33 


1. Hydrogen Peroxide .... 37 

2. The Metallic Peroxides ....... 39 

3. The General Character of the Per-acids and their Salts . 41 

4. The Persulphuric Acids and the Persulphates ... 42 

5. The Per-acids of Group V ....... 46 

6. The Percarbonates ........ 47 

7. The Perborates 50 




1. The Production of Active Nitrogen . . . . . 53 

2. The Properties of Active Nitrogen ..... 54 

3. The Influence of Impurities ...... 55 

4. The Effect of External Conditions upon Active Nitrogen , 57 

5. A Third Form of Nitrogen 58 


1. General . .59 

2. The Hartley Method 66 

3. The Problem of General Absorption . f , 67 

4. The Factors affecting Selective Absorption ... 76 

5. Valency and Absorptive Power . . . . .81 

6. The Spectroscopic Determination of Chemical Change . 83 

7. The Spectrophotometer 84 

8. The Calculation of Absorption Curves . 86 

9. Conclusion 88 


1. The Phenomena of Crookes' Tubes 90 

2. Characteristic X-rays ....... 91 

3. Corpuscular Kays 96 


1. The Phenomena of Diffraction . . . . . .98 

2. Space Lattices 102 

3. The Crystal as a Diffraction Grating .... 104 

4. Laue's Experiment ....... 105 

5. The X-ray Spectrometer 106 

6. The Analysis of Crystal Structure 107 


1. The Method of Measuring X-ray Spectra .... 114 

2. The Nature of the X-ray Spectra 115 

3. The Gaps in the Periodic Table 118 

4. The Anomalies of the Periodic Table . . . .119 

5. The Spectrum of Hydrogen 120 

6. Molecular Numbers .... ... 121 

7. Conclusion 122 


1. Introductory ......... 124 

2. The Physical Properties of the Bare Earth Elements . 126 

3. The General Chemical Character of the Bare Earth 

Elements 126 

4. Methods of Purifying the Bare Earths . . . .128 

5. Industrial Applications of the Bare Earth Elements . 133 




1. Introductory ......... 135 

2. The Treatment of Monazite Sand 136 

3. Fabrics and their Treatment 138 

4. Impregnation, Fixing, and Branding .... 138 

5. Burning, Collodionizing, and Testing .... 140 


1. The Apparatus 141 

2. The Theory of the Method .143 

3. The Interpretation of the Photographs .... 145 

4. The Substance X 3 149 

5. Conclusion ......... 152 


1. Historical 154 

2. Radium 155 

3. The Becquerel Rays 158 


1. The Disintegration Theory 166 

2. Multiple Disintegration . .... 169 

3. The Radioactive Series . .... 170 

4. Potassium and Rubidium 174 

5. Radioactive Recoil 174 

6. The Geiger-Nuttall Relation 175 


1. The Emanations of Thorium and Actinium . . . 177 

2. Niton . 179 

3. The Production of Helium 187 

4. The Disintegration Theory and the Age of Minerals . . 189 


1. Introductory 191 

2. The Work of Ramsay and Cameron . . . .191 

3. The Investigations of Collie and Patterson . . . 193 

4. Rutherford's Experiments 196 


1. Soddy's Theory of Isotopic Elements .... 198 

2. The a-Ray Change 201 

3. Fleck's Investigations of the Chemistry of the Radio- 

elements ......... 202 

4. The 0-Ray Change 204 

5. Soddy's Law 205 

6. -Rays and Ionic Charges 208 

7. The Atomic Weight of Lead 209 

8. Other Cases of Isotopy . .... 212 

9. Conclusion . . . .214 




1. Isotopes and Isobares 218 

2. The Structure of the Atom 225 

3. The Problem of Atomic Weight 234 


1. MendelSeff, Lothar Meyer and Crookes .... 236 

2. The Periodic Table and the Atomic Volume Curve . . 239 

3. Spiral Arrangements of the Elements .... 250 

4. The Periodic Surface ....... 251 

5. The Problem of the Bare Earth Elements . . . 253 

6. The Theory of Meta-elements 261 

7. Conclusion ......... 263 


APPENDIX. Notes on the Construction of the Periodic Surface . . 275 




AT all periods of the world's history certain problems have 
impressed themselves on men's minds as being of paramount 
importance. History affords many examples of this. From 
the doctrine of " the divine right of kings " to the question 
of " woman's suffrage " is a long step, but not a longer . one 
than from the once all-absorbing theme which exercised the 
minds of men of science as regards the true nature of 
phlogiston to the present contest regarding the structure of 
the cobaltammines. Science has, however, this advantage over 
politics that experiments devised to decide knotty questions 
are more easily carried out; and, further, that it is in the 
interest of no one to conceal the truth. 

In Dr. Stewart's Recent Advances in Physical and 
Inorganic Chemistry, he has treated of a number of subjects 
which are at present prominent in the minds of chemists. 
No doubt they will in time be succeeded by others equally 
engrossing; the essential features of the subjects here con- 
sidered will have been established or refuted, and new views 
will in some cases succeed those which he has expounded. 
Such new views, however, will be the result of maturer 
knowledge, gained by incessant experiment. At present 
these essays represent the state of knowledge which we 
possess; and rival theories have been set forth with fairness, 
and yet not without some guidance, from the author of this 
work, who has not scrupled to express his own opinion 
where he holds a decided view. 

The enormous mass of chemical literature which floods 
our tables monthly makes it impossible to do more than glance 
at the titles of the papers ; and each of these, it is sometimes 
almost sad to think, represents much patient and careful work 
which ousrht to receive, at all events, some consideration. The 


chemically educated public, too, is rapidly increasing; those 
who are not themselves actively engaged in furthering the 
science, and who have neither the leisure nor inclination to sit 
down and read the Transactions of the Chemical Society or the 
Zeitschrift fiir anorganische Chemie, are the better for having 
their chemical food prepared for the table instead of trying 
to assimilate indigestible masses of what is often very crude 
material. Dr. Stewart may be likened to a skilful cook, whc 
has trimmed his joint, rejecting all innutritious and redundant 
excrescences, and has served it up to table in a palatable form, 
Such essays, I venture to think, will do more to encourage 
a taste for chemistry than many text-books. They will be 
followed with pleasure by any one who has mastered the nomen- 
clature and is at home with the simpler conceptions ol 
chemistry. Each may be taken to represent in a readable form 
the latest work on the subject of which it treats. 

The book, it may be hoped, will have a still further use : 
it cannot be doubtful that some who peruse it will have 
suggested to them various directions in which they ma} 
profitably attempt to increase knowledge. Nothing is so sad 
as to see much time and labour spent, with patience and 
devotion, in the investigation of some matter which possesses 
no real importance. It may be retorted that every true 
statement is of importance, but this is not so. It is onlj 
statements which hold forth some prospect of contributing 
to an organic whole which can be held valuable. There 
may, perhaps, be a little more merit in ascertaining to the 
hundredth of a degree the boiling-point of sulphur than oi 
measuring the area of the wings of some particular butterfly 
but the difference is barely appreciable. One is as likely tc 
prove useless as the other. It would be well if enthusiasts 
anxious to carry on research would remember that it is 
much more stimulating to carry on an interesting than an un- 
interesting research. It is, I believe, not improbable that the 
reading of such a work as this may aid those imbued with the 
spirit of investigation to make a happy choice of a subject oi 
research. Should this hope be realised, Dr. Stewart will have 
done a most useful work. 




1. The Types of Electric Furnaces. 

THOUGH the suggestion of an electric furnace dates from Pichon 
in 1853, the first practical model was constructed by Siemens 
in 1878, with the object of fusing refractory metals and their 
ores. In 1885 the Cowles furnace made its appearance, initi- 
ally with a view to the reduction of oxides. Moissan, 1 in 1893, 
carried out a series of investigations on the action of the electric 
furnace upon various materials, and to him we owe much of our 
knowledge of the behaviour of substances at high temperatures. 
The carbides, borides, and silicides were placed by his work 
among the class of readily obtainable compounds. Since that 
time, the application of the electric furnace to scientific and 
commercial problems has become wider ; and at the present day 
the electric furnace industry has its ramifications through the 
most diverse fields. 

Naturally, when the instrument has been applied to so many 
different problems, the apparatus has required modification in 
many respects ; and at this point it may be well to give a brief 
survey of the chief types which are in use. 

Speaking generally, electric furnaces may be divided into 
two main classes : arc furnaces and resistance furnaces. In 
the arc furnace, the heat is derived from an electric arc which 
plays in the furnace ; whilst in the resistance furnace the heat 
is generated in a resistance through which the current runs, and 
from this resistance the heat passes to the charge in the furnace 
by conduction. 

1. Arc Furnaces. This grotip of furnaces may be divided 
into two sub-groups : furnaces with an independent arc, like 

i Moissan, Le Four Electrique. 

i tt 


Moissan's ; and furnaces in which the arc exerts a direct heating 
effect by coming in contact with the charge. In the indepen- 
dent arc furnace, the arc plays between two electrodes placed 
above the surface of the charge; and the heat is conveyed to 
the furnace contents by radiation, either direct or by reflection 
from the walls of the furnace. In the direct heating arc type 
two possibilities present themselves. The Girod furnace has 
one electrode buried in the furnace contents; so that the arc 
plays from the upper electrode on to the surface of the charge. 
In the. Heroult double arc furnace this effect is increased by 
placing both electrodes above the furnace contents, but so far 
removed from one another that no arcing is possible between 
them. The result of this is that in order to pass the current 
from one electrode to the other it is necessary to form an arc 
between the positive electrode and the charge and another arc 
between the charge and the negative electrode. Thus two arcs, 
one from each electrode, play upon the surface of the charge, 
which is included in the electric circuit. 

2. Resistance Furnaces. Here also we have two main 
types : those furnaces which are furnished with a special re- 
sistance ; and those which depend for the resistance upon the 
contents of the furnace. It will be advisable to deal with each 
class separately. 

In the first type it is possible to imagine two variations : 
for we might place the resistance which is to generate the heat 
either in the walls of the furnace or in the middle of the furnace 
contents. In the ordinary laboratory electric tube furnace the 
first method is adopted, the resistance employed being a ribbon 
of platinum or other metallic material coiled around the porce- 
lain tube which forms the actual furnace wherein refractory 
materials are placed. In the carborundum furnace, on the 
other hand, the a resistance is built of carbon which becomes hot 
on the passage of the current; and this carbon resistance is 
imbedded in the furnace charge, which is thus heated by con- 
duction from the resistance. 

Turning now to the second type of resistance furnace, viz., 
that in which no specially constructed resistance is used, it is 
clear that here also a number of modifications may be expected. 
In the first place, we may have furnaces devised purely for 
heating effect ; whilst in the second place, we may have furnaces 


in which both heating effect and electrolytic action come into 
play. In the case of purely heating furnaces, again, the nature 
of the furnace will differ in detail according as to whether the 
furnace contents are solid, semi-fluid, or liquid during the 

The following table 1 will give at a glance the relations 
between the different types of construction : 

f Independent arc 

AT?P "RVTRNTAnrQ J (Single arc 

ARC FURNACES 1 Direct heating arc L, * 

(Double arc 

Kesistance in walls 





With special resistance 1 1> * t, 

(Eesistance in charge 

Without special 



With electrolysis 

Solid charge 
Melting charge 
Liquid charge 

It is unnecessary to describe the various types in detail at this 
point, as the most important differences will be referred to in 
connection with the principal products of the electric furnace 
industry. In the remaining portion of this chapter the non- 
metallurgical applications of the electric furnace will be chiefly 
dwelt upon ; as the metallurgical side would occupy more space 
than can be spared in this volume. 2 

2. The Essentials of Electric Furnace Construction. 

Eeduced to its simplest terms, an electric furnace comprises 
three factors: (1) a source of current; (2) electrodes and their 
holders ; (3) a refractory casing. 

With regard to the current, this is usually supplied by a 
dynamo coupled with a transformer to reduce the voltage. 
Generally an alternating current is used ; but in the case of 
processes involving electrolytic action a direct current is neces- 
sary. Naturally the electric furnace industry tends to centre 
in districts like Niagara where water-power is cheap and current 

1 Modified from Stansfield, The Electric Furnace. 

2 For details on this part of the subject, see Stansfield, The Electric Fur- 
nace (1914); Wright, Electric Furnaces and their Industrial Applications 
(1910) ; and Bideal, Electrometallurgy (1918). 


can be obtained economically. It is unnecessary to give any 
figures with regard to quantities of current employed, since 
these vary from process to process. 

The electrodes usually employed in electric furnaces are 
carbon rods, for no other material is capable of standing the 
high temperatures which can be obtained. In certain types of 
furnace actual electrodes are not necessary, the current in such 
cases being generated in the furnace by induction. Provision 
is sometimes necessary to avoid the external ends of the elec- 
trodes becoming too hot. Connection between the electrodes 
and the cables carrying the current is made by means of metal 
clamps which also act as electrode holders. 

We now come to one of the most important points in an 
electric furnace : the materials of which the furnace casing is 
composed. It must be remembered that the temperature of 
the positive crater of an arc playing between carbon rods is 
about 3500 C., so it is clear that any material which is to be 
brought near such a focus must be of the most infusible type. 
In one of Moissan's experiments his furnace, which was built 
of quicklime, was destroyed in a few minutes : the quicklime 
melted and ran like water ; and a portion of it was even vola- 

Nor is it always sufficient merely to choose a refractory 
material for the furnace walls ; for it may be that the furnace 
charge, under the action of heat, may act upon the furnace 
lining, causing great deterioration. 

Amongst the most suitable materials for furnace linings are 
the following. Fireclay bricks will stand a temperature of 
1600 C., but they are rapidly attacked by slags containing 
metallic oxides. The same defect is noticeable in silica bricks, 
but these will remain unfused at 1700 0. Lime and magnesia 
will stand temperatures above 2000 C., but they are both 
crumbly materials which are apt to give trouble. On the 
other hand, they do not form fluxes with metallic slags. Chrome 
iron ore (chromite) is a neutral refractory material which is 
not easily acted on and stands temperatures up to 2100 C. 
Alundum is almost as infusible. Carbon itself is, of course, 
the most infusible material of all ; but it suffers from the dis- 
advantage that it is readily oxidized at high temperatures, so 
its use is restricted. Zirconia is also utilized. 


It is sometimes found convenient to construct the furnace 
wall in two layers, the front layer being refractory but not 
completely non-conducting, whilst behind it is placed a layer of 
some material of lower conductivity which prevents loss of heat 
from the furnace. 

In connection with the furnace construction, another point 
may be mentioned. Some furnaces are designed to run with 
a single charge ; and at the end of the process the furnace is 
allowed to cool and the charge is withdrawn. Other types 
have been devised which permit of continuous action, the 
finished material being withdrawn and fresh charges being 
introduced without any interruption in the action of the, 
furnace. Finally, by using a revolving furnace, it is possible 
to combine the advantages of both these types : the arc is kept 
continually at work while raw material is being added at a 
second region of the furnace, and the finished product is being 
cooled and extracted at yet a third point. 

3. Graphite. 

The action of the electric furnace upon carbon was studied 
by Moissan, who found that under suitable conditions amor- 
phous carbon could be transformed into either graphite or 
diamond. By dissolving carbon in molten iron in the electric 
furnace and then cooling the product in liquid lead, it is pos- 
sible to produce microscopic diamonds which can be separated 
from the iron mass by removing the metal with acids. The 
process is of no commercial value. 

The electric furnace used to produce graphite on a com- 
mercial scale is a resistance furnace in which the solid charge 
itself acts as the resistance. The furnace is about 30 feet long 
by 2 feet square, and its permanent portion consists of a base 
and two upright ends through which several electrodes pass. 
The charge, which is about 6 tons in weight, consists of sand 
and anthracite ground to the size of rice. Since cold anthra- 
cite is a poor conductor of electricity, it is necessary to provide 
for preliminary heating ; and this is accomplished by joining 
two of the electrodes by a thin carbon rod which runs through 
the centre of the charge and becomes heated by the passage of 
the current. The sides of the furnace are roughly built in 
with loose bricks before the run is commenced. 


The mechanism of the graphitization process is not fully 
understood. It is assumed that carbides are formed from the 
metallic constituents of the anthracite; that these carbides, 
formed in the hot zone, are almost immediately decomposed ; 
that the carbon thus liberated is in the graphitic form ; and 
that the volatile metals set free by the decomposition of the 
carbide again form carbides in another part of the furnace. 
In this way a comparatively small proportion of metal in the 
anthracite will suffice to convert the whole of the carbon 
present, via carbide, into graphite. 

A run of the furnace described above occupies twenty 
hours; and the temperature attained appears to be above 
2000 C. The material is extracted by unbuilding the loose 
sides of the furnace after it has cooled. 

The same type of furnace is employed to prepare graphite 
electrodes. A mixture of petroleum coke and pitch is moulded 
into the shape of the required electrodes, some carbide-forming 
material such as iron oxide or silica being added. After being 
baked, the rods are packed into the furnace transversely to the 
path of the current; and the graphitization is carried out as 

The most important manufacturers of graphite are the 
International Acheson Graphite Co., whose works are at 
Niagara Falls ; and to them is due the credit of having de- 
vised methods whereby graphite can be readily utilized as a 
lubricant. The Acheson Co. produce a very soft variety of 
graphite which, when reduced to an extremely fine state of 
division, is known as " Dag " (from the initials of Deflocculated 
Acheson Graphite). Mixed with grease, this comes on the 
market as Gredag ; suspended in oil it is known as Oildag ; 
whilst when water is substituted for oil the product is termed 
Aquadag. All three forms are excellent lubricants owing to 
their graphite content. 

4. Calcium Carbide. 

Calcium carbide is formed by the action of lime upon 
carbon in the electric furnace 

CaO + 3C = CaC 2 + CO 
Since under the conditions of the preparation the carbide is 


liquid, two possibilities present themselves : it is possible (1) 
to let the furnace contents solidify at the end of the run, 
or (2) to run off the liquid carbide from the furnace as it is 
formed. The furnaces used for the first process are termed 
ingot furnaces, whilst the second method is carried out by 
means of tapping furnaces. 

One of the most effective types of ingot furnace is that 
devised by Willson. The actual furnace is a sheet-iron box 
about 6 feet high, mounted on wheels so that it can travel 
along a railway. The two graphite electrodes are mounted 
vertically over the line of rails and can be lowered into the 
box when it is in the proper position. After the contents of 
one box have been treated, the electrodes are raised ; the box 
is rolled away ; and another box is brought under the electrodes 
in its turn. In this way charges can be treated in succession 
by the same pair of electrodes ; and no time is wasted in cool- 
ing the product. The furnace charge consists of burnt lime 
and anthracite or coke crushed to pieces about a quarter of an 
inch in diameter. Powdered material does not give such good 
results as granulated particles do. The charge is introduced 
down shoots and is kept on a level about 2 feet above the 
arc. It is stirred up at intervals to facilitate the escape of 
gases. The weight of the charge is about 10 cwt. and from 
this some 800 Ib. of commercial carbide is obtained. 

Other types of ingot furnace have no truck, but instead 
the bottom of the furnace opens outwards so as to allow the 
carbide to drop through. There is also the Horry type, in 
which the rotary principle is employed ; so that once the carbide 
is formed it is removed into a cooler region, by the slow 
revolution of the furnace, whilst the arc plays upon fresh raw 

Turning now to the tapping furnaces, these are furnished 
with a large vertical electrode, about 8 feet long, composed of 
rods held in a circular holder and having carbon paste rammed 
into the interstices of the mass. The furnace is permanent 
and is provided with a vent through which the molten carbide 
is drawn off (or "tapped") as it is formed. 

The choice of the raw materials for calcium carbide manu- 
facture requires care, as otherwise the process may be danger- 
ous, or the resulting product may be contaminated with 


undesirable matter. The presence of magnesia interferes with 
the carbide formation ; any phosphorus may give rise to 
phosphoretted hydrogen if the carbide is utilized for the 
production of acetylene ; whilst if the coal used contains much 
mineral matter the end-product may be of bad quality. 

The main commercial application of calcium carbide is the 
generation of acetylene ; but in recent times, as will be seen 
in the following chapter, carbides have been employed in the 
fixation of nitrogen. 

5. Silicon and Silicon Carbide. 

The electric furnace is now employed for the manufacture 
of silicon on a commercial scale from silica and carbon. The 
exact details of the furnace vary according to the process 
used ; but in general two carbon electrodes dip vertically into 
the charge. In some furnaces the two electrodes form the 
poles of the discharge ; whilst in others the current passes 
between the vertical electrodes and the carbon lining of the 

The reaction is simply one of reduction 

Si0 2 + = Si + C0 2 

but in practice there is considerable difficulty in avoiding the 
formation of some silicon carbide as well 

Si0 2 + 30 = SiC + 200 

The charge consists of coke and silica in the form of sand or 
crushed flint. The main practical difficulty is found in regu- 
lating the temperature of the furnace ; for if this be allowed 
to rise too high the silicon volatilizes and there is also some 
loss due to the formation of silicon carbide which contaminates 
the product. The heating effect is generally controlled within 
limits which allow the silicon, when formed, to melt and per- 
colate down through the charge to the bottom of the furnace, 
whence it is run off. 

Electric furnace silicon is a brittle crystalline body with 
a dark silver lustre. It contains about 2 per cent, of carbon 
along with lesser proportions of iron and aluminium. It can 
be cast into any form which may be required. 

Commercially, silicon is employed in several ways. In 


steel-making, it is used in certain cases instead of ferro-silicon. 
It is also utilized as a substitute for aluminium in the thermit 
process for manufacturing low-carbon ferro-alloys ; whilst a 
silicon-copper alloy, silicon -copper, plays a part as a deoxidizer 
in the making of copper castings. 

More important than silicon itself is its carbide, carbor- 
undum, which is formed when carbon and silica are heated in 
the electric furnace under certain conditions. On a commercial 
scale, the furnace charge is made up of silicious sand, coke, 
sawdust, and about 2 per cent, of common salt. The sawdust 
is used in order to make the charge porous and thus permit the 
escape of the carbon monoxide formed in the reaction ; whilst 
the addition of the salt is due to the fact that the furnace runs 
better when this flux is present. 

The furnace employed resembles that used for graphite 
manufacture. It is about 16 feet long, 6 feet wide, and 5 feet 
high. Between the electrodes a core of coke is placed, running 
from end to end of the furnace. This core acts as a resistance 
and heats the surrounding charge by conduction. The run of 
the furnace lasts about thirty-six hours. 

When the charge is broken up at the end of the run, it is 
found that the furnace contents are arranged around the core 
in a series of layers. Immediately surrounding the graphitized 
core lies a layer of crystallized carborundum about 1J- feet 
thick. This is the valuable part of the product. Outside this 
layer is found a greenish-coloured mass of impure amorphous 
silicon carbide which has no commercial value; whilst still 
further from the core the original materials of the charge are 
almost unaltered. These last two layers are usually worked 
up in the next charge. 

Carborundum thus prepared is a dark brown or black 
crystalline substance. Owing to its extreme hardness, it is 
employed as an abrasive ; its refractory nature makes it useful 
for lining steel and cement furnaces ; and it is also utilized as 
a deoxidizer in steel manufacture. 

Silundum is a compact form of carborundum which is 
obtained by impregnating carbon with silicon vapour in the 
electric furnace. The charge may be composed of silica and 
carborundum or of sand and coke. In either case, the object 
is to produce silicon 


Si0 2 + 2SiC = 3Si 4- 2CO 
Si0 2 + 20 = Si + 200 

The carbon articles which it is desired to convert into silundum 
are packed in the charge separate from the furnace core. 
During the run, silicon vapour is formed which attacks the 
carbon of the models and converts them into silundum shapes. 
Silundum owes its technical value to the fact that it is not 
readily oxidized, which renders it extremely useful as a 
material for electrical resistances intended to work at high 

Two other silicon derivatives deserve mention here. The 
first of these is siloxicon, which is a mixture of various com- 
pounds of carbon, silicon, and oxygen, and has the general 
formula Si n C n O. It is formed in the electric furnace by using 
a carbon-silica charge in which the carbon is insufficient to 
form silicon carbide with the whole of the silicon present. 
Siloxicon is a useful refractory material. It is a greyish-green 

Monox is a mixture of silica, amorphous silicon, and silicon 
monoxide, SiO. Under electric furnace conditions, silicon 
monoxide is a gas ; and if it be allowed to escape into an air- 
free chamber, a fine brown powder condenses. This is termed 
monox. The reaction appears to be as follows : 

Si0 2 + = SiO + CO 

Monox is extremely light ; a cubic foot of the powder weighs 
only 2^ Ib. It was employed as a paint and also as an in- 
gredient in printers' ink ; but its manufacture has now been 

6. Silicides. 

Just as a carbon yields carbides of the alkaline earth metals, 
so silicon can form silicides of an analogous type 

CaC 2 CaSi 2 

Calcium carbide. Calcium silicide. 

These silicides are obtainable in either of two ways ; for we 
may reduce the corresponding silicate with carbon in the 


electric furnace ; or we may heat silica and carbon with the 
oxide or salt of the required alkaline earth metal 

CaH 2 Si 2 6 , H 2 + 50 = CaSi 2 + 5CO + 2H 2 
CaO + 2Si0 2 + 50 = CaSi 2 + 500 

The formation of the silicides requires a temperature higher 
than that which suffices for carbide production. 

The silicides are bluish- white in tint and of metallic 
appearance. Even at ordinary temperatures they oxidize 
slowly in air, yielding silica and the oxide of the metal from 
which they are derived. When treated with water, they 
liberate hydrogen 

CaSi 2 + 6H 2 = Ca(OH) 2 . 2Si0 2 + 5H 2 

This property makes them extremely useful for military pur- 
poses, such as inflating airships : for the plant required in the 
hydrogen generation is extremely simple ; the efficiency of the 
silicides, weight for weight, is superior to that of metals and 
sulphuric acid ; and the hydrogen thus produced is free from 
contamination by heavier gases. 

Owing to this power of liberating hydrogen from water, 
the silicides are good reducing agents ; and barium silicide has 
been employed on a commercial scale for converting indigo 
into indigo-white. 

7. Alundum. 

This material is an artificial corundum or emery. It is 
manufactured by first calcining bauxite in an ordinary 
furnace to remove combined water from the material; after 
which some carbon is added for the purpose of reducing any 
metallic impurities, such as titanium or iron, contained in the 
bauxite. The final stage in the process is fusion in the electric 
arc furnace. The charge is allowed to cool slowly ; crystalliza- 
tion takes place; and the metallic impurities segregate in 
nodules, which can be picked out after the charge has been 

The crushed material is sorted out according to the size of 
the grains, and these are roasted in order to oxidize any metallic 
particles in them. After this the grains are mixed with fire- 
clay, moulded into wheels, and fired in kilns. 


8. Phosphorus. 

The electric furnace can be utilized to produce phosphorus 
either directly from bone-ash or from mineral phosphates. 
In either case, the reduction is carried out by mixing the 
phosphate with silica and charcoal, and subjecting the product 
to the action of an arc furnace with a vertical electrode. Phos- 
phorus vapour, carbon monoxide, and a silicate slag are pro- 
duced. The addition of the silica permits more rapid working ; 
as the slag can be run off in its molten form; thus time is 
saved in cleaning the furnace, which would have to be done if 
the by-products of the reaction were left to accumulate in the 
form in which they are produced. 

9. Carbon Disulphide. 

In the manufacture of carbon disulphide, the electric furnace 
appears likely to displace the older methods in a very brief 
time. The furnace actually employed is of a special pattern. 
Originally, horizontal carbon electrodes were used as terminals ; 
but these have now been dispensed with ; and in the newest 
types of furnace the electrical connection is made directly with 
the carbon of the charge. 

The actual furnace is a cylinder about 40 feet high and 16 
feet in diameter. The walls of this tower are hollow and the 
sulphur is fed in near the top, forming a non-conducting layer 
within the walls and thus preventing loss of heat from the 
furnace. As it sinks down under the action of gravity, the 
sulphur melts and finally emerges on to the furnace hearth. 
The central hollow of the tower is filled with a closely-packed 
mass of carbon which is replenished from shoots at the side. 
The current is supplied, on the one hand, through the carbon in 
the shoots, and, on the other hand, through the furnace bed. 
On the hearth of the furnace the heat acts upon the mixture 
of carbon and sulphur, forming carbon disulphide and sulphur 
vapour. The latter, in its passage upwards, reacts with some 
of the carbon and yields more disulphide. The disulphide 
vapour finally passes out of the top of the tower and is con- 
densed in suitable receptacles. The furnace is capable of 
running continuously for months at a time. 


10. Alkali Manufacture. 

Hitherto in this chapter we have been concerned with 
electric furnaces in which the heating effect of the electric 
current played the main part in the process ; but we must now 
turn to a different type wherein the action of the current is 
electrolytic in character. As an example of this, the manu- 
facture of caustic soda and metallic sodium may be chosen. 

In the Acker process for the commercial production of 
caustic soda and chlorine, fused sodium chloride is used as the 
electrolyte. Into this dips a series of carbon anodes ; whilst the 
cathode is a mass of molten lead. The whole arrangement is 
gas-tight ; there is an outlet for chlorine above the anodes and 
an injector through which steam can be forced. The inlet and 
outlet are separated from one another by a trap. 

The reactions which take place during the operation of the 
furnace are as follows. The sodium chloride is electrolysed, 
giving gaseous chlorine (which passes off through the outlet) 
and sodium (which forms an alloy with the molten lead of the 
cathode). This molten material then circulates through a trap 
into another compartment where it meets a blast of steam 
which attacks the sodium and produces caustic soda solution, 
as well as hydrogen. The hydrogen is emitted through a flue ; 
the alkaline liquid floats on the top of the molten lead and is 
drawn off through an outlet at the proper level ; whilst the lead 
passes again into the electrolytic chamber to take its place in 
the cathode mass. The hydrogen is utilized in burners for the 
preliminary fusion of the charge ; the chlorine is employed for 
the manufacture of bleaching powder. It will be seen that 
the successful operation of the method depends almost entirely 
upon the proper circulation of the molten lead ; for the sodium 
alloy decomposes if left long in contact with the fused sodium 

Turning now to the problem of manufacturing metallic 
sodium, the Castner process is the most successful on a large 
scale. The raw material used in it is caustic soda which is 
electrolysed in the fused condition. Apparently the following 
series of changes takes place during the electrolysis. In the 
first place, the fused caustic soda is electrolysed 

2]NaOH = 2Na (at cathode) + 20H (at anode) 


At the anode, the hydroxyl ions combine to form water and 

40H = 2H 2 + 2 

This water may then be supposed to become electrolysed, 
yielding hydrogen at the cathode and oxygen at the anode 

2H 2 = 2H 2 (at cathode) + 2 (at anode) 
Collecting these changes together in one equation we have 
2NaOH = Na 2 + H 2 + 2 

(at cathode) (at anode) 

A technical difficulty appears at this point. If we place 
the electrodes far apart from one another, we shall incur a 
considerable expense in current. On the other hand, if we 
place them close together, we run the risk of the sodium being 
attacked by the liberated oxygen with disastrous results. 

This danger has been obviated in an ingenious manner. 
The furnace is designed in the form of a cast-iron crucible with 
a furnace underneath for fusing the caustic soda at the be- 
ginning of the process. Through the bottom of the crucible, 
the cathode protrudes vertically. Surrounding this in the 
upper part of the crucible is an annular nickel anode, the dis- 
tance between the two being only an inch in some types of 
furnace. Between cathode and anode lies a cylinder of nickel 
gauze which extends upward above the top of the cathode and 
ends in a metal cylinder. This gauze screen guides the sodium 
as it floats upward from the cathode through the fused electro- 
lyte ; and at the same time it prevents the stream of oxygen 
from the anode from coming in contact with the newly-formed 
metal. It is found that the process works in a satisfactory 

The Ashcrof t sodium process depends upon a different series 
of reactions. The plant is in two parts connected by a pipe. 
In the first tank, fused sodium chloride is electrolysed, a molten 
lead cathode being employed as in the Acker process. This 
molten alloy is then circulated through the pipe to the second 
tank, which contains fused caustic soda. In this second tank 
there are two electrodes. Sodium is liberated at the cathode 
and hydroxyl ions at the anode. These hydroxyl ions then 
attack the sodium in the lead alloy, yielding caustic soda. In 


this way the loss of sodium from the fused caustic soda is con- 
tinually made good. The liberated sodium floats to the surface 
of the fused electrolyte and is removed by a cowl and tube. 

It is claimed that the Ashcroft process, when worked on a 
large scale, will be more efficient than the Castner process for 
three reasons. In the first place, it uses as a raw material 
common salt, which is cheaper than caustic alkali ; secondly, 
chlorine is a valuable by-product of the process ; and, thirdly, 
the current consumption in the electrolysis of the fused alkali 
corresponds to a production of sodium only ; whilst in the 
Castner process half the current is employed in producing 

Enough has now been said to give the reader some idea of 
the electrolytic furnace ; and it is unnecessary to deal with other 
types, as this chapter is not intended to contain a complete 
account of the electric furnace industry. Further information 
on the subject must be sought in works of reference. 1 It will 
suffice to mention the application of the electric furnace to 
metallurgical problems and to the glass and fused quartz 
industry ; whilst its use in the fixation of nitrogen will be de- 
scribed in the next chapter. 

1 See among others : Stansfield, The Electric Furnace ; Wright, Electric 
Furnaces and their Industrial Applications; Martin, Industrial Inorganic 
Chemistry ; Kideal, Electrometallurgy ; Pring, The Electric Arc. 



1. The Circulation of Nitrogen in Nature. 

ANIMALS have two fundamental needs: oxygen and nourish- 
ment. Oxygen they can derive direct from the air, but 
nourishment can only be obtained by a more roundabout 
process. At the basis of all animal structure lies a group of 
substances called proteins, which the animal, being unable to 
synthesize itself, obtains from vegetables. These proteins con- 
tain from 15 to 20 per cent, of nitrogen. It has been shown 
that non-nitrogenous vegetable products, such as sugars, are 
incapable of supporting life. Animals, therefore, subsist upon 
the nitrogenous bodies of the vegetable kingdom ; but this 
only carries us a step further back, and leaves us to discover 
how this nitrogenous matter is obtained by plants. It is found 
that part of the nitrogen in plants is derived from the air by 
means of certain bacteria, but the major part of it comes from 
the decomposition of plant tissue by one means or another. 
For example, a plant may be eaten by an animal ; a certain 
amount of the nitrogen is excreted in the form of urea deriva- 
tives and the like ; these break down to ammonia, which is then 
oxidized by bacteria to nitrous and nitric acids ; and these, in 
turn, are absorbed by the vegetable world once more. 

If one could enclose a few plants and a suitable series of 
animals in a bell-jar, this process of nitrogen circulation might 
conceivably be prolonged indefinitely; but, unfortunately, 
when we have to deal with the matter from a practical stand- 
point, the question of waste bulks very largely. The scheme 
below shows the main steps in the process of circulation, and 
we have now to consider at what points waste is to be expected. 






NITROUS ACID -^-Nitrifying bacteria AMMONIA 

In the first place, we must bear in mind that though the 
vegetable kingdom as a whole has its uses in the question of 
respiration, where it counterbalances the loss of oxygen due to 
the animal kingdom, it is only certain groups of plants which 
can serve as food. Thus any nitrogen which is absorbed by 
non-edible plants is, temporarily at least, lost from the point 
of view of nourishment. Again, we dispose of the greater part 
of our sewage by running it into the sea ; but in this way 
the larger portion of the nitrogen contained in it is lost to us 
and goes to build up sea-plants, for which we have at present 
no use. Further, the nitrous and nitric acids produced from 
ammonia in nature do not necessarily remain on the spot 
where they are produced; they may disperse and be lost, as 
far as practical purposes are concerned. 

In the past, these losses have been counterbalanced by two 
methods : either by inoculating plants with nitrifying bacteria, 
or by manuring the soil in which they grow. With the former 
method we have no concern here, but we must enter into some 
consideration of the second. 

At the present day, the nitre beds 1 of Chile and Peru 
supply the greater part of the cheap nitrogenous matter which 
is required in agriculture ; and the extent of the demand can 
be gauged from the following figures which give the amounts 
exported in various years : 

1 See Norton, Consular Report on the Utilization of Atmospheric Nitrogen, 
Washington, 1912 ; Martin, Industrial Inorganic Chemistry, I., 432. 




1890 . 1,000,000 

1900 1,400,000 

1910 . 2,274,000 

1912 2,542,000 

Within quite recent times, fresh nitre beds have been dis- 
covered in South America, so that there is no immediate pro- 
spect of exhaustion in that quarter. None the less there is a 
possibility of shortage, and within a measurable -time we shall 
be faced with the problem of producing nitrogenous substances 
upon a vast scale and at a reasonably low cost. 1 By adopting 
a less wasteful method of sewage disposal, and perhaps by 
utilizing as nitrogen sources such substances as the seaweed 
from the Sargasso Sea, we may hope to diminish the amount 
of synthetic material required ; but at the best we cannot hope 
to dispense with it altogether. 

During the war, the Germans were cut off from the natural 
sources of nitrates ; and in order to keep up their supply of 
munitions they were driven to utilize synthetic methods : so 
that it seems probable that a very large amount of suitable 
plant is now in existence. 2 

For synthetic purposes, the cheapest source of nitrogen 
is the atmosphere ; and various methods are now available, 
by means of which atmospheric nitrogen can be converted into 
compounds suitable for manures. This process has been termed 
the fixation of nitrogen ; and in the following sections of 
this chapter we shall deal with the main methods which have 
been devised with that end in view. 

2. The Production of Nitrogen Oxides, Nitrous and Nitric 


In any technical process we must bear in mind two things 
which govern the price of the final product: the cost of the 
starting materials, and the expenditure of energy necessary to 
convert them into the compound required. Now, in the manu- 
facture of nitrogen compounds, the cheapest source of nitrogen 

1 See Crookes, The Wheat Problem, for an account of the food question ; 
and for the chemical side, Knox, The Fixation of Atmospheric Nitrogen, and 
Martin, Industrial Inorganic Chemistry, Vol. I. 

2 For the British side, see Dobbie, Trans., 1920, 117, 430. 


is certainly the atmosphere : but we have not only to obtain 
nitrogen ; we have also to find some comparatively cheap sub- 
stance with which to combine it ; and, finally, we must devise 
an economical method of effecting the combination. 

It is obvious that oxygen is the cheapest element we could 
choose in order to form a compound with nitrogen, for the two 
occur together in the atmosphere in what is, for practical 
purposes, an unlimited quantity. The question then arises : 
Can we convert nitrogen into its oxides with an expenditure of 
energy low enough to make it pay in practice? Up to the 
present time, the methods applied to this question have been 
almost entirely electrical ; so that the problem, in this depart- 
ment at least, rests chiefly upon the cost of electricity. The 
electric power, in most processes, is utilized in the form of 
an arc. 

Crookes, 1 in 1892, showed that, at the temperature of a high 
tension arc, oxygen and nitrogen combined to form nitric 
oxide. Eayleigh 2 five years later proved that in this way 25 
c.c. of oxides of nitrogen could be produced by the expenditure 
of one watt- hour. The action of the arc appears to be due 
both to its heating effect and to electrical influences. 3 

At this point it may be well to recall the work of Nernst 4 
upon gas equilibria at high temperatures. If we take the case 
of the union of nitrogen and oxygen to form nitric oxide, we 
have the following equation : 

N 2 + 2 ^ 2NO 

An examination of this equation will show that the volume 
of gas at the beginning of the reaction is the same as that at 
the end. Now, if no change in volume takes place during 
a reaction, it is clear that the position of equilibrium of such a 
reaction would be uninfluenced by the pressure under which 
it is carried out. 

1 Crookes, Chem. News, 1892, 65, 301. 

2 Rayleigh, Trans., 1897, 71, 181. 

8 McDougall and Howies, Mem. Manchester Phil Soc., 1900, 4, 44 ; Muth- 
mann and Hofer, Ber., 1903, 36, 438; .Le Blanc and Niiranen, Zeitsch. 
Electrochem., 1907, 13, 297 ; Gran and Buss, ibid., 345, 573 ; Lee and Beyer, 
ibid., 701; Foerster, ibid., 1906, 12, 536; compare Haber and Koenig, ibid., 
1907, 13, 725 ; 1908, 14, 689 ; Warburg, ibid., 1906, 12, 540. A full summary 
of the evidence is to be found in Knox, The Fixation of Atmospheric Nitrogen. 

4 Nernst, Application of Thermodynamics to Chemistry, 1907. 


It is obvious that the position of equilibrium in a reversible 
reaction may be very largely dependent upon the temperature 
at which the reaction is carried out. Nernst has shown, for 
instance, that if air be passed through a tube heated to 1538 C., 
only 0'4 per cent, of nitric oxide is formed ; while at 1922 C. 
the quantity rises to 1 per cent. The results appear to be in 
contradiction to the fact that nitric oxide decomposes at a 
white heat. The matter will be made clear if we take into 
account the velocities with which the two reactions 

N 2 + 2 = 2NO ...... (1) 

2NO = N 2 + 2 ..... (2) 

proceed. According to the Guldberg-Waage Law of Mass 
Action, the amount of chemical action is proportional to the 
active masses of the substances reacting. Thus the chemical 
action corresponding to equation (1) will be proportional to 
the amounts of nitrogen and oxygen present; while equation 
(2) represents an action which will be proportional to the 
amount of nitric oxide used. Thus if v 1 be the velocity of 
formation of nitric oxide, then we have the equation 

! = h x (N 2 ) x (0 2 ) 

where (N 2 ) and (0 2 ) are the active masses of nitrogen and 
oxygen present. Again, if v 2 be the velocity of formation of 
nitrogen and oxygen from nitric oxide, then 

v 2 = L 2 x (NO) x (NO) = Jc 2 x (NO) 2 

Suppose, now, that the two reactions have proceeded until 
equilibrium has been reached. At this point, the amount of 
nitric oxide formed is equal to that decomposed. Conse- 

and the equilibrium constant, K, is the ratio of the one velocity 
constant to the other. In other words 

Now, the values of k and & 2 -can ^ e ascertained experimentally, 
so that we can easily determine the value of the equilibrium 
constant K. 


When we apply this line of argument to the concrete case 
of the production of nitric oxide from the air, we proceed as 
follows. It is evident that the equilibrium constant K will be 
dependent upon the active masses of the three gases present, 
nitrogen, oxygen, and nitric oxide, and will be expressed 


(N 2 ) x (0 2 ) 

Let x be the amount of nitric oxide produced at the equili- 
brium point; then, since the concentration of nitrogen at the 
start is 79 per cent., and that of oxygen is 21 per cent, (these 
being the proportions in which these gases occur in air), we 
have at constant temperature the following equation : 


Since we can determine K experimentally, we can solve the 
equation for x t and this will give us the quantity of nitric 
oxide formed when the equilibrium point is reached. 

Now, K alters with the temperature according to the 

d loge K _ Q_ 
dT 2T 2 

where Q is the change of energy in the system, and T the 
absolute temperature. Integrating this under the assumption 
that Q is independent of the temperature (as it apparently is 
for all practical purposes in the case we are considering), we 
then have for two temperatures T and TI the equation 

log, Kl - log e K = - 

Thus, if we know the equilibrium constant K for temperature 
TO, and also the energy change in the system, we can calculate 
from this last equation what the equilibrium constant will be 
for any other temperature TI. The energy change in the 
system can be determined by finding the equilibrium constant 
for two temperatures, and then substituting in the above 
equation. In this way, from a couple of measurements at 
different temperatures, we shall be in a position to foretell 


the equilibrium constant of the reactions for any temperature. 
The agreement between theory and experiment is quite close, 
as the following figures l show : 

Tm ^rafur, P er Cent. Per Cent, 

emperature. N0 Observed> N0 calculated. 

1811 0-37 0-35 

1877 . . . / 0-42 0-43 

2033 0-64 0-67 

2195* . . . . . 0-97 0-98 

2580 2-05 2-02 

2675 2-23 2-35 

Another point of importance arises in this connection. 
From what has already been said, it is clear that the velocity 
of decomposition of nitric oxide is very greatly influenced by 
the temperature at which the reaction is carried out. If we 
can bring the reactions into equilibrium at a temperature of 
about 2500 C., we shall have in the mixture, according to 
Nernst's data, about. 2 per cent, of nitric oxide. If we cool 
the gases down slowly, they will come into a fresh state of 
equilibrium for each temperature they pass through, and con- 
sequently when we reach ordinary temperatures the percentage 
. of nitric oxide will be negligible. On the other hand, if we 
produce a sudden drop in the temperature, the gases will not 
attain equilibrium quickly enough to keep pace with the fall 
in temperature, and consequently we shall be able to reach 
ordinary temperatures before the nitric oxide has time to de- 
compose. Once the ordinary temperature is attained, the de- 
composition of nitric oxide will proceed with such slowness as 
to be negligible from the practical point of view. Thus, if we 
could cool the mixture of gases instantaneously from 2500 C. 
to zero, we should preserve the 2 per cent, of nitric oxide 
which it contains at the higher temperature. Of course, in 
practice such instantaneous cooling is impossible; but if the 
gases be removed rapidly from the neighbourhood of the 
source of heat, it is possible to stave off the decomposition of 
the nitric oxide to a certain extent. 

In the McDougall and Howies process, the mixture of 
oxygen and nitrogen is passed, over a stationary electric arc, 
and the cooling is achieved by using a rapid stream of gas, 
so that the oxides are swept along in the current and carried 

1 Nernst, Application of Thermodynamics to Chemistry, p. 35. 


away from the arc as quickly as possible. An improvement l 
upon this was devised by Birkeland and Eyde, whose process 
has been utilized on a commercial scale at Notodden, in Norway. 
In this process advantage is taken of the effect which a strong 
electro-magnet exerts upon arcs. If we fix opposite the middle 
of an electric arc a strong electro-magnet, the arc will be 
diverted and finally broken ; after which it will reform be- 
tween the electrodes, be diverted and broken again. By this 
means we replace the simple arc by a series of arcs, which 
form, bend outwards, and break. If the apparatus be properly 
adjusted, it is possible by this means to produce an arc in the 
form of a thin disc instead of the usual narrow stream. The 
advantages of the Birkeland and Eyde method are twofold : 
in the first place, since the disc of the arc is thin, a gas passing 
through it is only heated for a very brief interval ; secondly, 
owing to the increase in area of the arc, it is possible to pass 
a much larger quantity of gas through the hot zone than is the 
case when the gases are passed across an ordinary arc. After 
the gases have been passed through the arc they are conducted 
into a cooling apparatus ; they are next led into an oxidizing 
chamber, where the nitric oxide becomes nitric peroxide. The 
final product of the process is basic calcium nitrate ; the basic 
salt is prepared in preference to the normal nitrate, owing to 
the fact that the latter is too deliquescent for commercial pur- 

In the Pauling process 2 an analogous type of arc is pro- 
duced without the use of an electro-magnet in the following way. 
The two electrodes are composed of metal tubing and are cooled 
by an internal stream of water. They are bent into a rough 
V-shape and placed so that the points of the V's are only about 
4 cm. apart, thus : > <. At the point where they are nearest, 
two thin iron blades jut out which approach within a couple 
of millimetres of one another ; and the arc strikes here. Below 
the V's, a strong air-blast is situated. The action of the appar- 
atus is as follows. The arc strikes between the iron points ; 
the air-blast then blows the arc in front of it and the arc runs 
along the arms of the V opposite to the jet-nozzle, growing in 

1 Birkeland, German Patent, No. 179882. 

2 See Scott, Jour. Boy. Soc. Arts, 1912,60, 645 ; also Zeitsch. Elektrochem., 
1907, 13, 225 ; 1909, 15, 544 ; 1911, 17, 431. 


area as it runs, since the points on the electrodes between 
which it plays are further and further apart as it flows up the 
arms of the Vs. Finally, the distance becomes too great for 
the current to bridge ; the arc breaks ; and a new arc is reformed 
at the iron points. 

Another modification is due to Schonherr. 1 In his process 
the arc is stationary and several metres in length. It is sur- 
rounded by a tube through which the air is driven in a spiral 
current, the spiral being produced by using a rotating jet to 
introduce the air current into the tube. In this way the mix- 
ture of oxygen and nitrogen passes alternately across the arc 
and into the cooler region on either side, so that it is subjected 
to a rapid succession of hot and cold periods. This process is 
said to be the most economical of its kind. 

3. The Manufacture of Ammonia. 

In the previous section we dealt with the processes sug- 
gested for the union of nitrogen with oxygen, and it was pointed 
out that the use of oxygen was dictated by its comparative 
cheapness ; in the present section the union of nitrogen with 
hydrogen will be considered, as, next to oxygen, hydrogen is 
the least costly element at least, when it can be utilized in 
the form of producer gas or some other such mixture. 

The work of Eegnault 2 on the spark discharge, and that of 
Donkin 3 upon the silent discharge, had shown that these two 
forms of electric energy were capable of causing the direct 
union of nitrogen with hydrogen. It was not till about ten 
years ago, however, that this process was brought into effec- 
tive touch with the technical side of the problem. In 1900, 
de Hemptinne, 4 using both the types of discharge, made an 
investigation of the proper conditions under which the action 
should be carried out. It had been proved by Eamsay and 
Young 5 that ammonia begins to decompose at a temperature 
of 500 C., and is completely broken up at a little below 800 C. ; 
so that it was to be expected that the best results would be 

1 Elektrotechnische Zeitschrift, 1909, 30, 365, 397. 

2 Kegnault, Traite de chemie, 1846. 

3 Donkin, Proc. Roy. Soc., 1873, 21, 281. 

4 De Hemptinne, Bull. Acad. roy. Belg., 1902, 28. 

5 Kamsay and Young, Trans., 1884, 45, 88. 


obtained by synthesis at comparatively low temperatures. 
This view is confirmed by de Hemptinne's results. He found 
that the yield was greatest when three conditions were ob- 
served : first, the temperature must be kept down to the point 
at which ammonia liquefies ; second, the pressure under which 
the experiment is carried out must be low ; and lastly, the gap 
across which the discharge acts must be narrow. Schlutius l 
has patented a somewhat analogous process in which, instead 
of a mixture of hydrogen and nitrogen, he employs " Dowson 
gass," which is obtained by passing a mixture of air and steam 
over glowing coal. The resulting gas has approximately the 
following composition : 

Per Cent. 

Hydrogen 14 

Nitrogen . 43 

Carbon monoxide . 39 

Carbon dioxide 4 

This mixture is acted upon by the silent electric discharge in 
presence of moist platinum, and the nature of the resulting 
compound depends upon the temperature at which the reaction 
is carried out. If the apparatus be kept below 80 C., ammonia 
is formed in accordance with the equation 

N 2 + 3H 2 + 2H 2 = 2NH 3 , aq 

On the other hand, if we allow the temperature to exceed 80 C., 
the carbon monoxide also reacts, with the formation of am- 
monium formate 

N 2 + 3H 2 + 2CO + 2H 2 = 2H . COONH 4 

We must now turn to the consideration of the action of 
catalytic agents upon the union of nitrogen with hydrogen. 
De Lambilly 2 has investigated this problem, and has taken out 
a patent for a process based upon his results. He was struck 
by the fact that the heats of formation of ammonia and am- 
monium hydrate are less than those of ammonium carbonate 
and ammonium formate. From this it follows that the ex- 
penditure of energy required to produce the salts is less than 
is necessary in the case of the free ammonia or ammonium 

1 Schlutius, English Patent, No. 2200 (1903). 

2 De Lambilly, German Patent, No. 74274. 


hydrate. De Lambilly therefore devoted himself to the pro- 
duction of the salts rather than the free substance. His process 
is as follows. Air and steam are led over white-hot coke, and 
from the mixture of gases thus produced, either carbon dioxide 
or carbon monoxide is removed, according as ammonium formate 
or bicarbonate is required. The gases are then passed through 
tubes filled with porous substances, such as platinized pumice, 
wood-charcoal, bone-charcoal, or spongy platinum. After they 
have passed once through the tubes, the gases are mixed with 
steam and again sent through the apparatus. The reactions 
which take place depend upon whether carbon monoxide or 
carbon dioxide be used 

N 2 + 3H 2 + 2C0 2 + 2H 2 = 2HO.CO.ONH 4 . . (I.) 
N 2 + 3H 2 + 2CO 4- 2H 2 = 2H.CO.ONH 4 . . (II.) 

In the case of equation (I.), it is found that the most favourable 
temperature lies between 40 and 60, while most ammonium 
formate (II.) is produced between 80 and 130. 

The main fault of the process thus outlined lies in the 
tendency of the Ammonium formate to decompose into water 
and hydrocyanic acid. In order to evade this difficulty, De 
Lambilly 1 utilized the same apparatus, but passed through it a 
mixture of carbon monoxide and ammonia, the latter being 
obtained from ammonium hydrate. In this case the best 
temperature seems to lie between 150 C. and 180 C. 

By far the most important of the modern synthetic 
methods of preparing ammonia is due to Haber and Le Kos- 
signol. 2 An examination of the equation for the reaction 

N 2 + 3H 2 ^ 2NH 3 

shows that four volumes of nitrogen and hydrogen give rise 
to two volumes of ammonia ; whence it is clear that high 
pressures will favour the conversion. A further point arises 
when we consider that the formation of ammonia in this way 
is an exothermic reaction; for this indicates that the lowest 

1 De Lambilly, German Patent, No. 78573. 

2 See Haber and van Oordt, Zeitsch. anorgan. Chem., 1905, 43, 111 ; 44, 
341 ; Haber and Le Eossignol, Zeitsch. Elektrochem., 1908, 14, 181, 513 ; ibid., 
1913, 19, 53; Haber, The Thermodynamics of Technical Gas Beactions ; 
Bernsthen, Eighth Internat. Congress of Applied Chemistry, 1912 (Abstracted 
in J. Soc. Chem. Ind., 1912, 31, 982). 


possible temperatures should be employed. But if low tem- 
peratures are used, then the velocity of combination of the 
nitrogen and hydrogen will be reduced; so it is evident that 
the reaction must be hastened by the use of catalytic agents. 
In other words, the conditions of success are : (1) high pres- 
sure ; (2) low temperature ; and (3) an efficient catalyst. 

On the technical scale, the outline of the process is as fol- 
lows. Nitrogen and hydrogen, in the proportions of one 
volume to three, and under a pressure of 200 atmospheres, are 
electrically heated to between 800 C. and 1000 C. The mixed 
gases then pass over a layer of the catalyst, which may be 
iron or uranium.* Ammonia is thus formed; and the product 
is removed by a liquefaction process. The unchanged nitrogen 
and hydrogen are then mixed with fresh supplies and are re- 
turned through the apparatus ; so that the process is a con- 
tinuous one. 

The Serpek process 1 is noteworthy in that along with am- 
monia it yields a valuable by-product. The outline of the 
method is as follows. A mixture of powdered bauxite and 
carbon is introduced into a rotating, inclined kiln ; and during 
its passage it is calcined by the action of heated gases derived 
from another stage of the process. The calcined bauxite is 
ejected from the kiln into a hopper, wherein it is mixed with 
the calculated quantity of carbon ; and from this hopper the 
mixture sinks into an inclined cylinder, one part of which is 
electrically heated to 1800 C. As the charge slides down the 
cylinder it is met and traversed by a stream of gas from a 
generator, the composition of the gas being about 30 per cent, 
carbon monoxide and 70 per cent, nitrogen. Aluminium 
nitride is collected at the lower end of the cylinder. The im- 
purities due to the bauxite appear to be oxidized and volatilized 
during the process; for the end-product is remarkably pure. 
The surplus gas is conducted upwards and serves to calcine 
the fresh bauxite. 

The reaction which takes place in the electrically heated 
zone appears to be this 

A1 2 3 + 30 + N 2 = 2A1N + 3CO 

* Osmium is a better catalyst, but is too rare for use on a commercial scale. 
1 For a fuller account of the Serpek process, see Knox, Fixation of Atmo- 
spheric Nitrogen. 


The aluminium nitride so produced can be broken up by acids 
or bases with the formation of ammonia and alumina ; and as 
the latter is pure, it is very suitable for conversion into 

A1N + 3H 2 = A1(OH) 3 + NH 3 

Temperature appears to be a weighty factor in the process ; 
and impurities also play a part as catalytic agents. 

The last method with which we need deal was originated 
by Kaiser. 1 Calcium, when heated to redness in a stream of 
hydrogen, is converted into calcium hydride ; and if this body 
is heated and nitrogen passed over it, calcium nitride is pro- 
duced. When this in turn is heated in a stream of hydrogen, 
ammonia is liberated. The process is expressed in the equa- 
tions below 

(I.) Ca + H 2 = CaH 2 
(II.) 3CaH 2 + 2N 2 = Ca 3 N 2 + 2NH 3 
(III.) Ca 3 N 2 + 6H 2 = CaH 2 + 2NH 3 

It appears to be immaterial whether the gases are allowed to 
enter the tube together or separately, the action seems to go 
on with the same ease in either case. 

4. The Synthesis of Cyanides. 

In the foregoing sections we have dealt with the prepara- 
tion of ammonia and of nitrates, both of which are naturally 
occurring substances ; the present section deals with another 
set of compounds which at first sight appear most unpromising 
from the agricultural point of view, owing to their poisonous 
character. In actual practice, however, the cyanides appear 
to offer many advantages when used as manures. 

The technical application of cyanides had led to the devising 
of methods for the production of these substances long before 
the question of nitrogen fixation became acute. The South 
African gold industry demanded a heavy output of cyanide 
for use in the cyanide process of extraction, and consequently 
a very considerable number of processes have come into exist- 
ence for the purpose of supplying this need. With these we 

1 Kaiser, French Patent, No. 350966 ; compare Haber and Oordt, Zeitsch. 
anorgan. Chem., 1905, 43, 111; 44, 341. 


have no concern here, but will confine ourselves to the chief 
methods which have been suggested within the last ten years 
or so. 

Next to oxygen and hydrogen, carbon is probably the 
cheapest element which we can employ to form compounds of 
nitrogen. Whether it be utilized in the form of coal or coke, 
it is comparatively cheap and easily procurable, so that it is 
not surprising that attempts have been made to employ it in 
the fixation of nitrogen. Before dealing with the later pro- 
cesses which depend directly upon coal, however, we must 
glance at one or two others which are based upon other sub- 

It was long ago shown by Berthelot that if a stream of 
mixed nitrogen and acetylene was raised to a high temperature, 
hydrocyanic acid was formed 

C 2 H 2 + N 2 = 2HCN 

Hoyermann 1 proposed to utilize this reaction in technical 
practice by passing a mixture of one part of acetylene and two 
parts of nitrogen over an electric arc. In this way 60 to 70 
per cent, of the acetylene was converted into hydrocyanic 
acid. The chief drawback of the method lies in the heavy 
deposits of carbon which take place on the electrodes. O'Neill 2 
employed a mixture of air with petroleum vapour or coal- 
gas, instead of acetylene and nitrogen. Gruszkiewicz 3 utilized 
" Dowson gas," which he passed over a flaming electrode. The 
best results were obtained when the percentage of carbon 
monoxide in the Dowson gas was raised to about 50. 

We must now turn to another set of methods, in which, 
instead of acetylene, metallic carbides are used. The work of 
Desfosses, 4 Fownes, 5 Bunsen and Playfair 6 showed that when 
carbon was heated with non-volatile alkalis, the resulting pro- 
ducts contained cyanogen compounds from which ammonia 
could be obtained. Attempts were thereafter made by various 
workers to utilize the reaction on a large scale, using barium 

1 Hoyermann, Chem. Zeit., 1902, 26, 70. 

2 O'Neill, Electrical World, 1902, 40, 1009. 

3 Gruszkiewicz, Zeitsch. Electrochem., 1903, 83. 

4 Desfosses, J. Chim. Pharm., 1828, 14, 280. 

5 Fownes, J. pr. Chem., 1842, 26, 412. . 

6 Bunsen and Playfair, J. pr. Chem., 1847, 42, 397. 


hydrate as the alkali, but none of them seem to have come to 
much, owing to the effect which the high temperatures required 
had upon the apparatus. 

The invention of the electric furnace, of course, gave a great 
impetus to technical research in this department, and as a 
result most of the difficulties experienced by the older workers 
have been evaded in modern practice. 

The Ampere Electrochemical Company 1 employ the follow- 
ing method. The starting materials are coke and a mixture of 
barium hydrate and carbonate. The quantities are so chosen 
as to provide an excess of coke over and above the amount 
required for the formation of barium carbide; the reason for 
this is that unless excess of coke is present, the mixture does 
not possess sufficient porosity to allow the passage of gases 
through it. The mixture is placed in a revolving electric 
furnace and the current is switched on. The action which 
takes place in the heated portion of the furnace is the forma- 
tion of barium carbide, which fuses and flows over the unaffected 
part of the coke. The revolution of the furnace now carries 
this into a colder zone, in which it cools down and solidifies. 
At this point in the revolution it encounters a stream of 
nitrogen, which converts the carbide into the cyanide 

BaC 2 + N 2 = Ba(CN) 2 

The further revolution of the furnace brings the formed cyanide 
into a cooler section, where it is withdrawn from the apparatus. 
The emptied part of the machine then passes another point, 
where it is refilled with the mixture of coke and barium salts, 
and the process continues as before. 

In this method 2 there is formed as a by-product the sub- 
stance barium cyanamide 

BaC 2 + Nj = BaN.CN + C 

This substance is a deriative of cyanamide, which in turn is 
an ammonia substitution product 

Ba : N . C=N H 2 N . C=N H 3 N 

Barium cyanamide. Cyanamide. Ammonia. 

1 Dingler's Polytech. Jour., 1903, 33, 524. 

2 French, Bericht d. V. Intern. Kongr. f. Angew. Chem., 1905, III., 727 ; 
Caro, Zeitsch. angew. Chem., 1906, 19, 1569. See also Zeitsch. angew. Chem., 
1903, 16, 520. 


On repeating the experiment with calcium salts instead of 
barium ones, it was found that the by-product of .the one 
reaction became the main product of the second; so that the 
process yielded chiefly calcium cyanamide. The substance, 
which is technically termed " nitrolim," when heated with 
water under pressure liberates ammonia, as shown in the fol- 
lowing equation : 

Ca : N . CEEN + 3H 2 = CaC0 3 + 2NH 3 

The same reaction takes place on long exposure to moist air ; 
so that nitrolim forms an excellent artificial manure. 

5. The Production of Nitrides. 

In the section on ammonia, we have already mentioned one 
of the processes by means of which nitrides have been made of 
service in the nitrogen fixation problem ; in the present section 
we may deal with another method which has been employed. 
In this department, as in some of the others, the invention of 
the electric furnace has made practicable, in technical work, 
methods which without its aid would never have been of any 
service on a large scale. 

The work of Schiitzenberger and Colson, 1 following upon 
that of Deville and Wohler, 2 showed that two nitrogen com- 
pounds of silicon could be obtained, one of which contained 
carbon in addition to the other two elements. The two sub- 
stances have the following compositions : 

Si 2 C 2 N Si 2 N 3 

From the technical point of view it does not much matter 
which of these substances is formed, or even if a mixture is 
produced. The nitride, of course, contains a greater percentage 
of nitrogen 'than the nitride-carbide, but the difference is not 
of sufficient importance to make it worth while to take special 
precautions in order to produce the pure nitride. 

Mehner has 3 patented the following process. An oxide of 
some nitride-forming element, such as silicon, boron, or mag- 
nesium, is mixed with coal or coke and submitted to the heat 

1 Schiitzenbeger and Colson, Compt. rend., 1882, 94, 1710. 

2 Deville and Wohler, Annalen, 1859, 110, 248. 

3 Mehner, German Patent, No. 67489. 


of an electric furnace. Nitrogen is then blown through the 
mass. As a result, carbon reduces the oxide, yielding the 
element itself, which is then attacked by the nitrogen to form 
the nitride 

2MgO + C = 2Mg + C0 2 

6Mg + 2N 2 = 2Mg 3 X 2 

When these nitrides are used as artificial manures, any 
weak acid in the soil, or even carbonic acid, is sufficient to 
decompose them 

Mg 3 lSr 2 + 6H 2 = 3Mg(OH) 2 + 2NH 3 

Their great advantage for agricultural purposes lies in their 
relatively high percentage of nitrogen, which is nearly double 
that found in other artificial nitrogenous substances. In this 
way the cost of carriage is reduced. On the other hand, the 
cost of manufacture seems at present to be too high to allow 
them to compete successfully with the products of other pro- 



THE minerals known as zeolites are hydrated silicates of alu- 
minium and calcium or the alkali metals. From the chemical 
standpoint they are of interest owing to the fact that the 
water-molecules in their structure appear to be very loosely 
attached to the rest of the complex, so loosely in fact, that 
they can be replaced under certain conditions by ammonia, 
alcohol, or sulphuretted hydrogen. Another property exhibited 
by the zeolites makes them of considerable technical importance. 
It is found that under proper conditions, the alkaline portion 
of their molecules can be displaced by other chemically allied 
groups. Thus from analcite, it is possible to produce the cor- 
responding silver compound 

Analcite Na 2 0, A1 2 3 , 4Si0 2 , 2H 2 

Silver derivative Ag 2 0, A1 2 3 , 4Si0 2 , 2H 2 

Since this replacement, in the case of artificial silicates, can 
be brought about by simply allowing the zeolite to remain in 
contact with a solution of a salt of the metal which it is desired 
to introduce into the silicate molecule, it is clear that the 
phenomenon may have a technical application in the removal 
of certain metals from solutions of salts. 

These artificial zeolites are known commercially as per- 
mutites; and their preparation is comparatively simple. For 
example, in one method 1 hydrated silicic acid is allowed to 
interact with an alkali aluminate ; the mixture is then heated 
to complete the reaction; and finally, calcium chloride solu- 
tion is employed to replace the alkali of the silicate by calcium. 
In this way, calcium aluminium silicates are produced which 

1 Gans, D.R.-R, 1905, 17409Y. 

33 D 


have the compositions CaO, A1 2 3 , 6Si0 2 , 6H 2 and CaO, A1 2 3 
4Si0 2 , 8H 2 0. Another method 1 depends on fusing together 
the proper quantities of sodium sulphate, carbon, kaolin, and 
quartz, whereby a sodium aluminium silicate is formed which 
has the composition 10Na 2 0, A1 2 3 , 10Si0 2 . The actual 
sodium permutite which is placed on the market has the com- 
position Na 2 0, A1 2 3 , 2Si0 2 , 6H 2 0. In order to yield the 
best results, it is essential that the permutite should be easily 
permeable by solutions. 

Now let us consider the practical applications of the per- 
mutites. When sodium permutite is allowed to remain in 
contact with a solution of, say, calcium chloride, it is found 
that calcium is removed from the solution and takes the place 
of the sodium in the permutite molecule ; whilst simultane- 
ously the sodium of the permutite goes into solution. The 
result is that calcium permutite and sodium chloride solution 
are produced. 

This at once suggests a means whereby calcium salts 
might be removed from hard water ; but if the matter stopped 
there, it would be a very expensive process. At this point, 
advantage is taken of the law of mass action. If concentrated 
brine be allowed to flow over the calcium permutite, it is found 
that, there being a greater active mass of sodium chloride 
present than there is of calcium permutite, the calcium of the 
permutite is replaced by sodium ; so that after a certain time, 
the sodium permutite is completely regenerated and is ready 
for further treatment with calcium chloride solution, from 
which it will remove calcium as before. 

The application of this to the softening of water will now 
be obvious. Suppose that the water-supply of a district con- 
tains an excess of the salts of calcium, magnesium, etc., which 
it is necessary to remove. The flow of water is directed 
through a bed of sodium permutite, which removes calcium or 
magnesium and replaces them by sodium. In this way, both 
temporary and permanent hardness are destroyed. As soon 
as the permutite bed shows signs of losing its powers, it is 
shunted out of the direct flow and a stream of 10 per cent, 
brine is passed through it. This treatment removes the 
calcium from the permutite in the form of calcium chloride 

1 Eiedel, D.E.-P. 1906, 186630 ; 1907, 200931. 


solution, so that no sludge is produced ; and at the same time 
sodium permutite is regenerated. When the regeneration is 
complete, the stream of brine is cut off; some pure water is 
run through to clear out the brine; and the permutite bed 
is re-shunted into the direct water-supply once more. There is 
practically no loss of activity in the permutite, provided that 
the water is filtered before entering the bed, so that no solid 
matter deposits itself on the permutite surface and chokes the 
pores in the material. 

The two reactions involved in the process are as follows : 

Na 2 0, A1 2 3 , 2Si0 2 , Aq + CaS0 4 

= CaO, A1 2 3 , 2Si0 2 , Aq + Xa 2 S0 4 
Sodium permutite. Calcium permutite. 

CaO, A1 2 3 , 2Si0 2 , Aq + 2NaCl = Na 2 0, A1 2 3 , 2Si0 2 , Aq + CaCl 2 

Calcium permutite. Sodium permutite. 

By a similar series of reactions, the solution of a potassium 
salt may be converted into a sodium salt solution; and this 
method is employed in the case of molasses. Since molasses 
is naturally contaminated with potassium salts which impart 
an undesirable taste to the product, the permutite process is 
utilized to replace the potassium by sodium, which is not so 

The removal of free alkali from water can also be accom- 
plished by a modification of the permutite process. Sodium 
permutite in presence of acids gives up its sodium, which is 
replaced by hydrogen. This hydrogen permutite possesses 
the property of absorbing alkali from alkaline solutions, form- 
ing sodium permutite which, by further treatment with acid, 
can be re-converted into hydrogen permutite ; so that in this 
case also the process is a cyclic one. 

Another troublesome impurity in some sources of water is 
iron. This also may be removed by the permutite process, 
though the method is more complicated than in the examples 
given above. The first step is the preparation of a manganese 
permutite, which is obtained from calcium permutite by the 
action of manganese chloride 

CaO, A1 2 3 , 2SiO 2 + MuCl a -= MnO, A1 2 3 , 2SiO a + CaCl 2 

Calcium permutite. Manganese permutite. 


The next stage in the process is the introduction of manganese 
in a higher state of oxidation ; whilst simultaneously calcium 
takes its place in the permutite molecule. This change is 
accomplished by treating the manganese permutite with 
calcium permanganate 

MnO, A1 2 3 , 2Si0 2 -f Ca(Mn0 4 ) 2 = CaO, A1 2 3 , 2Si0 2 

+ MnO, Mn 2 7 

The manganese oxides are precipitated upon the permutite, 
forming a dark-coloured mass. Now if any iron be present 
in the water, it is at once oxidized by this material ; and iron 
oxide is deposited from the solution. Apparently the action 
is partly catalytic. Special precautions are necessary to avoid 
the permutite bed becoming choked with particles of iron 

It will be seen that even when no iron is present, this 
method may still be of service ; for it enables water to be 
sterilized by means of a permanganate without any risk of 
manganese passing into the supply. Permanganate is added 
to the water ; any bacteria present are thus destroyed ; and the 
manganese is removed by passing the liquid through a per- 
mutite bed. Organic impurities may be similarly oxidized. 

The application of the permutite process to the problem of 
gold-extraction has been suggested. If a ferrous or stannous 
permutite be submitted to the action of a gold salt solution, 
the gold is removed and precipitated on the permutite surface 
in the form of purple of Cassius. 

Finally, we may refer to the utilization of permutites in 
the preparation of certain salts. Suppose that from potassium 
chloride and ammonium carbonate it is desired to produce 
potassium carbonate. A potassium permutite is formed; and 
this is then treated with ammonium carbonate solution. 
Potassium is liberated from the permutite and gives the 
required potassium carbonate. 

It will be seen that the applications of the permutites are 
far-reaching even at the present day ; and it is probable that in 
the near future they will be even further extended. 



1. Hydrogen Peroxide. 

OWING to its instability, hydrogen peroxide is only to be 
obtained in a pure state after a long series of operations. In 
the first place, a solution is prepared in the following manner. 1 
Sodium peroxide, produced by the oxidation of sodium, is 
added to 20 per cent, sulphuric acid, the reacting materials 
being kept at a low temperature during the process. Hydrogen 
peroxide and sodium sulphate are thus formed; and the 
sodium sulphate in part separates out from the solution/ 
taking with it a certain amount of water in the form of 
water of crystallization and thus leaving the solution more 
concentrated. The solid is then removed by filtration; and 
the solution is concentrated in vacuo ; or, as an alternative, 2 
alcohol may be added, which precipitates more of the sulphate 
and yields an alcoholic solution of the peroxide. This last 
method has the disadvantage that explosive ethyl peroxides 
may be formed during the process. Sometimes hydrofluoric 
acid 3 is used to decompose the sodium peroxide ; and the re- 
moval of the acid from the solution is accomplished by taking 
advantage of the low solubility of certain double fluorides. 
Thereafter the peroxide solution is concentrated, as before, by 
distillation under reduced pressure. When an 80 per cent, 
solution has been obtained, it is frozen, and practically pure 
hydrogen peroxide separates out. 4 If an absolutely pure pro- 
duct is required, it can only be prepared by a long process of 
repeated fractional distillation. 

1 Merck, D.E.-P., 152173. 

2 Dony-Henault, French Patent, 1909, 403294. 

3 Hulin-Schumann, D.R-P., 132090 ; and D.R-P., 253287. 

4 Arhle, Chem., 1909, (2) 79, 129. 



Pure hydrogen peroxide is fairly stable, and can be kept 
for a considerable time without decomposition. The solutions, 
on the other hand, are much more easily decomposed. Rough- 
ness of the walls of the containing vessel, the presence of free 
alkali, salts of heavy metals or suspended solid matter hasten 
the break-down. Colloidal metals especially have a marked 
influence upon the rate of decomposition. 1 

Hydrogen peroxide is a powerful oxidizing agent ; and it 
is therefore somewhat astonishing to find that it is capable of 
acting as a reducing agent in certain cases. Oxides of silver, 
gold, and platinum are reduced by it to the metallic state 

H 2 2 + Ag 2 = Ag 2 + H 2 -I- 2 

On the other hand, many basic oxides are converted into per- 

H 2 2 + Ba(OH) 2 = Ba0 2 + 2H 2 

Hydrogen peroxide has certain acidic properties ; for 
towards some salts it behaves as if it were an acid 

H 2 2 = Na 2 2 + C0 2 + H 2 

The constitution of hydrogen peroxide is still in dispute. 
Three formulae have been proposed 

H \ 

H 0-0-H >0 = H 0=0 H 

(I.) (II.) (III.) 

The first of these is the oldest. Its main defect is that it fails 
to express the fact that hydrogen peroxide splits off one atom 
of oxygen readily, but does not part with the other. The 
second suggestion is due to Kingzett, 2 and it expresses the 
readiness with which water is formed from the peroxide. 
Bruhl 3 is responsible for the third structure, which he found 
to agree with the refractivity of the substance better than 
either of the others. When the variety of reactions given by 
hydrogen peroxide is examined, it seems probable that the 
substance can exist in the tautomeric condition and that its 

1 Bredig, Zeitsch. physikal. Chem., 1899, 31, 258 ; 1901, 37, 1, 323 ; 1901, 
38, 122. 

2 Kingzett, Chem. News, 1884, 46, 141. 

3 Bruhl, Bar., 1895, 28, 2837. 


structure is determined by the outside reagent employed upon 
it. Possibly its constitution varies from (I.) to (II.) according 
to the conditions under which it reacts. 

2. The Metallic Peroxides. 

Oxides containing two atoms of oxygen are divisible into 
two classes : (1) those which liberate hydrogen peroxide when 
treated with acids ; and (2) those which yield no hydrogen 
peroxide. Clearly the two classes must have different consti- 
tutions ; and it is usual to distinguish them by naming the 
first type peroxides, whilst the second set are termed di- 




Barium peroxide. Manganese dioxide. 

We are here concerned only with the true peroxides. 

Since in hydrogen peroxide there are two hydrogen atoms 
replaceable by metals, it is clear that two substitution pro- 
ducts may be expected 

Me 00 H and Me 00 Me 

The mono-substitution products, termed hydroperoxides, are 
not of any technical importance. Tafel 1 prepared what ap- 
pears to be the sodium compound, Na . . . H, by shaking 
sodium peroxide with an ice-cold mixture of alcohol and a 
concentrated mineral acid (or even with alcohol alone). The 
product is a white, sandy, unstable powder which on heating 
evolves oxygen with explosive violence. The same substance 
has been obtained by Wolffenstein 2 by the action of sodium 
ethylate upon hydrogen peroxide 

C 2 H 5 .O.Na + H.O.O.H = C 2 H 5 .OH + Na.O.O.H 

Ammonium hydroperoxide is produced analogously 3 by the 
action of hydrogen peroxide upon ammonia in ethereal solu- 
tion at - 40 C. The calcium and barium compounds are 

1 Tafel, Ber., 1894, 27, 816. 

2 Wolffenstein, D.B.-P., 196369. 

3 Melikoff and Pissarjewsky, Ber., 1898, 81, 446. 


formed 1 by the action of hydrogen peroxide in excess upon 
the hydrates of the metals 


OH H.O.O.H /O.O.H 

+ = Ca< + 2H 2 

OH H.O.O.H M).O.H 

Turning now to the disubstitution products of hydrogen 
peroxide, the sodium compound is of some technical value. It 
is prepared 2 by oxidation of metallic sodium with air at a tem- 
perature of about 300 C. The air is first freed from water 
and carbon dioxide and is then driven down a heated passage. 
Up this passage the sodium is pushed, being contained in 
aluminium cases mounted on trucks ; and the method is so 
contrived that fresh sodium only meets the air-current after 
the latter has already been partially exhausted by acting upon 
half- converted sodium. In other words, fresh sodium is added 
to the line of trucks at the point furthest from where the air 
enters the passage. In this way violent reactions are avoided. 

Sodium peroxide reacts violently with water, yielding 
various hydrates. With organic matter, it forms explosive 
mixtures, though the peroxide itself seems devoid of explosive 
properties. It is employed commercially as an absorbent for 
carbon dioxide ; as an oxidizing agent ; and as a bleacher. 

"When we come to the peroxides of divalent metals, several 
formulae can be suggested to express their structure 

o /-\ 

Ba< I Ba : : Ba< >Ba 

X) \)-(K 

Since it is impossible to determine the molecular weight of 
such substances, we cannot even decide definitely whether they 
should be regarded as containing one or two atoms of the 
metal. The first of the above formulae is derived from the 
old formula for hydrogen peroxide; the second one expresses 
the ease with which the peroxides yield oxygen ; whilst the 
third structure is supported by the fact that compounds of 
the type Ba(0 . . C 2 H 5 ) 2 have been isolated. 3 

In general, the peroxides of divalent elements differ from 

1 Schone, Annalen, 1878, 192, 257. 

2 D.R.-P., 224480. 

3 Baeyer and Villiger, Ber., 1911, 44, 738. 


those of the monovalent metals in that the former can be pre- 
pared from oxides, whilst the latter can only be obtained by 
the direct oxidation of the metal. 

Barium peroxide is the most valuable representative of its 
class from the commercial standpoint. Its preparation and 
adaptation in the Brin oxygen process require no detailed 
description here. 

3. The General Character of the Per-acids and their Salts. 

Chemical nomenclature is by no means ideal; and in the 
case of the per-acids it is especially confusing, owing to a 
mingling of an older with a newer terminology. For example, 
when we speak of perchloric acid, we merely imply that this 
is an acid derived from chlorine in the highest state of oxida- 
tion which we find displayed in that class of substances. The 
name persulphuric acid, on the other hand, is given to the 
compound on account of the fact that this acid is derived from 
hydrogen peroxide by substitution. Thus the reader must be 
on his guard so as not to confuse the older " per-acids " with 
the newer type which is now regarded as the true claimant to 
the title. 

The true per-acids, then, are those acids which are derived 
from hydrogen peroxide by the substitution of an acidic radicle 
for one or more hydrogen atoms 

: B OH : B . . H 

Boric acid. Perboric acid. 

HO.S0 2 -OH HO.S0 2 -O.O.H HO.S0 2 -0.0-S0 2 .OH 

Sulphuric acid. Permonosulphuric acid. Perdisulphuric acid. 

As might be expected, the elements capable of giving rise to 
per-acids are confined to the third, fourth, fifth, and sixth 
groups of the Periodic Table. In Group III. boron stands 
alone, though lanthanum has been suspected of per-acid for- 
mation. 1 In Group IV. we find carbon, titanium, zirconium, 
and tin. In Group V. there are nitrogen, phosphorus, vana- 
dium, columbium, and tantalum; whilst in Group VI. the 
per-acid-forming elements are sulphur, chromium, selenium, 
molybdenum, tungsten, and uranium. It will be noticed that 

1 See Melikoff and Pissarjewsky, Zeitsch. anorgan. Chem., 1899, 21, 70. 


the number of elements per group increases as we go across 
the table, beginning with one in Group III. and ending with 
six in Group VI. Also, if we exclude the elements of the 
short periods, all the others which we have mentioned with 
the exception of tin and selenium belong to the even series. 
Finally, it appears that the stability of the per-acids in each 
group increases with an increase in the atomic weight of the 
parent element. 

In the succeeding sections of this chapter, a brief account 
will be given of some of the per-acids and their salts ; but for 
full details the reader must be referred to other works. 1 

4. The Persulphurie Acids and the Per sulphates. 

The history of the persulphuric acids dates back to Fara- 
day; but it is somewhat confusing, owing to the number of 
apparently contradictory data which were accumulated; and 
in this section it will be best to omit the earlier investigations 
and deal with the subject in non-chronological order. 2 

The following apparatus is used in the preparation of 
potassium persulphate: 3 A platinum dish forms the anode. 
This is filled with a saturated solution of potassium sulphate 
in dilute sulphuric acid ; and into it is dipped a porous pot 
containing dilute sulphuric acid. In the porous pot lies the 
cathode, which takes the form of a platinum wire. It is 
essential that the cathode surface should be small; for with 
a large cathode the yields are poor. 4 The platinum vessel 
forming the anode is kept cool by a stream of water passing 
over the outside. A current of about three amperes is used ; 
and after a day or two potassium persulphate begins to crys- 
tallize out. Ammonium persulphate can be obtained in an 
analogous manner. 

It appears that the stages involved in the persulphate 
formation are as follows. In the first place, the potassium 
sulphate reacts with the free sulphuric acid present to form 
potassium hydrogen sulphate. This then ionizes into H* and 

1 See especially Price, The Per-acids and their Salts. 

2 For a summary of the history of the subject, see Price, The Per-acids and 
their Salts, pp. 10 ff. 

3 Marshall, Proc. Boy. Soc. Edin., 1891, 18, 63 ; Trans., 1891, 59, 771. 
* Marshall, J. Soc. Chem. Ind., 16, 396. 


KSO' 4 . Molecular hydrogen appears at the cathode ; and the 
KSO' 4 ions join together in pairs to form K 2 S 2 8 . 

Now when sulphuric acid is added to a persulphate, we 
should expect to find the parent persulphuric acid liberated; 
but the matter is not quite so simple in practice. If dilute 
sulphuric acid be used, the resulting solution oxidizes aniline 
direct to aniline black ; but if the persulphate be dissolved in 
concentrated sulphuric acid and the liquid be then diluted, it is 
found that the addition of aniline produces, not aniline black, 
but nitrosobenzene. Evidently, two different substances can 
be formed : one being produced by the action of dilute acid on 
the persulphate ; the other being formed by the action of con^ 
centrated acid on the salt. 

It is now known that there are actually two persulphuric 
acids. The one obtained from a persulphate and dilute acid 
is termed perdisulphuric acid and has the formula H 2 S 2 08 ; 
whilst the other, produced when concentrated sulphuric acid 
is used, is termed Caro's acid or permonosulphuric acid, and 
has the composition H 2 S0 6 . 

Perdisulphuric acid can be obtained in the anhydrous state 
by allowing one molecule of hydrogen peroxide to act upon 
two molecules of chlorosulphonic acid, the mixture being kept 
cool. From this it would appear that perdisulphuric acid has 
the constitution shown below 

0-H 01 . S0 2 . OH 0-S0 2 -OH 
I + =[ +2HC1 

0-H Cl . S0 2 . OH 0-S0 2 -OH 

Caro's acid can be produced, 1 in addition to the method 
already given, by two reactions : by the electrolysis of a fairly 
concentrated solution of sulphuric acid; or by the action of 
concentrated sulphuric acid upon concentrated hydrogen per- 
oxide. Its composition was for a long time in doubt, the two 
formulae H 2 S 2 9 and H 2 S0 5 being proposed. The proof which 
enabled a choice to be made between the two suggestions is 
due to Willstatter and Hauenstein. 2 Sodium persulphate was 
treated with concentrated sulphuric acid and the resulting per- 
sulphuric acid was neutralized with sodium carbonate. The 

1 Baeyer and Villiger, Ber., 1901, 34, 853. 

2 Willstatter and Hauenstein, Ber., 1909, 42, 1839. 


sodium salt of Caro's acid was thus formed. When this was 
treated with benzoyl chloride in presence of sufficient sodium 
hydroxide to keep the solution neutral, a benzoyl derivative 
of Caro's acid was formed. Analysis of this proved that for 
each sulphur atom in the compound there was one benzoyl 
group present. Now if the formula of the acid be H 2 S 2 9 , 
the benzoyl derivative will contain two benzoyl radicles one 
for each sulphur atom according to the analysis. Its formula 
will therefore be (C 6 H 5 . CO) 2 S 2 9 , and it should contain no 
metallic atom. If Caro's acid be H 2 S0 5 , on the other hand, 
there is only one sulphur atom present; and therefore the 
benzoyl compound can contain only a single benzoyl radicle. 
Its formula would thus be C 6 H 6 . CO . S0 5 . H ; and it should be 
capable of forming a potassium salt. The product does actu- 
ally form a potassium salt C 6 H 6 . CO . S0 5 . K, which estab- 
lishes the formula of Caro's acid as H 2 S0 5 . Its constitution 
is probably expressed by : HO . S0 2 . . OH, which evidently 
represents a molecule of hydrogen peroxide wherein one 
hydrogen atom is replaced by a sulphonic radicle. , On this 
account, Caro's acid is correctly termed permowosulphuric acid. 
The constitution of ordinary " persulphuric acid," H 2 S 2 8 , has 
been shown to be analogous to that of Caro's acid. It also is 
derived from hydrogen peroxide by replacing hydrogen by the 
sulphonic group; but in this case both the hydrogen atoms 
are replaced : HO S0 2 .0.0. S0 2 . OH. The acid is there- 
fore known as perdisulphuric acid. 

The relations between perdisulphuric acid and its decom- 
position products were not made clear until a considerable 
amount of research had been carried out, most of which, at 
the time, appeared only to make confusion worse than before. 
It was finally established that the facts of the case were as 
follows. When perdisulphuric acid is freshly prepared by the 
electrolytic method it exists alone in solution ; and hence the 
solution containing it shows the reactions of pure perdisul- 
phuric acid. For example, it bleaches indigo and oxidizes 
ferrocyanides ; but with chromic acid and potassium perman- 
ganate it shows none of the reactions attributable to hydrogen 
peroxide. The ratio between "sulphuric acid formed" and 
"active oxygen liberated" in its reactions is 2 : 1. If the 
solution be allowed to stand for a time before being tested, 


it is found that its character alters. In any reaction to 
which it is subjected, the ratio of "sulphuric acid formed" to 
" active oxygen liberated " is now 1 : 1. Finally, on long 
standing, or in presence of colloidal metals, the solution under- 
goes yet another change. Hydrogen peroxide is formed, 
which reacts with the remaining acid, giving sulphuric acid 
and free oxygen. These various reactions are represented in 
the following equations : 

H 2 S 2 8 + H 2 = H 2 S0 4 + H 2 S0 5 
H 2 S0 5 + H 2 = H 2 S0 4 + H 2 2 
H 2 S0 5 + H 2 2 = H 2 S0 4 + H 2 + 2 

The persulphates derived from perdisulphuric acid are 
much more stable than the parent substance, those which 
crystallize anhydrous being less liable to decomposition than 
those which contain water of crystallization. All of them are 
soluble in water. In solution, they gradually decompose, like 
the parent acid, yielding hydrogen sulphates and oxygen 

K 2 S 2 8 + H 2 = 2KHS0 4 + 

The most striking property of the persulphates, however, 
is their oxidizing power. A persulphate liberates free halogen 
from halides 

K 2 S 2 8 + 2KI = 2K 2 S0 4 + I 2 

and it is found that this reaction is catalysed by the presence 
of iron or copper salts. 1 Ferrous salts, when treated with a 
persulphate, are oxidized to the ferric state ; and this reaction 
forms a convenient method for persulphate estimation, in con- 
junction with the usual permanganate titration. The oxida- 
tion of manganese salts to permanganates forms a simple mode 
of detecting the presence of a persulphate, since the colour of 
the permanganate can be seen in extremely dilute solutions. 

One of the most interesting reactions of the persulphates 
is that shown when their solutions act upon metals. 2 No 
marked evolution of gas occurs ; and the action appears to 
take the course shown below 

K 2 S 2 8 + Zn = K 2 S0 4 + ZuS0 4 

1 Price, Zeitsch. physikal Chem., 1898, 17, 459. 

2 Marshall, J. Soc. Chem. Ind., 1897, 16, 396 ; Tarugi, Gazzetta, 1903, 33 
(i.), 127 ; Turrentine, J. Physical. Chem., 1907, 11, 623 ; Levi, Migliorini, and 
Ercoliiii, Gazzetta, 1908, 38 (i.), 583. 


Magnesium forms an exception to the general rule ; for when 
it is treated with ammonium persulphate solution a violent 
reaction takes place and ammonia is evolved. 

The persulphates have found a most extensive application 
as oxidizing agents. In analytical work they are employed in 
iron and steel analyses for the estimation of manganese or 
chromium ; whilst in organic chemistry their uses are too 
numerous to mention, Caro's acid, especially, has proved of 
value in organic reactions, as it is one of the most easily re- 
gulated oxidizing agents which we possess. 

5. The Per-acids of Group V. 

Neither of these acids has been isolated in a pure state; 
and the evidence for their existence is therefore indirect. 

When nitrogen pentoxide interacts with anhydrous hydro- 
gen peroxide, the mixture being kept well cooled, a colourless 
liquid smelling like bleaching-powder is obtained. 1 The 
aqueous solution of this substance liberates iodine from potas- 
sium iodide and oxidizes aniline to nitrosobenzene. In dilute 
solutions, the compound is hydrolysed, hydrogen peroxide and 
nitric acid being liberated. The reaction may be assumed to 
take the following course : 

N 2 5 + H 2 2 = HN0 3 + HO.O.N0 2 

A solution having the same properties is produced when a very 
dilute aqueous solution of nitrous acid is oxidized with hydrogen 
peroxide 2 

H 2 2 + HN0 2 = HN0 4 4- 2H 2 

Further evidence on the possibility of the existence of pernitric 
acids is to be found in the results of electrolysing silver salts ; 3 
but up to the present it is the subject of dispute. 

1 d'Ans and Friederich, Zeitsch. anorgan. Chem., 1911, 73, 344 ; d'Ans, 
Zeitsch. Elektrochem., 1911, 17, 850. 

2 Raschig, Zeitsch. angew. Chem., 1904, 17, 1419 ; Ber., 1907, 40, 4585. 

3 See Berthelot, Compt. rend., 1880, 90, 653 ; Sulc, Zeitsch. anorgan. Chem., 
1896, 12, 89, 180; 1900, 24, 305; Mulder and Heringa, Rec. trav. chim., 1896, 
15.. 235 ; Mulder, ibid., 1897, 16, 57 ; 1898, 17, 129 ; 1899, 18, 91 ; 1903, 22, 385, 
388, 405; Tanatar, Zeitsch. anorgan. Chem., 1901, 28, 331; Bose, Zeitsch. 
anorgan. Chem., 1905, 44, 237 ; Watson, Trans., 1906, 89, 578 ; Brauner and 
Kuzma, Ber., 1907, 40, 3371; Luther and Pokorny, Zeitsch. anorgan. Chem., 
1908, 57, 290 ; Baborovsky and Kuzma, Zeitsch. physical Chem., 1909, 67, 48 ; 
Bose, ibid., 1909, 68, 383. 


The evidence for the existence of perphosphoric acid is 
stronger. It is not obtained by the action of hydrogen per- 
oxide upon phosphoric acid, nor by the electrolysis of phosphoric 
acid or the phosphates ; but when 30 per cent, hydrogen per- 
oxide is treated with phosphorus pentoxide, meta- or pyro- 
phosphoric acid, a solution is obtained which possesses strong 
oxidizing properties which cannot be accounted for merely by 
the presence of hydrogen peroxide. 1 The diluted solution will 
even oxidize manganese salts to permanganic acid. It appears 
probable that H 3 P0 5 is present. 

Pervanadic acid, percolumbic acid, and pertantalic acid are 
all known. The two first are obtained by the action of hydro- 
gen peroxide upon the corresponding lower acid, whilst per- 
tantalic acid can be prepared from its potassium salt by the 
action of sulphuric acid. 

6. The Per carbonates. 

Potassium and rubidium percarbonates are obtained by 
electrolysing concentrated solutions of the corresponding car- 
bonates in an apparatus similar to that employed in the case 
of the persulphates. 2 The temperature must be kept low ; and 
hence the method is inapplicable to carbonates which do not 
form strong solutions at zero or below it. 

The course of the reaction appears to be similar to that 
which is observed in the persulphates. Potassium carbonate 
dissociates into the ions K* and KCO' 3 ; and then two of the 
latter unite to form K 2 C 2 6 . From this it seems probable that 
the constitution of potassium percarbonate is that which is 
expressed below * 


1 + 1=1 I 

CO-0 -0 CO CO-0 0-CO 

An examination of this formula will show that it is derived 
from hydrogen peroxide by replacing the hydrogens by two 

1 Schmidlin and Massini, Ber., 1910, 43, 1162 ; d'Ans and Friederich, Ber., 
1910, 43, 1880 ; Zeitsch. anorgan. CTiew.,1911, 73, 343 ; d'Ans, Zeitsch. Elektro- 
chem., 1911, 17, 850. 

2 Constam and Hansen, Zeitsch. Elektrochem., 1896, 3, 137. 


carbonic acid nuclei and then forming the di-potassium salt of 
the acid thus produced. In other words, this per-acid is, pro- 
perly speaking, perdicarbonic acid, analogous to perdisulphuric 

The action of 3 per cent, hydrogen peroxide upon sodium 
carbonate gives rise to a substance 1 having the composition 
Na 2 C0 4 , liH 2 or Na 2 C0 3 , H 2 2 , iH 2 0. Whether this is a true 
percarbonate or merely, as the second formula suggests, a car- 
bonate with hydrogen peroxide of crystallization, is not clearly 
established. True per-acids liberate iodine from potassium 
iodide instantaneously ; but Tanatar's salt is almost without 
immediate effect upon iodides. The evidence, therefore, is 
against it being a true percarbonate. 

The action of carbon dioxide upon sodium peroxide and 
sodium hydroperoxide is interesting. Either one or two 
molecules of carbon dioxide can react with the peroxide, giving 
the following results : 

(1) Na 2 2 + C0 2 = ]STa 2 C0 4 

(2) Na 2 2 + 2C0 2 = Na 2 C 2 6 

At first sight, it seems probable that the compound ]STa 2 C 2 O ti is 
derived from the same acid as K 2 C 2 6 which is prepared by 
the electrolytic method. An examination of the effect of the 
two substances on potassium iodide shows that this is not the 
case. The electrolytically prepared salt liberates iodine quan- 
titatively, according to its nature as a true per-acid : but the 
sodium compound prepared from sodium peroxide only liberates 
half the theoretical quantity of iodine. Since the potassium 
salt has the structure (L), the only probable constitution for 
the sodium salt is (II.) 

OK OK 0-0 Na ONa 

CO-0 CO CO 0-CO 

(I.) (II.) 

The compound Na 2 C0 4 appears to be sodium perinonocarbonate, 
analogous to sodium permonosulphate 

NaO . . CO . . Na NaO . . S0 2 . ONa 

The action of carbon dioxide upon sodium hydroperoxide 

1 Tanatar, Ber., 1899, 22, 1544. 


also takes two courses, though in this case only one molecule 
of carbon dioxide is involved, and either one or two molecules 
of hydroperoxide may take part in the reaction 

Na . . . H + C0 2 = NaHC0 4 
2JSTa . Of . . H + C0 2 = Na 2 C0 5 , H 2 

The two products of the reactions are apparently compounds 
formed by the addition of hydrogen peroxide to the substances 
Na 2 C0 4 and Na 2 C 2 6 which were produced in reactions (1) and 
(2) above. 

The properties of potassium percarbonate may now be de- 
scribed. 1 It is a sky-blue * hygroscopic powder, which, when 
gently warmed, decomposes according to the equation 

2K 2 C 2 6 = 2K 2 C0 3 + 2C0 2 + 2 

Its solution at ordinary temperatures liberates hydrogen per- 
oxide and hence produces oxygen 

K 2 C 2 6 + 2H 2 = 2KHC0 3 + H 2 2 

The presence of potassium hydroxide accelerates the reaction, 
the decomposition taking place at zero with the production 
of hydrogen peroxide and potassium carbonate. With dilute 
sulphuric acid in the cold, hydrogen peroxide and carbon dioxide 
are liberated quantitatively 

K 2 C 2 6 -f H 2 S0 4 = K 2 S0 4 + 2C0 2 + H 2 2 

Potassium percarbonate is a good oxidizing agent in certain 
cases, such as the oxidation of lead sulphide to lead sulphate ; 
but it can also play the part of a reducing agent. For ex- 
ample, when it is treated with manganese dioxide, lead dioxide 
or silver oxide, oxygen is evolved and the corresponding metallic 
carbonate is formed. 

In technical practice, potassium percarbonate is utilized for 
bleaching purposes and also as a " hypo-killer " in photography 
under the name of "Antihypo." Sodium permonocarbonate 
is utilized as a source of hydrogen peroxide and also in medicine 
as an antiseptic. 

1 Wolffenstein and Peltner, Ber., 1908, 41, 280. 
* Rubidium percarbonate is colourless. 


7. The Perborates. 

Perboric acid, HB0 3 , does not exist in the free state ; but 
many of its salts are known ; and of these, from the com- 
mercial standpoint, sodium perborate is the most important. 

Sodium perborate 1 is manufactured by any of the three 
reactions represented below which take place in aqueous 

(1) 2NaOH + Na 2 B 4 7 , 10H 2 + 4H 2 2 4- H 2 

= 4NaB0 3 + 16H 2 

(2) Na 2 2 + H 3 B0 3 + HC1 + 2H 2 = NaB0 3 , 4H 2 + NaCl 

(3) Na 2 2 + H 2 2 + 2H 3 B0 3 = 2tfaB0 3 , 4H 2 

The possibility of preparing perborates by electrolysing con- 
centrated solutions of orthoborates appears to be still in doubt. 
Tanatar, 2 Pouzenc, 3 Bruhat and Dubois 4 claim that this 
method can be used ; but Constam and Bennet 5 state that no 
perborates are thus formed. 

When preserved with customary precautions, sodium per- 
borate decomposes only slowly, though the breakdown may be 
hastened by the presence of impurities. At times, however, 
it is liable to an unexplained type of decomposition which 
may take place after the substance has been kept quite un- 
affected for months. In this reaction, the perborate liquefies 
and froths ; and its behaviour resembles to some extent that 
of decomposing hydrogen peroxide. 

When sufficiently diluted, aqueous solutions of sodium 
perborate behave almost as if they contained free hydrogen 
peroxide, free boric acid, and free sodium hydrate ; and since 
alkaline solutions of hydrogen peroxide are readily decom- 
posed even at ordinary temperatures, it is clear that any 
reaction involving sodium perborate must be carried through 
rapidly before the evolved oxygen escapes. 

The commerical applications of sodium perborate depend 
upon its power of bleaching. In the laundry industry it has 

1 Tanatar, Zeitsch. physikal. Chem., 1898, 25, 265 ; 1898, 26, 132, 451 ; 
1899, 29, 152 ; Melikofi and Pissarjewsky, Ber., 1898, 31, 678 ; Jaubert, Compt. 
rend., 1904, 139, 796 ; Bruhat and Dubois, ibid., 1905, 140, 506. 

2 Tanatar, Zeitsch. physikal Chem., 1898, 26, 132. 

3 Pouzenc, French Patent, 411258. 

4 Bruhat and Dubois, Compt. rend., 1905, 140, 506. 

5 Constam and Bennet, Zeitsch. anorgan. Chem., 1900, 25, 265. 


gained a considerable vogue ; and it has been found more satis- 
factory than hydrogen peroxide in the bleaching of human 
hair, since its alkaline nature enables it to remove grease and 
thus ensures a more uniform effect. In the bleach-works it 
is largely used ; and it is generally employed here along with a 
catalyst such as the oxides of manganese and lead. 

The constitution of sodium perborate is still a matter of 
doubt. Four formulas may be suggested 

Na . . B/ | Na . . . B : NaB0 2 , H 2 2 

(I.) (II.) (HI.) 

Na.<\ /O.Na 

>B . . . B< 
H.CK M).H 


Formula (III.) is evidently a meta-borate with hydrogen per- 
oxide of crystallization ; and it should, by analogy with other 
similar compounds, lose hydrogen peroxide on heating. In 
actual practice, however, when sodium perborate is heated to 
60 C. under diminished pressure, water is lost but no hydrogen 
peroxide is found in the distillate. 1 Also, ammonium per- 
borate exists as NH 4 B0 3 , which throws doubt on the presence 
of hydrogen peroxide in the sodium salt. Formula (IV.) is 
also excluded ; as a compound of this type should give rise to 
a salt Na 4 B 2 6 on treatment with excess of alkali. No such 
reaction is known. 

This leaves us with the choice between (I.) and (II.) In 
favour of the second formula we have an argument by analogy 
from other per-salts the constitution of which is known. The 
first formula is supported by the fact that a stable potassium 
compound is known which has the composition KB0 4 , H 2 
and which may be regarded as a salt of potassium hydroper- 
oxide with perboric acid. This substance is remarkably stable 
and gives off only traces of oxygen when dissolved in water. 
This behaviour is not remarkable if perboric acid has the 
formula (I.) ; but if perboric acid is constituted as in (II.) the 
potassium salt would contain a chain of three oxygen atoms, 

* Bosshard and Zwicky, Zeitsch. angew. Chem., 1912, 25, 993. 


which, according to our normal views, would be unstable. 
The two possible formulae of the potassium salt are shown 

K.O.O.B | K.O.O.O.B:O 


Derived from perboric acid (I.). Derived from perboric acid (II.). 

The reader must decide for himself which formulation he 



1. The Production of Active Nitrogen. 

THE fact that some of the elements can exist in more than one 
form has been common knowledge for years; but it is only 
recently that nitrogen has been proved to be capable of exist- 
ing in allotropic modifications; and the results obtained 
during the investigation are of considerable interest since they 
display nitrogen in the guise of a. very reactive element. 

When electric discharges are passed through gases under 
low pressure it is sometimes found that the gases continue to 
glow for a time after the discharge has been interrupted ; and 
Strutt l has proved that this effect in the case of air is due 
to the mutual reaction of nitric oxide and ozone which are 
formed during the discharge. Similar phosphorescent com- 
bustions were observed in ozone when sulphur, sulphuretted 
hydrogen, acetylene, or iodine is present. 2 Pure nitrogen 
gives no after-glow when the simple induction coil discharge is 
employed ; 2 but if a jar discharge and a spark-gap be utilized 
the glow is observed as in the other cases. 

The practical details of the apparatus for producing active 
nitrogen are simple in the extreme. 3 Carefully purified nit- 
rogen from a gas-holder is drawn in a slow stream into a 
tube in which two electrodes are placed. The discharge 
between the electrodes activates the nitrogen, and the gas is 
then drawn on by the flow into another part of the tube, at 
which point it can be observed and its properties ascertained 

1 Strutt, Phys. Soc. Proc., 1910, 23, 66. 

2 Ibid., 1910, 23, 147. 

3 Ibid. ; Proc. Boy. Soc., 1911, A, 85, 219 ; 1913, A, 88, 539 ; 1915, A, 91, 



without any interference from the electric discharge. The 
flow of gas is maintained by means of a mechanical pump. 

When it is desired to test the effect of active nitrogen 
upon any other substance, the latter is placed in the observa- 
tion part of the tube so that the stream of active nitrogen 
flows over it in passage. 

2. The, Properties of Active Nitrogen. 

(a) Action on Non-metals. When active nitrogen comes 
in contact with the vapour of yellow phosphorus, some re- 
action occurs ; for the nature of the glow changes and red 
phosphorus is deposited as a film upon the surface of the tube. 1 
Closer investigation revealed the fact that the reaction is not 
immediate; the nitrogen reacts with the phosphorus mainly 
after the glow has died down. 2 Sulphur reacts with a blue 
flame or glow ; and a green material is deposited upon the glass. 1 
Apparently this is mainly nitrogen sulphide. 2 Selenium is 
unacted upon. 1 Arsenic shows a green glow during the re- 
action 1 and yields arsenic nitride. 2 Antimony, carbon, and 
hydrogen gave no results. 1 Oxygen destroys the glow of the 
nitrogen without producing any other luminosity. 1 Iodine 
exhibits a very marked light blue flame when its vapour 
mingles with active nitrogen. 

(b) Action on Metals. When glowing nitrogen is passed 
over a pellet of metallic sodium heated a little above its melt- 
ing-point, the sodium spectrum is shown brilliantly; and 
similar results were observed with other metallic elements. 
The phenomenon is due to the metal burning in the active 
nitrogen to form the nitride ; 3 and the spectrum is simply the 
flame spectrum of the metal. 4 Similar results were observed 
with zinc, cadmium, mercury, potassium, magnesium, and 
lead. 1 ' 2 In the case of mercury an explosive compound was 
produced. 1 

(c) Action on Compounds. When ammonia is mixed 
with active nitrogen the glow is extinguished, but no chemical 

1 Strutt, Proc. Roy. Soc., 1911, A, 85, 219. 

2 Ibid., 1913, A, 88, 539. 

3 Ibid., 1912, A, 88, 539. 

4 Ibid., 1911, A, 85, 219 ; Trans., 1918, 113, 200. 


effects have been observed. 1 Nitric oxide acts in a fashion 
quite different from what might have been anticipated. A 
greenish-yellow flame is observed ; and nitrogen peroxide is 
produced. 2 Why active nitrogen should convert nitric oxide 
into the more highly oxygenated derivative nitrogen peroxide 
is not yet understood. Copper oxide destroys the glow of 
active nitrogen abruptly ; but no chemical change was trace- 
able ; and it appears probable that the action is a catalytic one. 3 

Various salts were examined ; but the observations are 
limited to their effects upon the glow of the active nitrogen. 3 ' 4 
Stannic chloride and titanium tetrachloride give solid products 
containing nitrogen. 

Turning to the action of allotropic nitrogen upon the 
carbon compounds, a considerable amount of data no's been 
accumulated. 1 ' 2> 8 Acetylene yields cyanogen ; and a similar 
result was observed in the cases of ether, ethyl iodide, carbon 
tetrachloride, chloroform, bromoform, ethylene chloride, and 
ethylidene chloride. Benzene is much less reactive than 
acetylene, as might have been expected. Carbon disulphide 
produces a blue polymeric nitrogen sulphide and a polymeric 
carbon monosulphide. (Sulphur chloride yields ordinary 
yellow nitrogen sulphide. 5 ) Methane, hexane, and heptane are 
attacked with the formation of hydrocyanic acid. 

(d) Spectrum. The spectrum of active nitrogen is different 
from any other nitrogen spectrum. It contains in the visible 
a green, a yellow, and a red band. 6 

(e) Electrical Properties. Active nitrogen has a marked 
conductivity. Apparently a current of the gas displays a 
behaviour from the electrical standpoint which brings it into 
line with a salted Bunsen flame. 7 

3. The Influence of Impurities. 

At one time doubt was thrown upon Strutt's statement 
that pure nitrogen could give rise to the active form ; and it 

1 Strutt, Proc. Roy. Soc., 1911, A, 85, 219. 2 Ibid., 1911, A, 86, 56. 

3 Ibid., Proc. Boy. Soc., 1911, A, 85, 219. 4 Ibid., 1915, A, 91, 303. 

5 Ibid., 1913, A, 88, 539. 

6 Ibid., 1911, A, 85, 219 ; Fowler and Strutt, ibid., 377 ; Strutt and Fowler, 
ibid., 1911, A, 86, 105. 

1 Ibid., Proc. Roy. Soc., 1911, A, 86, 56; ibid., 1912, A, 87, 179. 


was suggested 1 that in order to obtain the phenomena de- 
scribed by him, a trace of oxygen was necessary. Strutt, 
however, showed that oxygen, instead of being advantageous, 
was actually deleterious, since the presence of 2 per cent, of it 
in nitrogen was sufficient to prevent the production of the 
glow. 2 None the less, the controversy continued. 3 

Koenig and Elod 4 suggested that possibly the other in- 
vestigators were at cross purposes in the matter owing to a 
lack of distinguishing between two different phenomena : the 
Strutt glow on the one hand and the chemical effects on the 
other. It seemed possible that these two effects were not 
necessarily co-existent ; and since Tiede and Domcke used the 
presence of the glow as a test for active nitrogen, it seemed 
possible that misunderstandings had arisen in this way. 

The main question was settled by Tiede and Domcke 
bringing their apparatus over to London so that the two pro- 
cesses could be tried side by side. The contradictory results 
were proved to be due to differences in the apparatuses ; and 
it was found that active nitrogen can be produced even when 
no oxygen is present. On the other hand, minute traces of 
oxygen appear to have a favourable effect upon the process. 5 ' 6 

The discovery of this catalytic effect of oxygen led Strutt 6 
to a study of the influence of other substances upon the pro- 
duction of active nitrogen ; and he established the fact that 
minute admixtures of other materials are capable of perform- 
ing the same function as oxygen itself. The following sub- 
stances all gave results, the most effective ones being placed 
first. Hydrogen sulphide, water, carbon dioxide, carbon mon- 
oxide, the three hydrocarbons, acetylene, ethylene, and methane, 
oxygen, mercury, chlorine, hydrogen, and the two inert gases, 
argon and helium. It was established that although the 
purest nitrogen will yield active nitrogen, yet in order to ob- 
tain the active farm in any abundance it is necessary to have 
about 0*1 per cent, of some foreign gas present. 7 Strutt 

Comte, Physikal. Zeitsch., 1913, 14, 74 ; Tiede, Ber. t 1913, 46, 340. 

Strutt, Physikal. Zeitsch., 1913, 14, 215 ; cf. Koenig and Elod, ibid., 165. 

Tiede and Domcke, Ber., 1913, 46, 4095. 

Koenig and Elod, Ber., 1914, 47, 516. 

Baker, Tiede, Strutt, and Domcke, Nature, 1914, 93, 478. 

Strutt, Proc. Roy. Soc., A, 1915, 91, 303. 

Cf. Strutt, Trans., 1918, 113, 200. 


suggests that the impurity acts by loading the electrons in the 
discharge and thus altering the character of their impact upon 
the nitrogen molecules. 

4. The Effect of External Conditions upon Active Nitrogen. 

The influence of temperature upon active nitrogen is v6ry 
well marked. When the portion of the observation tube 
through which the gas passes is heated, the glow of the nitro- 
gen is locally extinguished and only revives when the gas has 
gone further into a cooler part of the tube. From this, Strutt 
deduced that the change which produces the after-glow is ac- 
celerated by heating and retarded by cooling. 1 

Further investigation, however, threw doubt upon this 
view. 2 When a specimen of active nitrogen was sealed up 
in a bulb and excited by an electrodelees discharge, it was 
found that at ordinary temperatures the glow lasted for about 
a minute. A similar experiment carried out with the bulb 
cooled to liquid air temperature proved that the glow was ex- 
tinguished in fifteen seconds. Whilst the bulb was under the 
surface of the liquid air, the glow was very brilliant. 

Curiously enough, when the bulb was plunged in boiling 
water the duration of the glow was also shortened. 

Strutt then suggested that there were two factors to be 
taken into account. One of these was the direct temperature 
effect on the glow -transformation ; the other was the destruc- 
tive effect of the walls of the vessel, which had a catalytic 
action upon the phenomenon. The destructive effect would 
be increased by rise in temperature. 

On this hypothesis, the results are interpreted as follows : 
When cooled to - 180 C. the catalytic action of the glass 
is somewhat checked ; but, in compensation, the glow-trans- 
formation in the gas occurs more rapidly at low temperatures 
and the glow is given out with greater intensity during a 
briefer period. On the other hand, when the bulb is heated 
too 100 C. the catalytic effect of the glass is increased and its 
influence overbears the retardation of the transformation due 
to the higher temperature. 

1 Strutt, Proc. Boy. Soc., 1911, A, 85, 219. 

2 Ibid., 1911, A, 86, 262. 


If this hypothesis be correct, the transformation of active 
nitrogen is exothermic, but it is also unique among known 
reactions in that it proceeds more rapidly at low temperatures 
than at high ones. It must be emphasized that this is not yet 
more than a hypothesis ; but if it should prove correct it will 
have a considerable effect upon our ideas of reaction-conditions. 

Turning to another influence which might be supposed to 
be capable of affecting active nitrogen, it has been shown by 
Strutt l that an electric field has no action upon the properties 
of the gas. Neither the glow nor the chemical activity of the 
nitrogen seemed to be influenced. 

5. A Third Form of Nitrogen. 

Lowry 2 has detected the existence of nitrogen in a form 
which differs apparently both from ordinary nitrogen and from 
Strutt's active nitrogen. A current of air was passed through 
an ozonizer and then across a series of spark-gaps; and the 
gases were then examined in a spectroscope cell 64 feet in 
length, by means of which method the presence of minute 
traces of nitrogen peroxide could be detected. 

It was found that when the ozonizer or the spark-gaps were 
used singly, no trace of nitrogen peroxide was detectable ; but 
when the air from the ozonizer was passed over the spark-gap, 
the spectrum of nitrogen peroxide could readily be seen in the 

From these results it appears that the ozonizer and the 
spark-gap taken separately do not activate the nitrogen of the 
air ; but the successive action of them upon the stream of gas 
results in the formation of an active form of nitrogen which 
reacts with the ozone to produce nitrogen peroxide. 

The active form of nitrogen thus produced is not identical 
with Strutt's variety, since^ the latter is not oxidized by ozone. 

1 Strutt, Proc. Boy. Soc. t 1911, A, 85, 219 ; 1912, A, 89, 56. 

2 Lowry, Trans., 1912, 101, 1152 ; Phil Mag., 1914, 28, 412. 



1. General. 

THE physical properties of chemical compounds can be divided 
into two categories: colligative properties and non-colligative 
properties. The first of these groups comprises properties 
which are dependent purely upon the number of molecules 
present in the specimen we are examining ; and of this type a 
good example is afforded by the pressure which a gas exerts 
at constant temperature. Non-colligative properties are in 
turn divisible into two sections : additive properties and con- 
stitutive properties. It is difficult to draw a hard-and-fast 
line between these two sections, for most properties are partly 
additive and partly constitutive. When we speak of an additive 
property, we usually mean one in which the number of atoms 
of each element in the molecule is the preponderating factor ; 
while we regard as constitutive properties those in which the 
additive properties are masked by the greater influence of the 
mode of linkage of atoms in the molecule. 

The study of additive properties is a much more simple line 
of research than the examination of constitutive properties. 
For instance, refractive index appears to be an additive relation, 
and its laws are already marked out in their main outlines ; 
but when we take dielectric constant which is simply the 
refractive index for waves of infinite length the constitutive 
factor overwhelms the additive one, and up to the present time 
no definite relation has been traced between structure and 
dielectric constant. 

Of all the constitutive properties of chemical compounds, 
the most complex appear to be absorption spectra. 1 A few 

1 For a full account of absorption spectra, see Smiles' The Relation between 
Chemical Constitution and Physical Properties, 1910. 



rough rules, some of which, like Armstrong's quinonoid hypo- 
thesis, are either mere words or obviously incorrect, are all that 
we have to guide us at the present time in tracing a connection 
between what is called " colour " and the chemical constitution 
of compounds. Much time has been wasted in discussing the 
question of physiological colour, a phenomenon of very little 
scientific value, since it depends upon the crude perception of 
the human eye, which can be thrown out of gear by a dose of 
santonine, and which, at best, excludes more of the spectrum 
than it perceives. But even our finest instruments to-day 
include within their range only a portion of the complete 
electrical, thermal, and luminous spectrum of any substance. 

When we look at the sun, we see what is called the visible 
solar spectrum, that is, white light. If we hold up between 
our eye and the sun a beaker containing ferric chloride solution, 
we shall see a yellow colour. Now, it is obvious that the ferric 
chloride solution has added nothing to the sun's light ; so if 
it has produced a change in the light which reaches our eye, it 
is clear that the solution must have abstracted something from 
the solar spectrum. What actually happens is that some of 
the rays of the spectrum are absorbed in the solution, so that 
instead of the complete solar spectrum, we now see what is 
called the absorption spectrum of ferric chloride. 

Suppose that instead of merely looking through the beaker, 
we place it in front of a spectroscope and pass the sun's light 
through the ferric chloride solution. By looking through the 
spectroscope we shall then find that, instead of the whole solar 
spectrum being transmitted, some of the lines are absent, 
having been absorbed by the ferric chloride. By comparison 
with the usual solar spectrum we could easily determine 
exactly which lines have been absorbed, and in this way we 
should have a more or less accurate idea of the absorption 
spectrum of ferric chloride. In practice, however, this method 
suffers from many drawbacks, and instead of it we use the 
following arrangement : 

In Fig. 1 the apparatus is represented diagrammatically. 
E E are the electrodes of an arc light. C is a cell containing 
the solution whose absorption spectrum is to be measured. 
S i the slit of the spectroscope. P is the prism. is a 
photographic camera, containing a plate, Q. The electrodes, 


E E, may be of any metal or alloy which gives a well distributed 
series of lines in its spectrum ; an iron arc is sufficient for 
most purposes. The cell, C, nrust have quartz ends, through 
which the light of the arc may pass to the slit S. If glass 
ends were used, a very large proportion of the invisible ultra- 
violet part of the spectrum would be cut off; for though glass 
is transparent to the visible spectrum, it is almost 
opaque with respect to the ultra-violet region. 
The spectrum of the arc is thus thrown upon the 
plate P, and by means of a diaphragm placed 
before the slit it is reduced to a narrow band. 
With the usual form of photographic spectroscope, 
the band is about 5 inches long and a quarter of 
an inch wide. 

Before we can utilize the spectroscope for study- 
ing absorption spectra, however, it is necessary to 
prepare what we may call a standard plate. That 
is to say, we must have a photograph of the pure 
arc spectrum, and be able to identify the lines 
upon it. Suppose we are working with an iron 
arc. We first photograph the spectrum of the arc 
pure and simple, removing the cell C. This 
negative forms the basis of our standard plate. 
We then smear some salt, say lithium 
chloride, upon the arc, and take another 
photograph of the spectrum. In this new 
photograph we shall have not only 
the iron spectrum, but in addition 
to it the characteristic lines of the 
lithium spectrum. A comparison 
of the two negatives will enable 
us to detect the lithium lines. 
We now look up the tables of 

the lithium spectrum, and select one or two of the chief 
lines whose wave-lengths have been accurately determined. 
These lines we must next pick out on the lithium-iron plate. 
Having found them, it is clear that we shall be able to find 
some characteristic lines in the iron spectrum close to them. 
We look up the position of these iron lines in the table of 
the iron spectrum, and thus find the exact wave-length in that 


part of the spectrum. It is obvious that here we have a 
double control : the lithium spectrum and the iron spectrum. 
All that now remains to be done is to note with a mapping pen 
on the margin of the standard plate the exact wave-lengths of 
the lines in the iron spectrum whose position we have thus 
established. Other salts are smeared upon the arc, and a 
series of photographs containing iron spectra interspersed with 
characteristic lines of the spectra of other elements, is thus ob- 
tained. In this way we obtain the bearings, so to speak, of 
a series of points in the iron spectrum, and the intermediate 
points are filled in by interpolation. 

Having now got our standard plate, we must proceed to 
apply it in mapping absorption spectra. We fill up the cell C, 
and adjust it so that the light of the iron arc passes through a 
certain thickness, e.g. 80 mm., of a tenth-normal solution of, 
say, quinone in alcohol. We give an exposure of, say, fifteen 
seconds, and then develop the plate. On examination, the 
negative will be found to contain only part of the iron 
spectrum ; for the quinone solution will have absorbed a por- 
tion of the rays. We lay the negative side by side with the 
standard , plate, and are thus enabled to read off the wave- 
length at which the absorption begins.* 

Instead of using wave-lengths, it is more usual to reckon 
in frequencies that is, the reciprocals of the wave-lengths. 

It is clear that the absorption spectrum which we obtain in 
this way is conditioned by three things : the length of the 
exposure ; the strength of the solution ; and the thickness of 
the layer through which we photograph. Taking the case of 
a tenth-normal solution of quinone as an example, we shall 
find the following : If we begin with a layer of solution 
80 mm. in thickness, we find that all the spectrum beyond 
a frequency of about 2000 is absorbed (see Fig. 2, No. 1). 
When we reduce the thickness of the layer of solution in the 
cell to 70 mm., a somewhat peculiar phenomenon occurs. The 
part of the spectrum to the left of 2000 comes through as 

* The best modern spectrographs are provided with a scale of wave-lengths 
which can be photographed directly on the plate, side by side with the spec- 
trum. In this way the frequent use of a standard plate is rendered unneces- 
sary in actual practice ; though the scale must be checked by means of a 
standard plate at intervals to make certain that it has not got out of adjust- 


before, but in addition to it a small portion of light is trans- 
mitted at a frequency of about 2800. This is shown in Fig. 2, 
No. 2. When we reduce the thickness of the layer in the cell 
to 60 mm., we find that this band of light transmitted at 2800 
is a little broader, and it continues to increase in breadth as 

Red end of spectrum 




Ultraviolet end of spectrum 
2800 3200...frequencies 


Transmitted light 
I Absorbed light 

FlG. 2. 

we reduce the thickness of solution through which the light is 
passed. In the end (see Nos. 14 and 15) the light absorption 
between 2000 and 2800 becomes reduced to a minimum and 
finally disappears, so that we have complete transmission. 
The rough triangle of absorbed light extending from No. 2 to 


No. 14 in Fig. 2 is called an absorption land. It will be 
observed that this band has a portion of transmitted light on 
either side of it. When no second portion of the light is 
transmitted, as in No. 1 or below No. 14, we have what is called 
general absorption* The extreme apex of the absorption band 
which is shown in No. 14 is called the head of the band. The 
raggedness of the line running down from the frequency 2000 
is due to experimental error, and is brought about by the fact 
that the iron spectrum is not quite uniform.! Some of the 
lines are strong, and others are weak, so that the readings are 
influenced to some extent. A group of strong lines may be 
transmitted where the weaker ones would not appear, so that 
in some cases our readings are a little too high, in other cases 
a little too low. 

If, instead of using an 80 mm. layer of a tenth-normal 
solution of quinone, we repeated the experiment with an 
8 mm. layer of normal quinone solution, we should find the 
absorption exactly the same in both cases, since the light, in 
traversing the layer, passes in each case through the same 
number of molecules. This is an example of what is termed 
Beer's Law. 

Hartley was the first to see the value of results such as we 
have described. He pointed out that in this way we could 
obtain a map of the absorption spectrum of a compound which 
would serve to identify it completely, for each substance has 
its own particular type of absorption spectrum. Some give 
shallow bands, some deep, and some none at all, but only 
general absorption. Again, some substances have one band, 
some two, and others many. 

It is clear that the diagram in Fig. 2 would be too clumsy 
a mode of reproducing the results obtained, and Hartley 
therefore devised what are called curves of molecular vibration. 
The ordinates in this system represent the thickness in milli- 
metres of a given solution, say a tenth-normal one, and the 
frequencies are plotted as the abscissae. Baly and Desch 1 

* When an absorption band is present, we have what is called selective 

f Hartley has employed electrodes composed of an alloy of lead, tin, and 
cadmium ; the spark spectrum produced with this is practically continuous, 
and thus the disadvantages of the iron spectrum are avoided. 

1 Baly and Desch Trans., 1904, 85, 1029. 


suggested that instead of the thickness being plotted on the 
ordinates, it would be more advantageous from some points of 
view to use the logarithms of the numbers representing the 
thicknesses, and this is the method now generally employed. 
Fig. 3 shows the molecular vibration curve of quinone in 
alcoholic solution plotted in this manner. The data on which 
the curve is based are the same as those from which Fig. 2 
was drawn. The head of the band lies near the intersection 
of the lines 40 and 2200. There is one other term which it 

Oscillation frequencies. 

8 88888 888 8 8888 

1 44 

i 42 

o 40 
^ 36 
1 34 

1 32 
^ 30 
S 28 
3 26 
* 20 
S 18 

a M 


v 1 

D g |g|| || g S .8 

' Itelative thicknesses in mm. of N/10,000 solution . 



























FIG. 3. 

may be advisable to make clear before proceeding further, 
viz. the persistence of the band. By this term we mean the 
change of dilution over which the band remains in existence. 
An examination of the figure will show that the band comes 
into existence just above the line 48 and disappears just below 
the line 40. It has therefore persisted over a change of dilu- 
tion represented logarithmically by eight units. 

In the following section we shall discuss the application of 
this method of investigation to the determination of the con- 
stitution of substances which cannot successfully be dealt with 
by purely chemical means. 


2. The Hartley Method. 

Among the most interesting compounds in organic chem- 
istry are those which are termed tautomeric (or desmotropic, 
pseudomeric, phasotropic, merotropic, or isodynamically iso- 
meric). "Whatever name be applied to these substances, their 
chief characteristic is that an apparently homogeneous mother- 
substance can give rise to two isomeric derivatives according 
to the conditions under which the substitution reaction is 
carried out. 

Now Hartley found that the replacement of hydrogen by 
a methyl or ethyl radicle produced no marked alteration in 
the selective absorption of substances, except in some special 
cases. In general, the members of a homologous series show 
a slight increase in general absorption as we go up the series, 
but in other respects there is very little change in the spectra. 
But if we take the spectra of two such substances as the alkyl 
derivatives of ortho-hydroxy-carbanil, whose constitutions are 
represented below, we shall find that we have hit upon a case 
in which the replacement of a hydrogen atom by an alkyl 
radicle has made a considerable difference in the absorption 

-O.C 2 H 6 

Lactim form. 

Lactam form. 

The absorption spectra of hydroxy-carbanil and of the two 
ethyl derivatives (lactam and lactim forms) are shown in 
Fig. 4,* and it is evident on inspection that we have here two 
different types of absorption spectra. The spectrum of the 
lactam-compound differs completely from that of the lactim- 
ether. Now, when we compare the spectrum of the mother- 
substance with those of the derivatives, it is obvious that it 
resembles the spectrum of the lactam-body rather than that 
of the lactim-ether. From, this, Hartley deduces that the 
mother-substance must have a constitution resembling the 

* The curves are reproduced from Hartley's work, and it should be noted 
that the ordinates represent thicknesses of solution, and not logarithms of 


lactam-form rather than one akin to the lactim-type. We 
should, therefore, ascribe to hydroxy-carbanil the constitution 
(I.) rather than (II.) 

(II.) C 6 H / ^C-OH 

4000 4000 4000 

23456789)12 _2 3456789 | 12 23456789 (123 



4 I 




Lactam -ether 

O-Hyd roxy-carban i I 

FIG. 4. 

Lactim -ether 

This gives some idea of what is known as the Hartley 
method of determining the constitution of organic compounds. 
It has proved itself of great importance in many cases, and is 
certainly one of the most reliable in the solution of these 

3. The Problem of General Absorption. . 

In recent .years the study of absorption spectra has been 
for the most part confined to those questions which centre 


round the selective absorption of organic compounds, and it is 
not difficult to trace the causes which have led to this con- 
centration of attention upon a particular section of the sub- 
ject. The problem of the origin of colour in organic bodies 
attracted the attention of chemists both from the theoretical 
and, in the case of dye-stuffs, from the technical standpoint. 
Now, visible colour depends to a very considerable extent 
upon the presence of one or more definite, absorption bands in 
the visible region ; and when the work of Hartley had brought 
the investigation of the ultra-violet region into prominence, it 
was natural that the idea of colour should be enlarged to in- 
clude bands in the invisible ultra-violet as well as the corre- 
sponding phenomenon in the visible region. And since in the 
visible region the hardest-fought controversy from the chemical 
point of view had centred round the presence or absence of ab- 
sorption bands, it was only to be expected that similar discus- 
sions would arise as to the meaning of the numerous bands 
which were soon found to exist in the ultra-violet section of the 
spectrum. A second cause of the preference shown for the 
investigation of selective absorption is easy to detect. When 
we are dealing with substances which show absorption bands, 
a very slight alteration in their chemical structure is sufficient 
to produce a marked alteration in the character of the banded 
spectrum ; and it is this delicacy which makes selective 
absorption so fascinating to those who delight in theory- 
making. General absorption, on the other hand, is by no 
means so readily affected by minute constitutional changes. 
An examination of the curves on p. 67 will show that while 
the bands in the spectra of the lactam and lactim-ethers of 
hydroxy-carbanil are totally different in character, the general 
absorptions of the two substances differ but little from each 
other. Thus in the case of selective absorption we are dealing 
with a property which is susceptible to extremely minute 
influences ; while in general absorption we have to do with 
forces which must differ markedly from each other before we 
can detect their effects. 

When a compound is studied by the chemist there are 
seven chief questions which demand an answer: 

1. What elements does the molecule contain ? 

2. In what proportion are the various elements present ? 


3. What is the number of the atoms in the molecule ? 

4. In what manner are the atoms linked together ? 

5. Is the substance saturated or unsaturated ? 

6. How are the unsaturated centres of the molecule related 
to each other ? 

7. What is the arrangement in space of the atoms within 
the molecule ? 

Each of the factors indicated in the above questions enters 
into the problem of general absorptive power ; and for the sake 
of convenience we may take them up in turn. 

With regard to the influence of the various elements upon 
general absorption, it is impossible to lay down any hard-and- 
fast rules. Saturated open-chain hydrocarbons are usually 
distinguished by being extremely diactinic ; * and the intro- 
duction of halogen atoms into the molecule generally tends to 
increase the power of absorbing light. When we turn to other 
elements such as nitrogen, sulphur, or phosphorus, we com- 
plicate the problem by the introduction of unsaturation ; so 
that it is hard to estimate the exact part played by the new 
element. The question is of considerable importance, however, 
and it is to be hoped that some investigator may find time to 
throw some light on the point. 

From the point of view of spectroscopy, the second question 
is so closely connected with the first that no progress can be 
made in this direction until the first problem is in a much more 
advanced stage. 

Turning now to the influence of the number of atoms in a 
molecule, we come to the question of homologous series. It 
has been found that if we examine the spectra of the methyl, 
ethyl, propyl, and butyl derivatives of a given substance, the 
power of general absorption increases as we go up the series. 
This was established in 1879 by Soret and Killet 1 in the case 
of the alkyl nitrates ; and it appears to be a general property 
of homologous series. Each atom in a compound seems to 
exert a small amount of absorptive power, and by adding atom 
to atom we eventually obtain quite a marked effect on the 

We next come to the fourth point, the manner in which 

* A substance of very low absorptive power is said to be diactinic. 
1 Soret and Rillet, Compt. rend., 1879, 89, 747. 


the atoms are linked together within the molecule ; and here, 
again, if we confine our attention to saturated substances the 
data are very sparse. Crymble, Stewart, and Wright 1 have 
found that in the case of the alkyl bromides the normal sub- 
stance has a much weaker power of absorption than the iso- 
meric iso-form. For example, normal propyl bromide (I.) is 
much less absorptive than iso-propyl bromide (II.) 

CH 3 .CH 2 .GH 2 .Br >OH. Br 

CH 3 / 
(I.) (II.) 

Eesults of this type might indicate that a symmetrically built 
molecule has more absorptive power than an isomeric molecule 
in which groups are unsymmetrically distributed. 

When we take up the question of the influence of un- 
satufation upon the general absorptive power of molecules, 
the data are very full. 2 It has been established, largely by 
the monumental researches of Hartley, that if we compare the 
spectra of a saturated and an unsaturated hydrocarbon, the un- 
saturated substance will greatly surpass the saturated one in 
its power of absorbing light. Since cyclic carbon chains are 
really unsaturated bodies, it might be expected that this class 
would show a higher absorptive power than open-chain bodies, 
and this has been found to hold good. With regard to the 
influence of the triple bond upon absorption there are very few 
data available ; but it appears 3 that its effect is not so strong 
as that of the ethylenic linkage. 

We must now pass to the question of the effect produced 
by the introduction into the molecule of more than one un- 
saturated group ; and at this point we reach a branch of the 
subject in which chemical constitution plays a very important 

One of the best known characteristics of substances con- 
taining ethylenic bonds is their faculty of uniting with one 
molecule of bromine to form a dibromo- derivative, thus 

1 Crymble, Stewart, and Wright, unpublished observation. 

2 Hartley, Brit. Assoc. Report, 1903. 

3 Stewart, Trans., 1907, 91, 199 ; Ley and Engelhardt, Zeit. physikal. 
Chem., 1911, 74, 311. Compare Macbeth and Stewart, Trans., 3917, 111, 829. 



Br Br Br Br 

Let us now take the case of a substance containing two double 
bonds which are separated from one another by the interposi- 
tion of one or more methylene groups. If to a gramme mole- 
cule of such a compound we add one gramme molecule of 
bromine, we shall find that the halogen is absorbed by one of 
the double bonds, while the second double bond remains intact, 

CH 3 CH : CH CH 2 CH : CH CH 3 
Br Br 

= CH 3 CH CH-CH 2 CH : CH CH 3 

L L 

This is a perfectly normal behaviour ; but if no methylene 
group be interposed between the two double bonds we obtain 
quite a different result, for instead of the halogen atoms attack- 
ing one of the double bonds they attack both simultaneously 
and a new double bond is formed as shown below 

1 23456 

CHg-CH : CH-CH : CH-CH 3 
Br Br 

1 23456 

= CH 3 -CH-CH:CH-CH-CH 3 

Br Br 

It will be seen from the formulas that the original double 
bonds lay between the atoms 2, 3 and 4, 5 ; while in the final 
substance there is a new double bond in the 3, 4-position. 

Thiele, 1 from a study of many systems of this type, put 
forward his theory of "partial valencies." He assumes that 
in a double bond between two carbon atoms the whole of the 
available affinity of the atoms on each side of the double bond 
is not completely neutralized in the union, but that a certain 
amount remains free. This free affinity he terms a "partial 
valency," and he assumes that this partial valency serves to 
anchor any fresh atoms which come within range of the system. 

1 Thiele, Annalen, 1899, 306, 87. 


Once anchored, the presence of the new atoms causes a re- 
arrangement of affinity within the two molecules, with the 
result that the new atoms enter the system and reduce the 
double bond. If we represent the partial valencies by dotted 
lines, the steps in the process could be expressed by the formulae 
CH 3 -CH=CH-CH 3 CH 3 -CH=CH-CH 3 

Br Br 
Br Br CH 3 -CH-CH-CH 3 

Br Br 
When we come to a system of the type 

K CH : CH CH : CH K 

we have what Thiele terms a system of " conjugated double 
bonds," and he explains the abnormal behaviour of this system 
in the following way : Since, under his postulates, the addition 
of new atoms to a system can take place only after they have 
been anchored by the partial valencies of the system, it is clear 
from experimental results that in a compound of the following 


i 2 3 4 

only the partial valencies 1 and 4 are active. We are therefore 
driven to assume that the valencies 2 and 3, acting across the 
single bond, neutralize each other in some way, and are thus 
thrown out of action.* This state of affairs Thiele represents 
by the following scheme : 


Such a grouping as this is anomalous not only from the 
chemical point of view but also in its influence upon the physical 
properties of compounds in which it occurs. 1 Bruhl 2 found 

* The close agreement between this theory and actual experience is found 
in the case of benzene, where all the double bonds are conjugated in a single 
system with no free partial valencies. Hence, theoretically, benzene is not 
a true unsaturated body, and it is well known that experiment bears this out. 

1 For a full discussion of this question, see Smiles' Relations between 
Chemical Constitution and some Physical Properties, 1910. 

2 Bruhl, Ber., 1907, 40, 878. 


that it had a marked effect upon refractive power, producing 
higher values than the calculated refractivities ; Hilditch 1 
traced the same influence in the case of optical rotatory power ; 
analogous results were observed by Pascal 2 in diamagnetism ; 
and Sir W. H. Perkin 8 proved that similar effects could be 
detected in magnetic rotation. 

Crymble, Stewart, Wright, and others 4 have examined 
the spectra of numerous pairs of isomeric substances, one set 
containing isolated double bonds, while the isomeric bodies 
contained a conjugated grouping ; and they find that in every 
case the substance which contains the conjugated system is 
much more absorptive than the isomeric body in which the 
double bonds are not so completely joined together in a single 
system. For example, citraconic acid contains three double 
bonds joined together in a single conjugated system ; whereas 
in the isomeric itaconic acid one of the double bonds is separated 
from the rest by the interposition of the methylene group 


CH 3 -C-C=0 CH 2 =C C=0 


H-C-C=0 CH 2 -C=0 


Citraconic acid. Itaconic acid. 

Citraconic acid has a greater power of absorbing light than is 
shown by itaconic acid. 

From this it is clear that when two unsaturated groups 
occur in a molecule in such positions that they are able to 
exercise a reciprocal influence upon each other, their joint effect 
upon the absorptive power of the substance is very much greater 
than when they are separated from each other by neutral atoms. 
One centre of residual affinity increases the absorptive power 
to a considerable extent as compared with a saturated compound ; 
two isolated centres in turn augment the compound's general 

1 Hilditch, Trans., 1909, 95, 331, 1570, 1578 ; 1910, 97, 1091 ; Edminson 
and Hilditch, ibid., 1910, 97, 223. 

2 Pascal, Compt. rend., 1909, 149, 342. 

3 Sir W. H. Perkin, Trans t , 1896, 69, 1141. 

4 Crymble, Stewart, Wright, and Glendinning, Trans., 1911, 99, 451; 
Crymble, Stewart, Wright, and Miss Bea, ibid. t 1911, 99, 1262. 


absorption; while a still greater power is produced by the 
co-ordination of the two centres into one system. 

We must now pass to a consideration of the effect of stereo- 
isomerism upon general absorption. The spectra of compounds 
containing asymmetric carbon atoms have been examined by 
Magini x and by Stewart 2 with the following results : Two 
optical antipodes have exactly similar absorption spectra ; meso- 
tartaric acid has a greater power of general absorption than 
either active tartaric acid ; while the spectrum of racemic acid 
shows a power of absorption greater than that of dextro-tar- 
taric and less than that of meso-tartaric acid. Turning to the 
question of geometrically isomeric carbon compounds, the same 
authors have shown that maleic and citraconic acids possess 
much less light-absorbing power than the isomeric bodies 
fumaric and mesaconic acid 

Least absorbent. Most absorbent. 



Maleic acid. Fumaric acid. 



Citraconic acid. Mesaconic acid. 

Hartley and Dobbie 3 found that the spectra of the two stereo- 
isomeric oximes of benzaldehyde were identical 

^6^6 G BL C 6 H 5 C H 


Benzsynaldoxime. Benzantialdoxime. 

At first sight it appears strange to find this difference in 
behaviour between the isomers containing an ethylenic bond 
and those which contain the linkage N=C=, but a considera- 
tion of the formulae of the various substances in the light of 
what has already been described in the preceding paragraphs 
will throw some light upon the point. We have already 
learned that general absorption is greatly influenced by the 

1 Magini, J. Chim. ptys., 1904,-2, 403. 

2 Stewart, Trans., 1907, 91, 1537. 

3 Hartley and Dobbie, Trans., 1900, 77, 509. 


presence in the molecule of some un saturated centre ; and that 
when we have two un saturated centres in a substance, the 
absorptive power depends to a great extent upon their mutual 
interaction. Now the change from maleic to fumaric acid 
entails the shifting of one unsaturated carboxyl group with 
respect to the other ; and the same is true of the change of 
citraconic into mesaconic acid : so that in both these cases we 
are to a great extent altering the interplay of forces in the 
molecule. Similar arguments hold- good in the case of the 
change of active tartaric acid into the meso-form, for here also 
the relative positions of the carboxyl groups are altered. When 
we come to the two oximes, however, the case is different, for 
there the two groups which change their relative positions in 
space are the phenyl and hydroxyl radicles. The phenyl group 
certainly possesses a marked degree of residual affinity; but 
the influence of the hydroxyl radicle on general absorption is 
very feeble, as is proved by the very low absorptive power of 
the aliphatic alcohols. Hence we might expect that these two 
groups would influence one another but slightly, and that a 
change in their relative positions would produce no marked 
difference in the absorptive power of the molecule. Support 
has been lent to this view by an examination of the absorption 
spectra of brassidic and erucic acids 1 which were found to have 
almost identical powers of absorption 

CgTIjiy C M CgHjy C H 



Erucic acid. Brassidic acid. 

It will be seen that in this case we have altered the relative 
positions of the groups -rCOOH and C 8 H 17 ; the first of these 
is an unsaturated centre, but the second is a saturated group, 
so that no great change in absorptive power is to be expected 
on the above hypothesis. 

From the foregoing data it might be deduced that when 
the unsaturation of a molecule is arranged in a conjugated 
system, the greatest general absorptive power will be mani- 
fested when the free partial valencies are situated as far as 
possible from one another. In this connection, the cases of 

1 Crymble, Stewart, Wright, and Arbuthnot, unpublished observation. 


dimethyl-diacetylene, hexatriene and benzene are not without 
interest. 1 

C 6 H 6 Dimethyl-diacetylene CH 3 -C.=c^C=C-CH 3 
C 6 H 8 Hexatriene 

C 6 H 6 Benzene (/ \N 


\CH CH^ 

An examination of the molecular formulae of these sub- 
stances shows that hexatriene contains two hydrogen atoms 
more than either of the others ; and by ordinary standards it 
should therefore be regarded as the most saturated of the 
three. On the other hand, its free partial valencies are further 
removed from one another than those of dimethyl-diacetylene 
can be ; and benzene has no free partial valencies at all. If 
the suggestion put forward above be correct, hexatriene and 
dimethyl-diacetylene ought to show much greater absorptive 
power than benzene. When the spectra were compared, this 
was found to be the case. 

4 The, Factors affecting Selective Absorption. 

Among organic compounds, it is found that the presence 
in the molecule of certain atomic groupings produces an ab- 
sorption band in the spectrum of the substance. These parti- 
cular atomic groupings are termed chromophores, 2 and among 
them the most important are the following : CO , N : N , 
-N : 0, CgHg, and =C = NO.OH. 

An examination of the examples given above will show 
that all of them contain residual affinity ; and if this residual 
affinity be saturated, the chromophoric power is lost. Thus 
the spectrum of acetone, CH 3 . CO . CH 3 , contains an absorp- 
tion band ; but when the carbonyl group is reduced it is found 
that the resulting isopropyl alcohol, CH 3 . CH(OH) . CH 3 , ex- 
hibits general absorption only. 

1 Macbeth and Stewart, Trans., 1917, 111, 829. 

2 Witt, Ber., 1876, 9, 522. 


When a chromophoric group is found in the structure of a 
molecule, the molecule is termed a chromogen. 1 The nature 
of the chromogen sometimes influences the absorptive power 
of the chromophoric group which it contains; even in cases 
wherein the only unsaturation of the molecule is to be found 
in the chromophore. For example, the absorption band of 
camphorquinone is very much more persistent than that of 
diacetyl, 2 although both compounds contain the double chromo- 
phore CO . CO and the remainder of the molecule is 
saturated in each case 


CH 3 -C-CIL 

CH 3 .CO.CO.CH 3 


CH 2 1 

CH 3 

Camphorquinone , 

In general, however, the character of the absorption band 
shown by a substance depends mainly upon the chromophore, 
and is but little affected by the remainder of the chromogen 
when the latter contains no centres of residual affinity. 

The next question which presents itself is the influence 
exerted upon one another by two or more chromophores in 
the same molecule. This problem must be approached from 
two sides : for we may find the chromophores directly ad- 
jacent to one another in the molecular structure; or they may 
be separated from each other by saturated radicles. 

Taking the first alternative in the case of the ketones, the 
following data throw some light upon the point : 

Acetone . . CH 3 . CO . CH 3 Absorption band in ultra-violet 

Diacetyl . . CH 3 . CO . CO . CH 3 Bright yellow 

Triketopentane . CH 3 . CO . CO . CO . CH 3 Orange 

From this it will be seen that the accumulation of chromo- 
phores in adjacent positions in the molecule tends to bring 
the absorption band more and more towards the red end 
of the spectrum. It will be convenient to apply the name 
" conjugated chromophores " to such directly united chromo- 
phoric groups. 

1 Witt, Ber., 1876, 9, 522. 

2 Stewart and Baly, Trans., 1906, 89, 496, 503. 


Turning to the second type, wherein the two chromophores 
are separated from one another by a saturated grouping, we 
find quite different results. Acetone, CH 3 . CO . CH 3 , and 
acetylacetone, 1 CH 3 . CO . CH 2 . CO . CH 3 , both have bands in 
the same position in the ultra-violet ; but the acetone band is 
seen in concentrated solutions whereas the band of acetyl- 
acetone is only brought out when the solution is very highly 
diluted. Since there are fewer molecules of the ketone pre- 
sent in the weak solution than in the strong one, it seems 
reasonable to conclude that the process giving rise to the band 
in acetone is feebler than that which takes place in acetyl- 
acetone. It is probable, therefore, that when two identical 
chromophores are separated from one another by a methylene 
radicle they act independently; the band produced by them 
is exhibited at the same position in the spectrum as if either 
of them was acting alone ; but the dilution at which the 
band makes its appearance is increased by the joint action 
of the two. 

A very interesting example of both arrangements of 
chromophores is to be found in the case of some alkyl de- 
rivatives of iodine. 2 In this case the chromophore is the 
iodine atom attached to the carbon. The compounds ex- 
amined were 

One Chromophore, Two Chromophores. 

Methyl iodide . CH 3 . 1 Methylene iodide . CH 2 .I 2 

Ethyl . CH 3 .CH 8 .I Ethylene . I.CH 2 .CH 2 .I 

Three Chromophores. 

lodoform CHI 3 

Potassium tri-iodide KI 3 

an<j. their absorption spectra are shown in Fig. 5. 

Examination of the curve will show that ethyl iodide has 
an absorption band with its head at a frequency of 3900 and 
that this band occurs at the ordinate 3*7. Methyl iodide 
shows an almost exactly similar band, which has been omitted 
so as not to complicate the figure. 

As can be seen from the' figure, the spectrum of ethylene 
iodide contains a band whose head lies at 3900, just as that of 

1 Baly and Desch, Trans., 1905, 87, 766. 

2 Crymble, Stewart, and Wright, Ber., 1910, 43, 1183. 



the mono-iodo-derivative does. The main difference between 
the two curves (ethyl iodide and ethylene iodide) lies in the 
fact that in order to obtain the ethyl iodide band a much 
thicker layer of solution must be used than is necessary in 
the case of ethylene iodide ; for the one band occurs in the 

g 8 

O CM -* 

Oscillation Frequencies. 

\ 8 



______ lodoform. 

Pol-assium Iriiodide. 

Met-hylene iodide. 

______ Efhyl iodide. 

..___._ Ehhylene iodide. 
_ Iodine. 

FIG. 5. 

neighbourhood of the ordinate 3'7 whereas the other makes 
its appearance about ordinate 31 If we take a rough analogy 
from a piano, we might say that ethyl iodide and ethylene 
iodide both strike the same note ; but ethyl iodide strikes it 
with the soft pedal down whereas ethylene iodide strikes it 
with the loud pedal in operation. In other words, the same 
process is going on in the two cases; but it is much more 


intense in the case of ethylene iodide owing to the fact that 
the band-producting group E' CH 2 . 1 occurs twice over in the 
same molecule. 

Quite different results are found when the spectrum of 
methylene iodide is examined. Here there is no band at the 
frequency 3900 ; but instead a new band makes its appear- 
ance with its head at about 3400 i.e. much nearer the red 
end of the spectrum. Evidently in this case we have to do 
with a conjugation of the two chromophores ; and the band 
with its head at 3400 corresponds to the grouping : CI 2 . 

Turning to the spectrum of iodoform, CHI 3 , it will be seen 
that this also contains a band with its head at 3400 ; and this 
must be ascribed to the presence of the grouping : 0X2 in the 
iodoform molecule. But in addition to this band, another 
band makes its appearance with its head at 2900 ; and since 
no trace of this is to be found in the spectrum of CH 2 I 2 , its 
cause must be sought in the mutual influence of the three 
iodine atoms taken together. 

An examination of the potassium tri-iodide curve on the 
figure bring out a further point of interest. The spectrum of 
potassium tri-iodide contains two bands whose heads lie at 
approximately the same frequencies as those of the bands in 
the iodoform spectrum. In the case of each compound we 
have three iodine atoms attached to a central atom ; and this 
resemblance makes itself apparent in the absorption curves, 
notwithstanding the fact that the central atom in the one 
case is a carbon one and in the other case a potassium atom. 
This suggests that here also the influence of the non-chromo- 
phoric portion of the molecule has but little influence upon the 

We must now turn to yet another factor which may in- 
fluence absorption spectra: the nature of the solvent. There 
seems to be no doubt that solvent action plays a considerable 
part ; for when the spectra of substances in the form of pure 
vapour are compared with corresponding solutions in various 
solvents, it is found that the vapour spectra show a very 
large number of narrow absorption bands, whilst the solu- 
tions exhibit comparatively few bands of a much more dif- 
fuse character. In the case of aqueous solutions, ionization 
may produce different spectra at different concentrations ; and 


even in non-aqueous solutions there may be interaction be- 
tween the solute and solvent. 1 

5. Valency and Absorptive Power. 

In the course of his study of the relations between the 
physical properties of elements and their places in the periodic 
system, Carnelley noted that under certain limitations the 
colour of a compound depends to some extent upon the atomic 
weights of the constituent elements. For instance, if we take 
compounds AX, BX, CX, etc., where A, B, C, etc., represent 
successive elements belonging to the same sub-group in the 
periodic table, and X is any other element, then we shall find 
that, as we pass from the light element A through the heavier 
element B to the still heavier element C, the colours of the 
compounds deepen in tint. For example, let us take the case 
of the halogen compounds of cobalt and nickel 

CoCl 2 . . . Blue Ni01 2 . . . Yellow 

CoBr 2 . . . Green NiBr 2 , . . Golden yellow 

CoI 2 . . . Black NiI 2 . . . Black 

Here, as we pass from the chlorine atom, with an atomic 
weight of 35*5, to the iodine atom, with an atomic weight of 
127, we find that the colour deepens from blue to black in the 
case of cobalt salts, and from yellow to black in the case of the 
corresponding nickel derivatives. Again, if we take the case 
of salts in which the acidic part remains constant through- 
out the series while the metal is replaced in rotation by the 
next higher member of the same sub-group, we get similar 

Nal . . White Agl . . Light yellow 

Cul . . Cream coloured Aul . . Golden yellow 

This rule seems to be of fairly general application, but as it 
refers to substances in the solid condition, it is possible that 
the state of aggregation exercises a considerable influence. 
This might account for several of the exceptions to the rule 
which have been observed. 

1 See, for example, the behaviour of iodine in alcohol, chloroform, and 
other solvents (Stewart and Wright, Ber., 1911, 44, 2819 ; Trans., 1917, 111, 



Crymble l approached the subject from a standpoint which 
evades this difficulty. On photographing the absorption spectra 
of potassium chloride and potassium sulphate, he found that 
10 millimetres of a normal solution of either salt was diactinic. 
This proves, of course, that all the ions present in the solution 
under such conditions are diactinic. Hence it is clear that 
the colour of solutions of copper chloride or copper sulphate 
must be due to the copper ion. Taking as a standard a 
10-millimetre thickness of a normal solution of a salt or 
a length of solution containing an equivalent amount of the 
metal, Crymble regarded a compound as "diactinic" when 
under these conditions light is transmitted up to a wave- 
length 2300. 

He found that under this definition metallic ions .may be 
grouped into three classes 

A. Non- absorptive ions. This class includes the following 
metals : 

Li, Na, K, Be, Mg, Zn, Cd, Ba, Al, and Th. 

B. Ions showing general absorption only. In this class 

Hg, Tl, Sn, Pb, Sb, and Bi. 

C. Ions showing selective absorption. In this class may 
be placed 

Cu, Ce, Mn, Fe, Au, Or, U, Co, Pt, Ti, and V. 

Now, if we examine the metals placed in Class A, it is 
obvious that all of them possess a fixed valency : lithium, for 
example, forms only one chloride. The metals which show 
absorption (Class B and Class C), are all capable of exhibiting 
a change of valency : tin forms stannous and stannic chlorides, 
while copper forms cuprous and cupric salts. Thus it is 
established that metallic elements which are capable of exert- 
ing two degrees of valency have a much greater power of 
absorption than those metals which do not yield more than 
one chlorine compound. 

A comparison of the majority of the elements in Class B 
and Class C will show that here also we can trace a general 

1 Crymble, Proc,, 1911, 27, 68. Lieutenant Crymble was killed in action 
in France in 1914. 


rule ; for the elements in Class C, whose ions show selective 
absorption, yield two series of salts in which the valency of the 
metal changes its valency by a single unit 

FeCl 2 and FeCl 3 Cr01 2 and CrCl 3 MnS0 4 and Mn 2 (S0 4 ) 3 

whereas in the case of the metals in Class B, which show 
general absorption only, the change of the " ous " form to the 
"ic" form entails a change of two units in the valency 

T1C1 and T1C1 3 SnCl 2 and SnCl 4 SbCl 3 and SbCl 6 

It is not yet certain whether this second generalization 
will hold good in every case; but the first rule put forward 
by Crymble marks a great advance in our knowledge of the 

If we apply to this problem of the metallic ions the results 
which we arrived at from our examination of the effect of 
unsaturation upon absorptive power in organic substances, we 
might be led to assume that the metallic atoms contain certain 
centres of residual affinity which are called into play by the 
combination of one atom with another. In the case of atoms 
such as sodium and potassium, which are capable of exhibiting 
one valency only, we might imagine that they contained only 
one such centre. In the case of atoms showing more than one 
valency, however, it seems possible that the atom may include 
two centres of residual affinity, which may act like two con- 
jugated double bonds in an organic compound, and by their 
interplay may increase the absorptive power of the substance 
just as we found in the case of organic bodies. 

6. The Spectroscopic Determination of Chemical Change. 

Dobbie, Lauder, and Tinkler 1 have shown how spectro- 
scopic measurements may be used to throw light upon the 
extent to which chemical change takes place in a solution, and 
also upon the relative strengths of two reagents which have 
the same chemical effect. The substance used by them in this 
research was cotarnine, which is capable of existing in two 
forms. What the structures of these two forms are is of no 
importance to us at the present time ; considerable controversy 

1 Dobbie, Lauder, and Tinkler, Trans., 1904, 85, 121. 

8 4 


has centred round the point. All that we need remember for 
our present purpose is that the change of one form into the 
other takes place under the influence of alkalis, while the 
reverse change is produced by acids. 

In aqueous solution, cotarnine is yellow ; that is to say, it 
has an absorption band in the visible region. Now, when the 
aqueous solution is treated with alkali, the yellow colour dis- 
appears, the amount of colour change being proportional to 
the quantity of alkali present. To compare the strengths of 
various alkalis, therefore, it is only necessary to add fixed 
quantities of them to a given solution of cotarnine, and esti- 
mate spectroscopically the change which they bring about in 
the substance. 

In this way, Dobbie, Lauder, and Tinkler arrived at the 
results shown in the table below. The figures represent the 
percentages of the ammonium form of cotarnine which is con- 
verted into the carbinol form in presence of the corresponding 
amount of alkali. 

Strength of base. 




Ba(OH) 2 . 

Ca(OH 2 ). 

NH 4 OH. 





































































The possible applications of this method are, of course, 
almost inexhaustible ; and it would be quite feasible to enlarge 
its scope and make it a means of measuring reaction velocities. 

7. The Spectrophotometer. 

Hitherto in this chapter we have confined our attention to 
the results which have been attained by the use of Hartley's 
curves of molecular vibration ; but it is necessary to say a few 


words with regard to a newer method 1 of measuring absorption 
spectra which has been brought into practice in comparatively 
recent times. This method yields results to which a definite 
physical meaning can be attached. 

The object of the spectrophotometer is to measure quanti- 
tatively the proportion of light of a given wave-length which 
is absorbed when it passes through the substance under 
examination. Hartley's method, on the other hand, was a 
more or less qualitative one ; for in it no attempt was made 
to estimate the proportion of light absorbed. In the Hartley 
method we merely measure the number of molecules of the 
substance which is required to reduce the light traversing 
them to such an extent that it will fail to impress a photo- 
graphic plate after it has emerged from the substance ; and 
it is obvious that this is a purely arbitrary standard to set up. 

The simplest form of the spectrophotometer can be very 
briefly described. Between the ordinary quartz spectrograph 
and the light source is introduced the photometer proper. 
This consists of two revolving discs, one pierced with a fixed 
aperture, the other having an aperture which can be varied at 
will. Behind the fixed aperture is placed a cell containing a 
solution of the substance under examination. The light from 
the source travels in two rays, one of which passes through 
the variable aperture to the slit of the spectrograph whilst the 
other passes through the fixed aperture and the solution and 
thence to the slit. At the slit the two beams are received 
on the two halves of a biprism and are directed on to the 
photographic plate, where they are registered side by side. 

A series of photographs is taken, the aperture of the disc 
being varied for each exposure. During the exposure, the 
discs are rotated at a uniform speed by means of a motor. 
On examining the plate, we find the photographs arranged 
in pairs. In one of each couple, the amount of light passing 
to the plate has been reduced by the variable aperture ; in the 
case of the other the reduction in the amount of light has been 
produced by the absorption of the substance in the cell. If at 

1 Bielecki and Henri, Compt. rend., 1912, 155, 456, 1617; 1913, 156, 550, 
884, 1322, 1860; 157, 372; 1914, 158, 567, 866, 1022, 1114; Ber. y 1913, 46, 
2596, 3627 ; 1914, 47, 1690. A full account of the work is to be found in 
Henri's Etudes de photochimie (1919). 


any part of the spectrum we find that the two photographs 
show equal darkening on the negative, it is clear that the 
intensity in the two cases has been equal. We know what 
proportion the aperture of the sector during this experiment 
bore to its full aperture ; so that we know the proportion of 
light excluded from the plate in this instance. Since the same 
amount of light was obviously excluded by absorption in the 
cell for the effects of the two beams on the plate are the 
same it is clear that we can read off directly the proportion 
of light absorbed by the solution of the substance which we 
have used. For example, if we find that when the sector is 
only one-tenth open we get lines of equal blackness in the two 
spectra at frequency 3450, we can say that for that particular 
frequency the substance absorbs nine -tenths of the light which 
enters the cell, and only allows one-tenth to pass through to 
the plate.* 

Since photographic plates vary from batch to batch as re- 
gards their sensitiveness to different portions of the spectrum, 
it is necessary to calibrate the instrument carefully for each 
batch of plates used. 

8. The Calculation of Absorption Curves. 

V. Henri l has shown that in certain cases it is possible to 
calculate the curve of absorption of substances, provided that 
certain fundamental data are known. In the first place, it is 
necessary to know what chromophoric groups the compound 
contains ; secondly, the relative position of these groups in the 
molecule must be ascertained ; and thirdly, the frequencies of 
the infra-red absorption bands produced by these chromophoric 
groups must be known. 

When these various factors have been ascertained, the 
absorption curve is calculated from the following scheme : 

v = v rvi c = ae 

where v and * are the corresponding frequencies and absorptions 
of the compound under consideration, v and are the values of 

* In actual practice the sector opening is calibrated accurately by photo- 
graphing a substance of a known absorptive power and applying the necessary 
corrections from the data thus acquired. 

1 V. Henri, fitudes de photochimie, chapter vi. 


frequency and absorption coefficient for the chromophore, v f is 
the frequency of the infra-red band of the chromophore, r is an 
integer which rarely exceeds 5 and a is a numerical factor. 

To make the matter clear, a concrete case may be given. 
Suppose that it is desired to calculate the absorption curves of 
oxalic acid. The first step is to ascertain the absorption 
spectrum of acetic acid, which serves as a basis. In this 
particular example, the value of v is 240 x 10 12 and a is 2, 
according to Henri's work. .We select some particular fre- 
quency in the acetic acid spectrum and calculate the corre- 
sponding frequency in the oxalic acid spectrum by subtracting 
240 x 10 12 . To get the height of the ordinate corresponding to 
this abscissa in the oxalic acid spectrum we take the value of e 
at the original frequency in the acetic acid spectrum and multiply 
it by 2 ; and we then use this value as the ordinate in the oxalic 
acid spectrum we are calculating. Having thus ascertained the 
abscissa and ordinate for one point on the oxalic acid curve, we 
proceed to calculate the other points in an analogous manner 
until the curve is complete. 

It will be seen that the calculation is a somewhat laborious 
one ; but the results are in many cases wonderfully accurate 
the calculated and observed curves practically coinciding with 
one another. Even in the case of such complex substances as 
uric acid (I.) it is possible to obtain a close approximation 
between the experimental and calculated curves when the latter 
is based on the spectrum of isobarbituric acid (II.) 


I 1 II. 


il >co i ii 


(I.) (II.) 

There can be no doubt that the work of Henri represents 
almost as great a stride in advance as the investigations of 
Hartley did in their time. If absorption spectra can thus be 
brought within the group of " calculable " values, a new and 
extremely practical weapon will be placed in the hands of the 
synthetic dye industry; while on the purely scientific side, 
great advances will be possible. It is perhaps too soon to 
anticipate results in the near future, as Henri has pointed out 


the extremely laborious nature of the work entailed in ascer- 
taining the fundamental constants ; but the mere inception of 
the method is in itself a very important advance. 

9. Conclusion. 

In absorption spectra at the present time, one of the most 
promising fields for investigation appears to be the influence 
of solvents. If a long series of experiments were carried out 
in which attention was paid to the physical properties (such as 
dielectric constant) of inert solvents, one can hardly doubt 
that some important results would follow. Another branch of 
the same subject is the influence of the solvent upon reaction- 
velocities, and here it might be possible to detect a parallelism 
between the action of the solvent upon the reacting bodies and 
its action upon their spectra. 

Again, spectroscopy furnishes us with a means of watching 
the changes which take place while a reaction is going on in a 
solution ; and it -seems probable that interesting results might 
be acquired by a study of reactions which may or may not 
involve the formation of an unstable intermediate product, or 
of those which result in intramolecular change. 

A field which appears to have been barely t6uched is the 
examination of analogous compounds of similar elements. For 
example, it might be of interest to discover whether there was 
any relation between the absorption spectra of an alkyl ether, 
a sulphide and a selenide and the chemical properties of the 
three substances. Parallel results might be found in the 
examination of derivatives of tri- and penta-valent nitrogen, 
phosphorus and arsenic. In the " onium " salts, also, the 
spectra might throw some light upon the cause of the ioniza- 
tion of the negative radicle. 

At the present time we are ignorant of the cause of colour 
in many of the simplest substances. For example, if nitrogen 
and hydrogen, both colourless gases, be combined, we get a 
colourless gas, ammonia, formed. But if for hydrogen we 
substitute oxygen, we get a brown gas, N0 2 , which on poly- 
merisation gives us a much more weakly coloured body, N 2 ^4- 
It is clear that in cases such as these one atom has the effect 


of stimulating another into a certain state of vibration, while 
other atoms have not this power. 

At the present day, the problem of what is termed residual 
affinity is rapidly being forced to the front in many widely 
different lines of investigation, and it is becoming more and 
more evident that our ordinary graphic formulse are reaching 
the end of their usefulness. As an expression of the purely 
chemical behaviour of compounds they have played a wonder- 
ful part on the chemical stage, and in the future they will no 
doubt aid us in this as well as any substitute which can at 
present be imagined. But when we come to the relation 
between chemical constitution and physical properties the 
ordinary graphic formula, devised as it was to suit a totally 
different "set of circumstances, cannot be blamed if it fails to 
throw much light on the question. What will be required in 
future is a standardization of what we describe as residual 
affinity, and it appears possible that spectroscopy may aid us 
in this line. Assuming the electronic basis of matter, we are 
driven to regard the chemical, electrical, and physical proper- 
ties of a compound as different facets of a whole, or as inter- 
pretations of the electrons' motions in terms of various crude 
measurements with different units. At the present time our 
measurements of residual affinity are in the atomic stage, that 
is to say, we ascribe to both a compound containing an ethyl- 
enic bond and one containing a carbonyl group the same 
amount of residual affinity since they are both capable of com- 
bining with two hydrogen atoms. If it were possible to 
apply some spectroscopic measure of residual affinity, we should 
be in possession of a much finer measuring instrument, and it 
is not improbable that the differentiation between] one com- 
pound and another in this way would lead to valuable results. 
Ordinary chemical means can carry us only to the confines 
of the atomic world, but for a true understanding of the 
whole field we must be able to penetrate into the subatomic, 
or electronic universe, for it is in this region that the real 
changes occur which we measure in the ordinary chemical re- 



1. The Phenomena of Crookes' Tubes. 

WHEN an electric current passes through a highly evacuated 
tube, three types of rays may be produced, according to the 
experimental conditions ; cathode rays, positive rays, and X-- 
rays. As a parallel set of phenomena will be met with when 
we come to deal with the radioactive elements, it seems best to 
deal with all three classes here, though strictly the first two 
types lie outside the scope indicated by the title of the present 

From the cathode of a Crookes' tube, cathode rays are 
thrown off at right angles to the surface. They are streams 
of electrons travelling at very high velocities in straight lines ; 
and are capable of penetrating thin sheets of aluminium or 
gold leaf. They ionize gases through which they pass; and 
produce an image on a photographic plate. When they im- 
pinge upon certain materials they set up fluorescence or phos- 
phorescence; and they also heat up any substance by their 
impact. Finally, owing to their electrical nature, the particles 
which compose the stream can be deviated from their straight 
paths by magnetic or electric fields. Each electron represents 
a charge equal in quantity but opposite in sign to that carried 
by a hydrogen ion. 

The second type of rays generated in a Crookes' tube is 
known by various names : canal rays, Kanalstrahlen, Gold- 
stein rays or, in recent times, positive rays. They are streams 
of positively charged molecules of the residual gas in the tube 
which are produced when the cathode is pierced with one or 
more apertures ; and they travel in straight lines in a direc- 
tion opposite to that taken by the cathode rays. They have 



less power of penetrating matter than cathode rays have ; and 
they are not so susceptible to deviation by electrical or mag- 
netic fields. 

Finally, we come to the X-rays, which are produced when 
cathode rays impinge upon matter. For long after their 
discovery, the nature of the X-rays remained in doubt; but 
it is now known that they are vibrations similar to light, 
though their wave-lengths are extremely small as compared with 
those of light-waves. Like light, an X-ray can be diffracted, 
and polarized. When passed into a solid, an X-ray pencil is 
" scattered " just as light is scattered in a foggy atmosphere. 

2. Characteristic X-rays. 

When an X-ray beam impinges upon matter it may produce 
three types of radiation : scattered X-rays, characteristic X-rays, 
and corpuscular rays. The proportions in which these three 
vibrations are generated vary with the atomic weight of the 
element upon which the parent X-ray strikes. Elements of 
low atomic weight radiate mainly scattered X-rays ; an element 
of higher atomic weight, such as zinc, emits for the most part 
characteristic rays; whilst the heavy elements give off more 
corpuscular radiation than the lighter ones, mass for mass. 
The heavy elements scatter incident X-rays to a greater degree 
than the lighter elements, when account is taken of their rela- 
tive masses. 

Scattered X-rays are of no great importance from the 
chemical point of view ; and we shall deal in a later section 
with corpuscular radiation ; so for the present we may confine 
our attention to the characteristic rays which are emitted by 
different elements when X-rays strike their surfaces. 

An ordinary X-ray tube generates X-rays of many different 
qualities simultaneously; so that the beam arising from it is 
composed of a mixture of different vibrations just as white 
light is. If such a mixture of X-rays be allowed to fall upon 
a metallic surface, it is found that the metal is stimulated and 
gives off characteristic X-rays of definite wave-lengths. 1 The 
nature of the new radiations depends solely upon the metal 

1 Barkla and Sadler, Phil. Mag., 1908, 16, 550 ; Kaye, Phil. Trans., 1909, 
A, 209, 123. 


from which they are generated and has no relation to the pro- 
perties of the original X-ray, provided always that the parent 
X-ray is harder * than the characteristic radiation of the metal 
used. If the parent rays are too soft, no characteristic ray 
is excited by them. A rough rule is that the characteristic 
radiation of a given atom can excite the corresponding radiation 
of a lighter atom, but cannot produce any such effect in the 
case of a heavier atom. 

An analogy to this phenomenon is to be found in the case 
of fluorescence excited by light. If we allow white light, 
which is a mixture of many wave-lengths, to fall upon a solu- 
tion of quinine, we get a blue fluorescence produced which is 
characteristic of quinine. The same white light falling upon 
a fluorescein solution generates a green fluorescence, which is 
characteristic of fluorescein. 

It has been shown that some elements give out several 
kinds of characteristic radiations, and these types are known 
as the K, L,, series. The K radiation is about three hundred 
times more penetrating than the L type. Both of these radia- 
tions conform to 1 an empirical relation, known as Owen's Fifth- 
power Law, which connects quality of the radiation with the 
atomic weight of the metal emitting it. This law states that 
the penetrating power of the radiation is roughly proportional 
to the fifth power of the atomic weight of the metal from which 
the rays are generated. 1 

A simple relation has been established between the penetrat- 
ing powers of the K and L radiations and the atomic weights 
of atoms. 2 If an element of atomic weight A L emits a soft (L) 
radiation of a certain degree of hardness, then the atomic 
weight AS of a particular element having a K radiation of the 
same hardness is given by 

A K = J(A t - 48) 

These characteristic X-rays are governed entirely by atomic 
factors and are not affected by molecular conditions; for an 
element has been shown to give exactly the same quantity of 

* The terms " hard" and " soft " applied to X-rays indicate the penetra- 
tive power of the rays, hard rays being more penetrating than soft ones. The 
hard rays are probably of shorter wave-length than the soft ones. 

1 Owen, Proc. Boy. Soc., 1912, A, 86, 426. 

2 Chapman, ibid., 439 


characteristic rays whether it exists in the free state or in 
combination as a salt. 1 For example, Fe, Fe 2 3 , Fe 3 4 , FeS0 4 , 
and K 4 Fe(CN) 6 all emitted the characteristic X-rays of iron. 2 

It has been found that strong characteristic radiations may 
be obtained in the following manner. 1 The cathode stream of 
a Crookes' tube is allowed to fall upon an anti-cathode of the 
metal whose characteristic radiation it is desired to produce. 
Screens of the same metal are then placed in the path of the 
X-rays generated at the anti-cathode; and in this way any 
non-characteristic radiations are converted into those character- 
istic of the metal under examination. Whiddington 3 deter- 
mined the minimum velocity which the cathode particles must 
possess in order to excite the characteristic radiation ; and this 
also proves to be a function of the atomic weight. If V K is 
the minimum velocity of the cathode rays in cms. per second ; 
and A is the atomic weight of the metal composing the anti- 
cathode : then in the case of the K radiations the following 
relation holds: 

V K = A . 10 8 

This has been tested for a range of elements extending from 
aluminum to selenium. 

The transformation of the energy of cathode rays into that 
of X-rays has been shown by Beatty 4 to be influenced by the 
atomic weight of the material of the anti-cathode. The follow- 
ing relation appears to hold : 

X-ray energy _ = ^ ^ ^ 
Cathode ray energy 

where A is the atomic weight and )3 is the velocity of the 
cathode rays expressed as a fraction of the velocity of light 
(3 X 10 10 cms. per second). 

For example, if a platinum anti-cathode be excited by 
cathode rays with a speed of 10 10 cms. per second, we have 
A = 195 and )3 = J so the fraction becomes equal to 6*61 X 
10 ~ 3 . These results refer only to primary X-rays; and if, in 

1 Kaye, Phil. Trans., 1908, A, 209, 123; Chapman, Phil. Mag., 1911, 21, 

2 Glasson, Proc. Camb. Phil. Soc., 1910, 15, 437. 

8 Whiddington, Proc. Roy. Soc., 1911, A, 85, 323. 
4 Beatty, Proc. Roy. Soc., 1913, A, 89, 314. 


addition, characteristic X-rays are excited, their effect must be 
added to that which is given by the formula above. 

We must now turn to the question of the absorptive power 
which elements exhibit when penetrated by characteristic rays. 1 
The absorption coefficient X is derived from the expression : 
I = I^" 1 ^, wherein I represents the transmitted X-radiation, I 
represents the incident X-radiation, and x stands for the thick- 
ness of the layer of material through which the X-rays pass. 
Since the density of the material is a factor in the problem, it 
is usual to take it into account and define the absorption by 

the fraction - where p is the density of the substance under 


The absorption phenomena are rather complicated ; and to 
make them clear it will be best to take a concrete example. 
Suppose that we use a nickel cathode as the source of X-rays 
and that we find it gives off three types of characteristic radia- 
tion which we may denominate the K, L, and M series of rays. 
The K type are harder than the L-rays and these in turn are 
harder than the soft M-rays. We choose the K-rays to work 
with ; and we allow them to fall upon equal masses of a series 
of metals of increasing atomic weight. 

On an earlier page it was pointed out that in order to excite 
the characteristic radiation of any element it was necessary to 
throw upon it X-rays of greater hardness than that of the 
element's own characteristic radiation. We begin by throwing 
the nickel K radiation on elements of an atomic weight less 
than that of nickel ; and we find that the absorption of the 
element increases with the atomic weight, as shown by the 
rising curve marked K + L + M in Fig. 6. Over this range, 
the metals tested all emit their three types of characteristic 
radiation, K, L, and M. Now, when we come to nickel itself 
(atomic weight 61*3) we find that there is a sudden drop in the 
absorption, as is indicated by the dotted line on the diagram. 
What has happened is that the nickel K radiation is now no 
longer able to excite the K radiation of the element on which 
it is falling, since this element (nickel) would emit a K radia- 
tion as strong as that with which we are trying to excite it. 

1 Barkla and Sadler, Phil. Mag., 1909, 17, 739 ; Barkla and Collier, ibid., 
1912, 23, 987. 



All that the incident nickel K radiation can now do is to excite 
the next element to emit the softer L and M radiations. 
The absorption of these increases with rise in the atomic weight 
of the absorbing element, as shown by the second up-grade of 
the curve (marked L + M in the figure) ; until we reach an 
element with an atomic weight of approximately 160. This 
element repeats the behaviour of nickel itself, for its L radiation 
is as strong as the nickel K radiation ; and consequently all 
that the latter can do is to excite the soft M-rays. The results 
can be seen from the figure. 

Q 40 80 \20 160 200 2*0 

Atomic Weigur of Absorbing ElemenT 
(From Kaye's " X-Rays ".) 

FIG. 6. Graph showing relation between the absorption of nickel (K) radia- 
tion by various elements and the atomic weight of the absorbing element. 
The absorption passes through a minimum for a screen of nickel and also 
for one of atomic weight of about 164 (whose L radiation is identical with the 
nickel (K) rays). 

It is self-evident that if we used the L radiation of the 
element at atomic weight 160 as our exciting radiation, 
the results would be the same as those given above, since the 
hardnesses of the nickel K and of the L radiation from the 
160 element are equal. 

Benoist J applied the absorption of X-rays to the solution 

1 Benoist, Compt. rend., 1901, 132, 325, 545, 772 ; J. de Physique, 1901, 
iii.), 10, 653. 


of chemical problems in the following manner. If we 
cast a decigramme of a substance into the form of a right 
cylinder, the area of whose base is 1 sq. cm., and if we allow 
X-rays of a given quality to fall normally upon the base and 
pass through the cylinder, the " specific opacity " of the 
material is measured by the fraction of the rays which is 
absorbed during their passage through the cylinder. Another 
constant is the " equivalent transparency " which is measured 
as follows. A cylinder of paraffin wax, 1 sq. cm. in cross-sec- 
tion and 75 mm. high is chosen as the standard. The material 
to be examined is cast into a cylinder of equal cross-section. 
By experiment, the height of the second cylinder which will 
give an absorption equal to the wax is determined. The mass 
of the material in this size of cylinder is termed the " equi- 
valent transparency " of the substance under examination. Both 
the specific opacity and equivalent transparency are properties 
independent of the state of aggregation and the temperature : 
and they are not affected by the state of combination in which 
the elements under examination exist. 

Now Benoist showed that the specific opacity increases in 
a regular manner with the rise in atomic weight; so that if 
we plot specific opacity against atomic weight for a series of 
elements we get a smooth curve. Thus in order to determine 
whether an element has an atomic weight A or nA, it is only 
necessary to establish its specific opacity and that of its pos- 
sible neighbours. If the specific opacities lie on a smooth 
curve, the correct atomic weight has been found ; if not, some 
multiple of this atomic weight must be selected and the ele- 
ments with atomic weights near that must be examined until 
a smooth curve is obtained. Evidence of this kind has been 
found useful in establishing the position of indium in the 
Periodic Table. 

3. Corpuscular Rays. 

The corpuscular rays may be dismissed very briefly. It 
was shown by Curie and Sagnac 1 that when an X-ray beam 
strikes a plate, part of the secondary radiation from the plate 
carries a negative charge. The intensity of the atomic 

1 Curie and Sagnac, Compt. rend., 1900. 


corpuscular radiation appears to be proportional to the fourth 
power of the atomic weight of the element from which the rays 
emanate 1 ; and the radiation becomes abundant when the 
material begins to emit its characteristic X-rays. 

"When the incident X-ray is a characteristic X-ray that is 
to say when it is a homogeneous radiation the corpuscular 
ray excited by its impact on any given metal is also homo- 
geneous in character. In each case, the corpuscular ray has 
a definite absorption coefficient which is proportional to the 
atomic weight of the metal whose characteristic rays were 
utilized as an exciter. 

The fourth power relation mentioned above leads to in- 
teresting deductions. 2 If we write Jc for the absorption 
coefficient and A for the atomic weight of the metal emitting 
the exciting characteristic X-rays, we have the relation 

k . A 4 = constant 

Now by utilizing a formula deduced theoretically by Sir J. J. 
Thomson, the maximum velocity ~V d with which a cathode 
ray can leave a material of thickness d is given by 

V 4 - V 4 * = a . d 

where V is the initial velocity of the ray and a is a constant. 
From this it follows that 

Jc . tf = constant 

where v is the velocity of the particle at the moment of pro- 
jection from the radiator. Combining the two expressions, it 
is clear that v varies with the atomic weight. 

Measurements by Beatty have shown that this velocity v is 
equal to 10 8 . A ; and this turns out to be the critical velocity 
which a cathode ray must possess in order to be able to 
generate a characteristic radiation, as has been mentioned on 
an earlier page. Thus it appears that the corpuscular radia- 
tions have, approximately, the same velocity as the original 
incident cathode ray. 

1 Whiddington, Proc. Roy. Soc., 1912, A, 86, 360 ; compare Moore, Proc. 
Roy. Soc., 1915, A, 91, 330, and Proc. Phys. Soc., 1915, 27, 432. 

2 Ibid. 



1. The Phenomena of Diffraction. 

IN order to make clear the method by which the arrangement 
of atoms in crystals has been determined, it will be necessary 
to give a very brief account of the principles which under- 
lie the action of diffraction gratings. In its simplest form, a 
diffraction grating may be regarded as a plane sheet of trans- 
parent material upon which are ruled a number of close-set, 
equidistant, and parallel lines. This construction results in 
the formation of a series of fine strips capable of reflecting or 
transmitting light, each strip being separated from its neigh- 
bours by a line which neither reflects nor transmits the light- 
rays which fall upon it. 

In Fig. 7 let us assume that AB, CD, and EF represent 
sections of the opaque parts of the grating ; whilst BC and 
DE are transmittent portions of the surface ; and let us 
imagine that a series of parallel rays of homogeneous light 
fall from a slit at S upon the surface of the grating. When 
these light-rays reach the apertures BC and DE each wave- 
front becomes a centre of new vibrations in the ether which 
spread outward in all directions from the apertures as centres. 
Some of these rays will continue in a straight line perpendicu- 
lar to the grating, as shown at BN. Others, however, will 
diverge in all directions ; and for our present purpose we 
must consider the rays BL and DM, one of which starts from 
B and the other from D. These two rays are parallel and 
make an angle with the direct-passing ray BN. From D 
draw DK perpendicular to BL. 

Now imagine that these rays are intercepted by a screen 




LM placed at right angles to the direction in which the rays 
are travelling. It will be evident from the figure that the 
ray from B has to travel further than the ray from D has to 
do before reaching the screen ; since BL = BK + KL = BK + DM. 
Thus a ray starting from B has the extra distance BK to pass 
over as compared with the ray from D. Suppose that BK is 
exactly equal to half a wave-length of the light used. Then 
clearly the ray from B will arrive at L in opposite phases of 
vibration from that in which the 
ray from D reaches M : if one ar- S 

rives at L in its " crest " phase, 1 4 

the other will reach M in the 
" trough " phase. And if these 
two beams be then concentrated A B 
by a lens, the two vibrations of 
the ether will neutralize one an- 
other and darkness will result. 
On the other hand, if the length 
BK be equal to a complete wave- 
length of the light, the two vi- 
brations will arrive at L and M 
respectively in the same phase ; 
and if they are condensed by a 
lens they will form together a 
bright image of the source. 

This reasoning can be applied more generally. If BK has 
a length equal to an odd number of half wave-lengths, extinc- 
tion will result when the lens is applied : whilst if BK be equal 
to any number of complete wave-lengths the two vibrations 
will reinforce one another. 

We must now consider the effect of the angle 6 on the 
problem. From the figure, it is clear that the angle NBL is 
equal to the angle CDK; so that CDK also is. 6 and BK is 
equal to BD sin 0. BD is a fixed length, for it corresponds to 
the interval between two similar points in the grating system ; 
but since we are at liberty to choose any of the rays which 
start in all directions from B we can select any angle which 
we desire. If we pick out a particular angle which makes 
BD sin equal to an odd number of half wave-lengths, then 
we shall get extinction when the two rays from B and D are 


concentrated together ; whilst if we choose so that BD sin 
equals a complete number of wave-lengths, we shall get an 
image formed by the lens. Further, if be chosen so that 
it lies between these extreme values, partial interference will 
result; and only a faint image will be formed by the lens. 
Finally, if the wave-length of the light which we use is long 
as in red light BK also will be long when it is equal to a 
multiple of a complete wave-length ; and hence sin 6 will be 
large and 9 also will be large : so that the light of long wave- 
length will be diffracted far away from the direct beam BN. 
A ray of short wave-length, on the other hand, will be much 
less diffracted, since in its case BK is small if it is equal to a 
multiple of the wave-length. 

In practice, the light used is not homogeneous; so that, 
after diffraction, it is broken up into a series of spectra, each 
wave-length of which has its own angle for which the reinforce- 
ment of the various beams from the different apertures is 
greatest. This can be seen from a consideration of the formula 
BK = BD sin 0. Call the distance from B to D 3 and let X x 
and X 2 be the wave-lengths of light employed, then obviously 
the same value of 9 will not satisfy both conditions ; and we 
have the light diffracted at the angle X in one case and at the 
angle 9% in the second instance 

A! = S sin 9i 
X 2 = S sin 9 2 

In the foregoing reasoning we considered only one side of 
the apertures, but in practice diffraction takes place at each 
side of an aperture; so that two sets of spectra are formed, one 
on each side of the central beam. 

Finally, we shall have one spectrum formed if the. distance 
BK is equal to one wave-length ; another spectrum is produced 
when BK is equal to two wave-lengths, and so on 

nX = S sin 9 

When n is unity the corresponding spectrum is said to be of 
the first order; when n equals 2, the spectrum is termed a 
second order one, etc. 

The results obtained with an actual grating are shown in 
Fig. 8. 

The line 00 is the image formed by the direct, undiffracted 



beam BN" and those parallel to it. Since all the wave-lengths 
of the source are represented in this beam, it is of the same 
nature as the light from the source itself. On either side of 
the central beam lie the two first order spectra, VE. Outside 
them lie the second 'order spectra, VE'; and beyond them 

FIG. 8. Grating Spectrum. 

again, and partially overlapping them, will be seen the third 
order spectra, V"K". 

Hitherto we have dealt with diffraction caused by 'trans-; 
mission through the grating ; but it will be seen that the same 
reasoning holds good in the case of reflection from the surface 
of the grating. Turning again to Fig. 7 we need only imagine 
that light falls upon the grating from the source S' instead of 
from S; and that this light is absorbed by the portions AB 
and CD whilst being reflected by the parts BC and DE. The 
succeeding deductions are exactly analogous in this case to 

T -Q 

p4 \^s, 


G v 

FIG. 9. 

those which we have already drawn ; so there is no necessity 
to go into them again in detail. 

There is a third type of diffraction grating, different in 
nature from the two we have mentioned, which is specially 
important from the point of the present chapter. Let us 
assume (Fig. 9) that we have a series of parallel surfaces PQ 
RS, TV, etc., which lie a certain distance d apart from each 


other; and that they are composed of a material which is 
capable of reflecting some light and transmitting the rest of 
any beam which may fall on them. Assume, further, that 
parallel beams of homogeneous light fall upon this series of 
surfaces from the points A and C. The 'beam from A strikes 
the upper surface at B : the beam from penetrates the upper 
surface and is reflected from the second surface at F. Both 
beams will then travel along the line BX. Produce CF to G. 
G is obviously the image of B in the plane ES. Draw BD from 
B perpendicular to CG. In the figure, FB = FG ; and the angle 
ABP = the angle DBG. 

Now consider the distances travelled by the light-rays passing 
from A and C to X. In the one case, the path is AB + BX ; 
in the other case it is CF + FB 4- BX. Cancelling out the 
common part BX and substituting CD for AB and FG for FB, 
we find that the difference between the two paths is DG, which 
may be expressed as 2d sin 61. Now if X, the wave-length of 
the light we are using, be chosen equal to 2d sin 9, then the 
ray from F will reinforce the ray from B ; but if X has a value 
equal to a fraction of 2d sin 0, then interference will result and 
no clear image of the source will be formed at X. 

From this reasoning, it will be seen that if we know and 
X we can calculate d ; or, knowing d and we can calculate X. 
The application of this will be seen in later sections. In the 
meantime, it will be sufficient to point out that here also various 
orders of spectra will be formed according to the value of the 
integer n in the formula 

nX = 2d sin 

2. Space Lattices. 

The distinction between an amorphous and a crystalline 
body lies in the fact that in the former the atoms and mole- 
cules are distributed at random, whilst in a crystal there is a 
certain regularity of arrangement according to a definite system. 
The investigation of such systems has been made possible by 
the work of Laue 1 and especially by the researches of W. H. 
and W. L. Bragg. 2 

1 Laue, Friedrich, and Knipping, Sitzber. K. Akad. Munchen, 1912, 303, 363. 

2 W. H. and W. L. Bragg, X-Eays and Crystal Structure, 1915. A sum- 
mary of this is to be found in a lecture by W. H. Bragg, Trans., 1916, 109, 252. 


Before entering into the subject, it will be necessary to 
obtain a clear conception of the meaning <rf the term " space 
lattice " since these space lattices form the basis upon which 
our ideas of crystal structure will be built up. A space lattice 
is an arrangement of points in space such that the environment 
about every point is the same as that about any other point. 
Suppose, for example, that we take a series of small cubes and 
assume that the points we are considering lie at the corners of 
the cubes. If we build the cubes together into a large block, 
it will be evident that the point at one corner of one cube has 
the other points of the system arranged about it in space in a 
certain pattern ; and if we choose any other corner of any 
other cube it in turn will have other points arranged with re- 
gard to it in exactly the same manner as we found in the first 
case. In other words, each of these points has an environment 
of other points arranged in a particular way ; and since the 
whole arrangement has a certain symmetry, it does not matter 
which point we fix upon to start with. 

Now according to modern crystallographic theory, in any 
crystal the atoms of the same element are arranged so as to 
occupy the points of one or more such space lattices ; and the 
whole crystal is made up by the interpenetration of a number 
of lattices. For example, in the case of a crystal of rubidium 
chloride, the positions of the rubidium atoms will represent 
points on one space lattice ; the positions of the chlorine atoms 
will correspond to points on another space lattice: and the 
crystal will take the form of an arrangement built up by the 
interpenetration of these two space lattices. 

We may look at the matter from another point of view. 
Imagine that we plot upon several planes an identical, sym- 
metrical arrangement of points; and that we then place the 
planes parallel to one another in such a way that all the points 
form part of a symmetrical system. The points now make up 
a space lattice. If we arrange the points at the corners of 
imaginary squares upon the surface of the planes and then lay 
the planes parallel to each other so that a straight line can be 
drawn perpendicular to the planes through a set of points ; and 
if we further arrange matters so that the distance between the 
planes is equal to the length of the side of the imaginary 
squares with which we started, then obviously the points in 


the three-dimensional figure will lie at the corners of cubes. 
This is what is called a cubic lattice. On the other hand, it is 
possible to conceive of an arrangement in which the lengths 
of the sides of the parallelepiped are all different and in which 
the lines joining similar points on different planes do not cut 
the planes at right angles. In this case we have the space 
lattice corresponding to a triclinic crystal. 

3. The Crystal as a Diffraction Grating. 

For the sake of simplicity, we shall confine the present 
section to the consideration of a cubic crystal; though the 
reasoning, with appropriate modification, will apply to any 
type of crystal. 

As we have seen, a cubic crystal is assumed to be built up 
by the interpenetration of a number of space lattices, the 
points " of any one lattice being occupied by the same type 
of atom. Suppose that a crystal contains sodium chloride. 
Then, in order to determine its structure, we begin by fixing 
the relative positions of, say, the sodium atoms. The atoms 
of chlorine will be arranged symmetrically with regard to the 
sodium atoms throughout the crystal structure ; and we can 
determine their arrangement by a subsequent operation. In 
the case of the cubic system, the problem is simplified by the 
fact that only three types of lattice are possible: the simple 
cube with the atoms at its eight corners; the centred cube, 
which has an additional atom at its centre; and the face- 
centred cube with atoms at each corner of the cube and in 
the centre of each face, but with no atom in the cube's centre. 

Imagine that we could pare off from the face of the crystal 
a slice of molecular thickness. In this slice a " net " in 
crystallographic terms we should find a regular grouping of 
atoms and spaces. Suppose that we were able to cut off from 
the slice a section, also of molecular thickness a " row " as 
crystallographers name it we should find its structure to be 
roughly representable by the following : 

Each of the points would represent an atom; and the spaces 
between the points would correspond to the spaces between 
the atoms in the crystal. This arrangement at once suggests 


(From Kaye's " X-Rays"} 

FIG. 10. Pattern of Laue spots obtained by Friedrich and Knipping when 
X-rays are diffracted by a zinc-blende crystal. The incident rays are parallel to 
.a cubic axis of the crystal. 


a diffraction grating. If we could pass a ray of light through 
the crystal, the spaces between the atoms would play the part 
of the apertures in a diffraction grating ; whilst if we reflected 
a ray of light from the crystal the solid atoms would act as 
the reflecting parts of a grating and the spaces would have 
no reflecting power. Thus it appears that in crystals we have 
a series of natural diffraction gratings lying parallel to each 
other and finer than any of our artificial structures. 

In actual practice, when white light is thrown upon a 
crystal face we do not find the face acting as a diffraction 
grating. The grating is too fine as compared with the wave- 
lengths of ordinary light to allow us to obtain the effects which 
we seek. It is therefore necessary to find some type of 
vibration with a wave-length very much shorter than that 
of even ultra-violet light. This vibration was discovered by 
Laue in the form of the X-rays.* 

One important difference between the ordinary grating 
and the crystal grating must be mentioned. Owing to the 
fact that the crystal grating is three-dimensional, the rein- 
forcement of one X-ray vibration by another is not always 
possible. A ruled grating, as we have seen, reflects a con- 
tinuous spectrum when white light falls upon it ; a crystal 
grating, on the other hand, reflects only certain specific wave- 
lengths ; so that it produces a discontinuous spectrum. 

4. Laues Experiment. 

At the suggestion of Laue, Friedrich and Knipping investi- 
gated the effect of passing a fine pencil of X-rays through a 
crystal of zinc blende, the incident pencil being introduced 
parallel to an axis of symmetry of the crystal. Behind the 
crystal was placed a photographic plate ; and when this was 
developed after the exposure, it was found to bear a sym- 
metrical pattern of spots which is shown in Fig. 10. 

These spots are produced in the following manner. As the 
X-ray pulse passes over the atoms of the zinc blende, each 
atom becomes a centre of new vibrations, just as each point 

* The discovery was really due to the fact that Laue was searching for 
some means of diffracting X-rays and bethought himself of crystals as likely 
to fulfil his purpose. 


on the wave-front of a light-pulse is a centre of fresh ethereal 
movement An X-ray spectrum is thus thrown out by the 
three-dimensional grating of the crystal; but owing to the 
extra dimension, as has been pointed out, this is not a con- 
tinuous spectrum, for it consists of a few lines only. Each 
spot on a quadrant of the figure represents a particular wave- 
length ; and the central dark disc is formed by the undiffracted 
portion of the X-rays. 

Laue put forward an explanation of the phenomenon which 
was based on the assumption that the X-ray pencil was made 
up of a limited number of different wave-lengths : but his 
views have since been shown to be erroneous : and for the 
true theory of the matter we are indebted to W. L. Bragg. 

5. The X-ray Spectrometer. 

Just as an ordinary spectroscope serves for the examina- 
tion of light spectra the Bragg X-ray spectrometer 1 enables 
us to investigate the X-ray spectra of elements and compounds. 
The following- brief description will make its construction 
clear (see Fig. 11) : 

The X-rays are generated by a bulb in which the anti- 
cathode is placed perpendicular to the cathode stream; and 
the particular X-rays used are those which leave the anti- 
cathode at the grazing angle i.e. those which are travelling 
almost parallel to the surface of the anti-cathode. These rays 
pass through a fine slit in the lead casing with which the bulb 
is screened ; and a second slit is sometimes introduced in their 
path in order to obstruct any scattered radiations which may 
have filtered through the first protection. In the path of the 
ray, after it has passed through the slits, is placed the crystal 
to be examined. It is mounted on a revolving table by means 
of a lump of wax. The pencil of X-rays emerging from the 
crystal is passed into an ionization chamber which revolves on 
the same axis as that which carried the crystal. The ioniza- 
tion chamber is closed and contains a gas easily ionizable by 
X-rays, such as sulphur dioxide ; and the X-ray enters the 
chamber through an aluminium window before which a third 
slit is placed. The ionization chamber is insulated and maintained 

1 See Bragg, X-Bays and Crystal Structure, chapter iii, 


(Front Kaye s " X-Rays.") 

FIG. 11. Photograph of Bragg's X-ray spectrometer. B is a lead box 
containing an X-ray bulb. Sj and S 2 are adjustable slits which direct a beam 
of X-rays on to the face of the crystal C. The reflected beam passes through 
the slit S ;5 into the ionisation chamber I, where it is recorded by the tilted 
electrocsope in the metal box E. K is an earthing key ; M a mirror for 
illuminating the electroscope. C and I can each be rotated about the axis of 
the spectrometer. 


at a high potential by means of storage cells. It contains an 
electrode mounted just out of the path of X-rays entering 
the chamber ; and this electrode is connected with a Wilson 
tilted leaf electroscope. 

The operation of the apparatus is almost self-evident. 
Suppose that the crystal be kept fixed and the X-ray stream is 
turned on. The ionization chamber is then moved slowly 
round the axis ; and whenever a strong X-ray penetrates the 
chamber, the gas in the latter becomes ionized and the 
electroscope is affected. The angle of the ionization chamber 
is read off; and from it the wave-length of that particular ray 
can be determined in terms of the crystal grating. The 
ionization chamber is then rotated through a further angle 
until the electroscope is again affected, showing that a new 
X-ray is passing through the chamber. It will be evident that 
the base of the ionization chamber is moving along the 
circumference of a circle over which the X-ray spectrum is 
being distributed by the crystal grating. 

The position in which the ionization chamber rests when 
ionization is observed gives a means of calculating the angle 
at which the X-ray has been refracted from the crystal ; and 
the degree of ionization determined by the rate of discharge 
of the electroscope is a measure of the intensity of the parti- 
cular line which is entering the- spectrometer. 

6. The Analysis of Crystal Structure. 

Again, for the sake of simplicity, we shall confine our at- 
tention to cubic crystals. Such crystals are assumed to be 
built up from a series of elementary cubes ; and the pattern 
on which they are constructed must fulfil the requirements of 
cubic symmetry. Now if we examine the conditions which 
fulfil these demands, we shall find that there are three pos- 
sible arrangements of the characteristic points within the 
elementary cube 

Type 1. The points lie at each corner of the cube. 

Type 2. There are points at each corner of the cube and 
also a point at the centre of the cube. 

Type 3. There are points at the corners and also in the 
centre of each cube face. 


In order to determine the nature of a crystal we must de- 
fine two factors 

A. The space lattice of each element in the crystal. 

B. The nature of the interpenetration of the various space 


The results which were obtained in the cases of sodium 
chloride and the chloride, bromide, and iodide of potassium 
will serve as an example of the method employed. From the 
similarity in chemical character of the substances, it might ap- 
pear probable that all the salts would give the same Laue 
pattern ; but when the examination of them was made it was 

(From Kaye's " X-Rays ".) 

FIG. 12. Representation of two types of diffracting centres in a 
cubic crystal. 

found that they differed markedly from one another. Potas- 
sium chloride gave rise to a simple pattern of spots such as 
would be formed by a cubic lattice of the first type. Potas- 
sium bromide and potassium iodide produced a spot-pattern 
characteristic of the face- centred lattice described in Type 3. 
Sodium chloride yielded a pattern more complex than either 
of the others and apparently intermediate * between them. 
How can we explain this ? 

Let us examine the system illustrated in Fig. 12. 

* Strictly speaking, sodium chloride is more like potassium bromide ; but 
the small difference between the diffracting powers of sodium and chlorine 
atoms produces an apparent resemblance to the potassium chloride struct ure. 


In this diagram, the characteristic points are represented 
by light and dark circles. Let us first ignore the distinction 
between the two and look at the figure as a whole. It will be 
evident that it represents a space lattice of the first type, 
since all the points are situated at the corners of cubes. Now 
omit from consideration the white circles, and it will be found 
that the black circles are placed at the corners of the large 
cube and also in the centres of the faces of that cube. In 
other words, this arrangement represents a space lattice of the 
third type. Further, the white circles taken alone form parts 
of two exactly similar groupings. 

We must now endeavour to interpret this in terms of the 
results given by the salts. Let us make a single assumption : 
that a heavy atom forms a better centre of diffraction than a 
light atom does. This assumption leads us to draw a distinc- 
tion between the constituents of the various salts which have 
been described ; for we should expect to find a marked differ- 
ence between the diffraction effects of potassium (atomic weight 
39) and iodine (atomic weight 127) ; but only a slight differ- 
ence between those of potassium and chlorine (atomic weight 

Take first the case of potassium chloride. Here the two 
constituent atoms are of approximately equal weight; and 
their diffractive powers, in accordance with our assumption, 
will be of almost equal value. The Laue figure produced will 
depend upon the joint action of the two. If we return to 
Fig. 12 we shall see that when the black and white circles 
are of equal value, the whole system becomes a Type 1 lattice ; 
so that it is not a far-fetched assumption if we regard the 
potassium chloride crystal as being built up in this manner, 
the two different kinds of circle representing the potassium 
and chlorine atoms. 

Now take the case of potassium bromide and potassium 
iodide. Here we have the light potassium atom and the com- 
paratively heavy bromine and iodine atoms. The Laue figure 
produced by these salts will be influenced mainly by the heavy 
atoms. We find that the pattern actually produced is cha- 
racteristic of a face-centred lattice; and from this we must 
deduce that the halogen atoms occupy the places indicated by 
the black circles in Fig. 12. 


Finally, in the case of sodium chloride, the difference in 
atomic weight between metal and halogen is not so great as 
in potassium iodide but is greater than in potassium chloride ; 
and hence, according to our assumption, there will be an im- 
press of both sodium and chlorine effects upon the plate in the 
Laue pattern. But the result of this will be to give us a 
pattern which will be akin to the face-centred variety, if we 
regard the chlorine atoms only, or will be related to the cubic 
lattice if we assume that chlorine and sodium have approxi- 
mately equal effects. In other words, we shall have a mixture 

(From Bragg's " X-Rays" (Oeo Bell & Sons, Ltd.).) 
FIG 13. 

of patterns which corresponds exactly to the experimental 
result. Eubidium bromide and caesium iodide should give the 
same pattern as potassium chloride ; and an examination of 
these two substances will be of much interest. 

There seems to be little doubt that the assumption of the 
space lattice shown in Fig. 12 as a representation of the alkali 
halides, is fully justified. 

The results obtained with other crystals may be described 
very briefly. Curiously enough, the face-centred lattice ap- 
pears to be a very common one ; for the crystals of copper, 
silyer, zinc blende, iron pyrites, fluorspar, and the diamond 
all appear to be constructed upon this system or by the inter- 
penetration of more than one such system. 

The zinc blende crystal is composed of a face-centred 



lattice of zinc atoms combined with a face-centred lattice of 
sulphur atoms, the latter grouping proceeding from coincidence 
with the zinc one along a diagonal of the face-centred cube to 
approximately a quarter of the length of the diagonal. 

Diamond is an exceptionally interesting crystal from the 
point of view of chemical theory. In the model illustrated in 
Figs. 13 and 14, it will be seen that every atom stands at the 
centre of gravity of a tetrahedron formed by its four nearest 
neighbours ; and, further, that the atoms are arranged in rings 

(From Bragg' s " X-Rays " (Geo. Bell & Sons, Ltd.).) 
FIG. 14. 

of six throughout the structure. The suggestiveness of this 
grouping in connection with the benzene theory 1 and the 
tetrahedral arrangement of optically active carbon compounds 
requires no elaboration here. From the point of view of the 
space lattices, it' is sufficient to note that the diamond arrange- 
ment may be considered as derived from the zinc blende con- 
struction by the substitution of two similar atoms for the 
dissimilar zinc and sulphur atoms of zinc blende. 

In calcium fluoride, the fluorine atoms lie at the centres of 
small cubes into which the larger face-centred cubes of calcium 

1 See Collie, Trans., 1916, 109, 561. 


can be divided. The grouping in iron pyrites cannot be made 
clear without a figure. 1 

The copper crystal is perhaps the simplest of all. It is 
based on a face-centred lattice ; and the copper atoms occupy 
the same positions relative to one another as would be taken 
up by a series of shot piled upon one another. 

We must now turn to another unknown in the problem : 
the several distances which separate the atomic planes lying 
parallel to the crystal faces. The reasoning in this case is 
rather more complicated. 2 We begin by deciding to use 
throughout the experiments one particular wave-length of 
X-ray, say the B peak from a platinum anti-cathode. In the 

X = 2d sin 9 

the only factor which we can determine directly is 0, which 
we can measure by means of the X-ray spectrometer ; the other 
two factors, d and X are left unevaluated for the present. 

Now if we apply the X-ray successively to the three 
primary planes of a crystal, we can obtain three different 

values of the expression ^ Call these three values -A ~^ and 

A A A 

^- 3 . Since X is constant throughout, these values give us the 

relative lengths of the sides of the elementary parallelepiped 
from which the crystal is built up. 

Further, the volume of this elementary parallelepiped is 

d-, x d 2 x do .. , V T . 

given by .3 - which is simply p. Let us take 

another step. If the density of the crystal be p, then the 
mass associated with each elementary parallelepiped will be 
V/o ; and if M be the molecular weight of the substance, the 

number of molecules in each parallelepiped is ^. Write this 

in the form X 3 (~ 1L); then the quantity within the brackets 
\X 3 ' M/ 

is proportional to the number of molecules in each parallele- 

1 See Trans., 1916, 109, 265. 

2 For the following simple method of presentation I am indebted to Kaye's 
X-Rays, p. 221 ff. 


When the value of ( -p . ^= j is calculated from the experi- 

mental data for such different substances as sodium chloride, 
zinc sulphide, calcium fluoride, calcium carbonate, and iron 
pyrites, it is found that in each case it approximates to unity. 
In the case of sodium chloride, for example, it is 1*22. This 
points to there being only a single molecule associated with 
each diffracting centre ; and if we combine this with the reason- 
ing already given with regard to the nature of the space lat- 
tices of the alkali salts, it seems reasonable to deduce further 
that there is only one heavy atom associated with each charac- 
teristic point in the lattice. Support for this view is found 
in the case of potassium chloride, for- when its constant is 
calculated it is found to be 0'605; which implies that there is 
only half a molecule i.e. either a potassium or a chlorine 
atom at each characteristic point in the lattice. It will be 
recalled that this conclusion has already been drawn from 
data of a totally different type. 

If the foregoing reasoning be correct, then obviously 

~ = 1 ; and we can write 


In the case of sodium chloride we have 
Molecular weight, M = 58'5 X 1/64 X 10' 24 grs.* 
Density p = 2'15 gr./c.c. 

= 33*3 (determined experimentally). 

A 3 (33-3 x 215) = 58-5 x 1-64 X 10' 24 

and X = 1-10 X 10- 24 . 

This gives the wave-length of the homogeneous B platinum 
X-ray. From this we can now calculate the value of d in the 
equation A = 2d sin 9. 

In the case of graphite, a measurement made on the cleav- 
age basal plane showed that in this instance d is 3'42 X 10~ 8 cm. 
In the calcite crystal a measurement of the plane (100) yielded 
3-04 x 10" 8 cm. for the value of d ; whilst for the planes (110) 
d was found to be 1 917 X 10~ 8 cm. 1 

* The mass of a hydrogen atom is taken as 1*64 x 10- 24 grammes. 
1 Bragg, X-Rays and Crystal Structure, pp. 115-16. 




1. The Method of Measuring X-ray Spectra. 

THE work of W. H. and W. L. Bragg, described in the last 
chapter, provided a means of determining the wave-length of 
any particular X-ray, and thus opened the way to an examina- 
tion of the X-ray spectrum emitted when a given substance 
is excited by the ordinary methods. The investigations of 
Moseley 1 in this field resulted in the discovery of a new 
atomic relation which appears to be of fundamental impor- 
tance ; and their extension by other workers 2 has thrown 
light upon a number of points of great interest. 

The method employed by Moseley was as follows. The 
stream of electrons from the cathode of an X-ray tube was 
concentrated upon a small area of an anti-cathode composed 
of the element whose high-frequency spectrum it was desired 
to examine. The characteristic rays thus set up were allowed 
to pass through a slit in a platinum plate, after which they 
emerged from the X-ray tube through an aluminium window. 
In their path, inclined at a certain angle, was placed the 
cleavage face of a crystal of potassium ferrocyanide which 
analysed the X-ray beam. After being diffracted by the 
crystal, the rays were allowed to fall upon a photographic 
plate ; and in this way photographs of the X-ray spectrum were 

1 Moseley, Phil. Mag., 1913, 26, 1024 ; 1914, 27, 703. Lieutenant Moseley 
was killed during the Dardanelles campaign, 1915. 

2 de Broglie, Compt. rend., 1913, 156, 1153 ; 1916, 165, 87, 352 ; de Broglie 
and Lindemann, ibid., 1913, 156, 14B1 ; Herweg, Verh. deut. phys. Ges., 1914, 
16, 73; Siegbahn and Friman, Physikal Zeitsch., 1916, 17, 17, 48, 61; 
Siegbahn and Stenstrom, ibid., 318 ; Siegbahn, Ber. deut. phys. Ges., 1916, 
18, 39, 150, 278; Siegbahn and Friman, Phil. Mag., 1916, 31, 403; Friman, 
ibid., 1916, 32, 497 ; cf. Rutherford and Andrade, ibid., 1914, 27, 854. 



*- Increasing Wave Length. 

(From Kayt? s " X-Rays.") 

FIG 15. Moseley's photographs of the X-ray or high-frequency spectra 
of a number of metallic anticathodes. The spectra, which are the third 
order, are placed approximately in register in the figure. For each metal, the 
more intense line, with the longer wave-length, is the K characteristic radiation. 
The hrass shows the Zn and Cu lines ; the cobalt contained both nickel and 
iron as impurities. 


obtained. Owing to the weakness of certain of the radiations, 
the spectrometer was constructed so as to be capable of evacua- 
tion; and it thus became possible to photograph even those 
X-rays which would have been intercepted by a layer of air. 

From the photographs thus obtained, the wave-lengths (or 
the frequencies) of the X-rays emitted by the anti-cathode 
could be conveniently measured. 

2. The Nature of the X-ray Spectra. 

It will be remembered that the characteristic rays emitted 
by various elements fall into series which are termed the K, 
L and M series. Moseley showed that the K spectrum of all 
the elements contains two strong lines ; and an examination 
of Fig. 15 will show that a certain regularity can be 
traced as we pass from element to element. The strongest 
line in the spectrum is termed the a-line and the weaker one 
is named the j3-line. Inspection of the figure will suffice to 
prove that in the case of the cobalt spectrum the original 
specimen has been contaminated with nickel and iron (since 
the spectrum contains weak lines characteristic of these ele- 
ments) ; whilst the nickel spectrum contains traces of the 
copper spectrum also. 

Later work has shown that the spectra are not quite so 
simple as they were at first supposed to be; for more lines 
have been detected which were missed by Moseley. The 
spectra obtained when X-rays of the L series are used instead 
of the K series are still less simple. But even the most 
complex X-ray spectra are simplicity itself compared with 
ordinary light spectra; and the mathematical analysis of 
them is therefore much more likely to yield us definite in- 
formation with regard to the systems of vibration which 
produce them. 

In the case of gaseous elements such as chlorine, it might 
be supposed that experimental difficulties would hinder the 
determination of the X-ray spectrum ; but it has been found 
that such elements can be dealt with by utilizing a compound 
instead of the element itself. Potassium chloride, for example, 
shows the line spectrum of potassium and in addition it 
exhibits certain non- potassium lines which are evidently due 


to the chlorine atom in the molecule. In this way it is pos- 
sible to establish the high-frequency spectra of gaseous and 
liquid elements ; though of course the method breaks down in 
the case of the inert gases which form no compounds. 

We must now turn to the problem presented by the 
frequencies of the a-lines in the spectra of various elements. 
If we calculate for each element the factor Q from the 

Q = 

I 2 2 2 


in which v is the frequency of the a-line in the X-ray spectrum 
of the element in question, whilst v is the fundamental 
frequency of ordinary line spectra which was obtained by 
Bydberg,* there is obviously a certain regularity in the 


Atomic Weight. 

Atomic Number. 

















Iron . 




Examination of the figures shows that there is no direct 
relation between the factor Q and the atomic weight for any 
given element.! At the same time, there is a certain regularity 
in the values for Q; for each of them differs approximately by 
unity from its neighbours above and below. Further, the 
elements, when arranged in order of their Q values, lie in the 
order which is suggested by their chemical properties. 

If we now turn to the Periodic Table and, beginning with 

* It may be noted in passing that Moseley used Rydberg's earlier result, 
(v = 109,720) instead of the more recent value v = 109,675, but the differ- 
ence between the two is slight. 

f The relationship : Atomic weight = 2Q + 6-12, which might be deduced 
from the figures given above, breaks down when the series is extended. E.g. 
for silver, Q = 46-6, whence 2Q + 6-12 would be 99'32 instead of 107-88. 
Harkins and Wilson (J. Amer. Chem. Soc., 1915, 37, 1396) have proposed the 
formula : Atomic weight = 2(N + ri) + % + ( l)*-i in which N is the 
atomic number and ri is zero for the lighter elements. 


hydrogen as No. 1, count along the Table until we arrive at 
titanium, we shall find that it is the twenty-second element in 
the series ; vanadium is the twenty-third ; chromium is the 
twenty-fourth, and so on. Here also we have a series of 
figures each differing from its neighbours by unity; and the 
arrangement of the elements in each of the two series is the 
same. The numbers obtained by counting along the Table 
from hydrogen are termed Atomic Numbers. 1 Inspection will 
show that for any element in the above series the factor Q 
and the atomic number N are approximately related thus 

N = Q + 1 

Looking at the matter from a more general point of view, 
it is clear that Q, v, and VQ in the equation given above are 
related to one another in the following manner : 

v = |y. x Q 2 

In the case of the L series spectra, the relationship takes 
the form 

,2 2 3 2 

so that in that case the relation between Q and the other 
factors become 

v = & . v X Q 2 

Evidently the most general way of writing the relationship 

and since Q is a factor which increases concurrently with the 
atomic number N, we are entitled to substitute N - b for Q, if 
6 is a constant. The relation between the frequency of the 
a-line of an element and the atomic number of the element 
thus becomes simply 

and in the case of the K line it can be seen from the figures 

given above that 6 = 1; so that the expression is reduced to 

v = A(N - I) 2 

1 van den Broek, Phi/sikal Zeitsch., 1913, 14, 33 ; Nature, 1913, 93, 373, 
476; Soddy, ibid., 399, 452 ; Butherford, ibid., 423 ; Bohr, Phil. Mag., 1913, 


Moseley 1 suggested that the integer N" was identifiable 
with the number of free positive units of electricity contained 
in the atomic nucleus, which brings his work into line with 
previous view of Soddy 's that the chemical nature of an 
element depends, not on its atomic weight, but upon its 
electrical character 2 : the mass of the atom being regarded as 
of only secondary importance. 

3. The Gaps in the Periodic Table. 

"We must now turn to another region of the subject. It is 
clear that if we determine the atomic numbers of any two 
elements from their X-ray spectra we are in a position to state 
the number of other elements which are interposed between 
them. For example, in the table given on p. 116 titanium is 
the twenty-second element and iron is the twenty-sixth ; there- 
fore there must be three elements between titanium and iron, 
corresponding to the numbers 23, 24, and 25. These elements 
are of course, vanadium, chromium, and manganese. 

The atomic number of uranium has been found to be 92, so 
that there can be only ninety elements between it and hydrogen. 
Of these ninety, we already know eighty-four, so that there are 
still six to be accounted for. A glance at the Periodic Table 
will show that there is a blank space between molybdenum 
and ruthenium, where an element with an atomic number 43 
may be expected. Another space lies between tungsten and 
osmium, corresponding to the atomic number 75. A third 
blank in the table is located between polonium and niton ; and 
a fourth between niton and radium. This leaves us with a 
couple of unknown elements still to be allocated. Moseley 
assumed that thulium was a mixture of two elements, one of 
which is thulium proper and the other is the celtium of Urbain ; 
and he placed the remaining missing element, No. 61, between 
neodymium and samarium. 3 

The importance of X-ray spectra in this field can hardly be 
over-estimated. The rare earth group of elements has long 

1 Moseley, Nature, January 5th, 1914. 

2 Soddy, Jahrb. RadioaUiv. Elektronik, 1913, 10, 193 ; Nature, December 
4th and 18th, 1913 ; Chemistry of the Radio-Elements, II., 6 (1914). 

8 See Soddy, Chemical Society's Annual Reports, 1914, XI., 279, for 
information supplementing Moseley's papers in the Phil. Mag. 


been one of the puzzles of chemistry ; and the definite proof of 
the exact number of these substances still remaining to be 
discovered has removed the possibility that much time and 
labour might be wasted in seeking for fresh members of the 
group after the roll-call was complete. The rare earth group 
stands apart from the remainder of the Periodic Table and 
contains within itself no indication of the number of elements 
which it includes ; so that without Moseley's key we might 
have seen many fruitless attempts to discover " missing " 
members long after the whole flock was gathered together in 
the fold. 

4. The Anomalies of ike Periodic Table. 

It is we]l known that the Periodic Table is not free from 
blemishes in its arrangement. Argon and potassium do not 
follow one another in the order of their atomic weight ; nor is 
the chemical sequence of iron, cobalt, and nickel the same as 
the order of increasing atomic weight. Iodine and tellurium 
are also inverted in the usual form of the Table. 

Moseley's work has shown that the atomic numbers run 
parallel to the chemical sequence of the elements and not to 
their order of atomic weights. For example, the X-ray spectra 
show that the value of N for potassium is 19, which brings it 
into Group I as usual, instead of placing it in Group as the 
atomic weight sequence does. Also, the atomic order of the iron 
group is : iron (26) ; cobalt (27) ; and nickel (28) : which agrees 
with their chemical behaviour and not with their arrangement 
by increasing atomic weight. Tellurium and iodine also fall 
into their chemical order. This supports what has already been 
said with regard to the importance of the atomic numbers from 
the chemical standpoint. 

Another point of interest arises in connection with certain 
" elements " which, on stellar spectroscopic evidence, have been 
supposed to exist. Coronium, nebulium, and asterium have all 
been taken for granted on the strength of certain lines which 
were observed in the spectra of the heavenly bodies ; and places 
were assigned to elements of this class in a special series be- 
tween hydrogen and lithium. If Moseley's conclusions are 
correct, there can be no such series ; and evidence of the 


existence of the celestial elements will in future be regarded 
with considerable suspicion.* 

A crucial test of the correctness of Moseley's atomic num- 
bers was furnished by the case of the isotopes.f These are 
elements having different atomic weights, but possessing pro- 
perties so similar on the chemical side that they cannot be 
separated from one another by any reactions. If the atomic 
number is a function of that portion of the atomic structure 
which manifests itself in chemical properties, then clearly the 
spectra of two isotopes ought to be identical. The matter has 
been put to the test by Eutherford and Andrade. 1 Lead and 
radium-B are chemically inseparable from one another ; and 
when their X-ray spectra were compared, it was found that 
they were identical. Thus each group of chemically insepar- 
able elements has a single atomic number covering the whole 
group and the individual isotopes are not entitled to separate 

5. The Spectrum of Hydrogen. 

An interesting point arises when we consider the X-ray 
spectrum of hydrogen. This has not been observed; and it 
cannot be calculated exactly from Moseley's formula because 
when N is put equal to unity in that equation, the quantity 
within the bracket becomes zero and no physical meaning can 
be attached to the result. The information can, however, be 
obtained by another method which leads to an approximation. 

Moseley suggested that the square root of the frequency of 
the radiation was probably proportional to the nuclear charge 
on the atom ; and for nuclear charge we may substitute atomic 
number. In the present case it will be convenient to substitute 
for the frequency its reciprocal, the wave-length of the vibration. 

* It must, however, be borne in mind that if the " astronomical elements " 
are isotopes (see Chapter XVII.) of the elements in the first series of the Table, 
the above conclusion could not be sustained, since two isotopes have the same 
atomic number ; but if the spectra of two isotopes be identical (see p. 200), the 
existence of the new celestial elements isotopic with known elements can 
hardly be supported by spectroscopic evidence. 

t See Chapter XVIII. 

1 Eutherford and Andrade, Phil Mag., 1914, 27, 854 ; see also ibid., 1916, 
32, 49. 


We thus get the following equation, which is approximately 
true : 

Now in this equation we replace N 2 by 74, the atomic number 
of tungsten ; X 2 we replace by 0'185 X 10~ 8 cm., which is the 
wave-length of the K radiation of tungsten ; and we substitute 
1, the atomic number of hydrogen, for N r This gives us the 
value of A! which proves to be 101/3 /^u.* In other words, 
the wave-length of the K radiation emitted by the hydrogen 
atom should approximate to 101*3 JJ./UL. 

Now the shortest wave-length discovered by Lyman in the 
remote ultra-violet spectrum of hydrogen is 91 '2 ^ ; and 
there is good reason for supposing that this is the shortest 
vibration of which a hydrogen atom is capable. Bearing in 
mind the approximate character of the calculation we have 
made, the agreement in value between the experimentally de- 
termined ultra-violet wave-length and the wave-length calcu- 
lated for the K radiation of hydrogen seems to suggest that 
the ultra-violet spectrum of hydrogen is simply the X-ray 
spectrum of the gas. For the L radiation a similar calculation 
can be made, which brings out a value closely approximating 
to the ordinary visible spectrum of hydrogen. 

From this it follows that in all probability the ordinary 
radiations of hydrogen are neither more nor less than the 
element's X-ray spectrum. Very little thought will show how 
important this conclusion may be from the point of view of 
our knowledge of the relations between atomic constitution 
and spectral series. 

6. Molecular Numbers. 

Allen J has suggested that just as we have atomic weights 
and molecular weights, so we ought to consider atomic numbers 
and " molecular numbers," the latter being obtained by adding 
together the atomic numbers of the atoms in any molecule just 
as the molecular weight is got by taking the sum of the atomic 
weights of the atoms composing the molecule. The molecular 

* The expression /*/* represents a millionth of a millimetre. 
1 Allen, Trans., 1918, 113, 389. 


number would represent the number of positive charges on the 
various atomic nuclei in the molecule. Allen points out that 
if we consider the four compounds : CH 4 , NH 3 , H 2 0, and HF 
which are formed by the combination of hydrogen with a 
typical element belonging to successive groups in the Periodic 
Table, all four compounds have the same molecular number, 10. 
Similarly all the compounds : SiH 4 , PH 3 , H 2 S, and HC1 have 
the molecular number 18 in common. The two series differ 
by eight units ; and this difference, or one of eighteen units, 
makes its appearance frequently in comparisons among mole- 
cular numbers. 

A point of some interest is found when the molecular 
number of ammonium is calculated. It is found to be 11, 
which is identical with the atomic number of sodium. In 
view of the resemblances between the sodium and ammonium 
ions, the coincidence of the values is striking. 

It is too early yet to express any ojAiion on the value of 
the molecular numbers in chemical problems ; but when they 
have been more exhaustively studied it seems possible that 
valuable results may be deduced from their relations. 

7. Conclusion. 

Enough has now been said to indicate the importance of 
Moseley's work ; and it seems probable that we cannot at pre- 
sent appreciate the developments which may arise from it in 
the near future. It is true that X-ray spectra have, on further 
investigation, proved to be more complicated than was at first 
supposed ; but even at their greatest complexity they are much 
simpler than the corresponding emission spectra of the ele- 
ments ; and their mathematical treatment, complex as it may 
prove to be, will certainly be a much easier problem than the 
attack upon the luminous spectra has been. 

The relation between the X-ray spectra and the model 
atom of the physicists is of considerable interest; but a dis- 
cussion of the question would lead us far beyond the limits 
laid down for the present chapter. 

Allen 1 has shown that there is a simple relation connecting 
the atomic number N, the characteristic atomic frequency v, 

1 Allen, Proc. Boy. Soc., 1917, A, 94, 100 ; Phil. Mag., 1917, 34, 488. 


and a fundamental frequency V A = 21'3 X 10 1 sec" 1 . He finds 
that this relationship can be expressed by the equation 

Nv = nv A 

The value n, which he terms the frequency number, is always 
an integral. For the applications of this to problems of atomic 
heats and photo-electric phenomena, the original papers should 
be consulted. 

On the purely chemical side, the support lent by Moseley's 
work to the Periodic Arrangement of the elements cannot but 
give rise to much speculation. To what particular property 
does the atomic number correspond ? 

In the first place, it has no traceable relation to the 
maximum valency exhibited by an element ; for in the rare 
earth group we find a set of elements all of which are ob- 
viously trivalent, and yet their atomic numbers run in sequence 
just as if they were characteristic elements of the first two 
series in the Periodic Arrangement. 

Secondly, the fact that lead and radium-B have the same 
X-ray spectrum, though their atomic weights are not identical, 
appears to prove that there is no direct relationship between 
the atomic weight and the atomic number. 

Thirdly, since the radio-active properties of lead and 
radium-B are different, it is clear that the atomic number 
is not affected by those factors in an atom which give rise to 
radio-active phenomena. 

This leaves us with only one characteristic of the elements 
undiscarded : the chemical properties. But here we are still 
left in doubt. The chemical properties of ferrous iron and 
ferric iron are quite different ; so are the properties of copper 
in the cuprous and cupric forms. Can we assume that the 
atomic number of ferrous iron is the same as that of ferric 
iron ? If so, then the atomic number cannot be regarded as a 
guide to chemical sequence in the strictest form. It will be 
of considerable interest to see the results which are obtained 
with the X-ray spectra of elements in the " ous " and " ic " 
varieties ; for it seems probable than an investigation in this 
direction will enable us still further to limit our search for 
the property which corresponds most closely to the atomic 



1. Introductory. 

IN the atomic order between barium and tantalum there is a 
group of elements whose atomic weights lie between 130 and 
180. These substances, together with scandium and yttrium, 
make up the greater part of what are termed the rare earths.* 
These rare earths are found for the most part in the Scandi- 
navian peninsula, but deposits have also been brought to light 
in the Ural Mountains, and in certain parts of America and 
Australia. It appears that rare earths are usually found 
among eruptive rather than among sedimentary deposits, and 
that they are most likely to occur in ancient igneous rocks, 
particularly in granite. 

The rare earths are chiefly made up of a series of basic 
oxides, but it is usual to divide them into three groups : the 
cerium group, the terbium group, and the ytterbium group. 
This division is arbitrary, for the rare earth elements resemble 
each other very closely indeed, and we have really a single 
series of substances with properties changing slightly as we 
pass from member to member, rather than three groups whose 
properties differ sharply from each other. f Still, for the sake of 
convenience, the following arrangement is usually employed : 

* The term " rare earths " is a misnomer in modern times, as it has been 
found that these substances are really very widely distributed in nature and 
extensive deposits of them are known at the present day. Detailed accounts 
of the rare earth group are to be found in Friend's Textbook of Inorganic 
Chemistry, vol. iv., by H. V. F. Little, and in Spencer, The Bare Earth. 

t The solubility of double salts is the chief criterion applied in arranging 
the elements into these three groups. 




Cerium Group. 

Terbium Group. 

Ytterbium Group. 

Atomic Atomic 
Number. Weight. 
57. Lanthanum 139-0 
58, Cerium . 140-25 
59. Praseodymium 140-9 
60. Neodymium 144*3 
61. ? 
62. Samarium 150'4 

Atomic Atomic 
Number. Weight. 
63. Europium . 152*0 
64. Gadolinium 157-3 
65. Terbium . 159-2 

Atomic Atomic 
Number. Weight. 
66. Dysprosium 162-5 
67. Holmium 163-5 
68. Erbium 167-7 
69. Thulium 168-5 
70. Ytterbium 173-5 
71. Lutecium 175 
72. Celtium- ? 
39. Yttrium 88-7 

Assuming that Urbain's celtium has the atomic number 72, 
this leaves only one rare earth still unknown, and it should 
correspond to the blank position, No. 61. 

The group of the rare earths has considerable importance, 
whether we look upon the question from the point of view of 
theory or from the practical standpoint. With the theoretical 
problem involved we shall deal in Section 5 of Chapter XIX, 
so we need not enter here into the question of the position 
which must be assigned to these elements in the Periodic 

On the practical side, the rare earths were without any 
importance until a quarter of a century ago ; but in 1884, 
Auer von Welsbach took out a patent in which the present- 
day incandescent gas mantle was foreshadowed, and at once 
the question of the rare earths became a commercial one. At 
that time the known deposits were limited, both in extent and 
number, and it consequently was doubtful whether the raw 
material for. the manufacture of mantles could be produced 
at a price sufficiently low to allow of commercial success. 
Fortunately, large deposits of monazite sand were soon after- 
wards discovered in Carolina and in Brazil. These beds 
rendered it possible to obtain the rare earths comparatively 
cheaply, and the incandescent gas-lighting industry has grown 
up on them. 

At this point the theoretical side has shared in the profit, 
for the mantle industry has produced vast quantities of by- 
products which have been placed at the disposal of investi- 
gators, who would otherwise have been forced to work up the 
crude material for themselves at the cost of much time and 


labour. At the present day, these by-products form one of 
the most fruitful sources in the investigation of the elements 
of the rare earth series. 

2. The Physical Properties of the Rare Earth Elements. 

Muthmann and Weiss * have carried out a comparison 
between some of the members of this group of substances, and 
have obtained the following results : 

As regards colour, lanthanum is a white metal similar to tin 
in appearance ; cerium is rather more like iron ; neodymium 
shows a faint tinge of yellow; while praseodymium is more 
pronouncedly yellow in ti,;fc. 

When the hardness of these elements is compared with that 
of lead, tin, and zinc, they lie in the following order : lead, 
tin, cerium, lanthanum, zinc, neodymium, praseodymium. 
Samarium is harder than any of the foregoing. 

The melting-points are shown in the following table. Alu- 
minium melts at about 660 C., and silver at about 960 C., 
which gives a rough standard of comparison : 


Cerium. . . . . 623 

Lanthanum .... .... 810 

Neodymium 840 

Praseodymium . 940 

The specific gravities of the four metals are 

Lanthanum 4 6-15 

Praseodymium ,6*48 

Neodymium 6*96 

Cerium 7*04 

Cerium, therefore, has about the same density as tin (7'3). 

3. The General Chemical Character of the Rare Earth 

The metals of the rare earths are comparatively reactive, 
and enter into direct combination with various elements. They 
burn in air,* giving rise to oxides of the type Me 2 3 ; but in 
addition to this series, some of the metals can be converted 

1 Muthmann and Weiss, Annalen, 1904, 331, 1. 

* Cerium burns with a brilliance exceeding even that of magnesium. 


into oxides having the composition Me0 2 . For example, cerium 
oxide of the formula Ce 2 3 is unstable, the stable oxide being 
Ce0 2 . The corresponding oxide of praseodymium has also been 
isolated, but it is much less stable than the cerium one. Ter- 
bium and neodymium also yield higher oxides having the com- 
positions Tb 4 07 and Nd0 2 . It is a curious fact that there are 
two forms of the oxides Me 2 3 , one form being prepared by 
heating the nitrate of the metal, the other form being obtained 
by igniting the hydroxide. The two varieties are different in 
both physical and chemical properties. It appears likely that 
this case is parallel to that of the two forms of calcium oxide 
(unslaked lime and overburned lime) and to the two varieties 
of magnesium oxide. The origin of the difference in properties 
probably lies in the existence of two different polymorphous 
forms of the oxides. 

The oxides of the rare earth metals are bases of medium 
strength ; the order of basicity appears to be the following, the 
strongest bases being placed first : 

La 2 3 , Pr 2 3 , Nd 2 3 , Ce 2 3 , Y 2 3 , Sm 2 3 , Gd 2 3 , Tb 2 3 , 
Ho 2 3 , Er 2 3 , Tm 2 3 , Yb 2 3 , Sc 2 3 , Ce0 2 

The behaviour of ignited lanthanum oxide resembles that of 
quicklime ; it hisses when placed in contact with water, and ab- 
sorbs carbon dioxide from the air. The oxides of the ytterbium 
series are much less active in this respect. From the fact that 
their salts with strong acids are not measurably hydrolysed 
even in dilute solution, it follows that the oxides of the cerium 
group are the strongest bases derived from trivalent elements. 

The rare earth metals combine directly with hydrogen and 
nitrogen as well as with oxygen. The hydrides have the 
general formula MeH 2 or MeH 3 , and are obtained either by 
passing hydrogen over the metals at 200 C. to 300 C., or 
by the action of magnesium upon the metallic oxides in a 
stream of hydrogen. The nitrides are formed by the action 
of nitrogen upon the oxide in presence of magnesium, or from 
the carbides by the action of ammonia. They have the com- 
position MeN". 

The carbides of the rare earth series have the general 
formula MeC 2 . They are produced by the electrolytic reduction 
of the oxides in presence of carbon ; and when treated with 


water they yield mixtures of hydrogen with acetylene and 
other hydrocarbons. 

It is unnecessary for us to enter into details with regard to 
the salts of these metals. They present no special character- 
istics. The rare earth elements form salts with nearly all the 
ordinary organic and inorganic acids, and these salts possess to 
a great extent the ordinary properties of salts of strong bases. 

4. Methods of Purifying the Rare Earths. 

In the present section no attempt will be made to enter 
into any great detail in the description of the various methods 
employed to divide the rare earths from the minerals among 
which they occur, and to separate the elements of this group 
from one another. For these details the reader is referred to 
larger treatises. 1 All that we can do in this place is to indicate 
briefly the main lines of the separations, choosing as far as 
possible the more characteristic of these. 

There are four chief stages in the isolation of the rare earths 
from the deposits in which they occur naturally 

A. The decomposition of the mineral, the isolation and 

conversion into oxalates of the rare earth class. 

B. The conversion of the rare earth oxalates into soluble 


C. The separation of the rare earth group into three sub- 

groups by means of double salt formation with potas- 
sium sulphate. 

D. The isolation of the various elements from the cerium, 

terbium, and ytterbium sub-groups. 

Section D we cannot enter into in detail, as each element 
requires special methods for its treatment. We shall content 
ourselves with indicating in outline the principles underlying 
the methods actually employed. 

A. In nature, the rare earths occur among silicates, which 
are usually easily broken up by evaporating the ore to dryness 
with concentrated hydrochloric or sulphuric acid. The powdered 
residue is then added to water, and any insoluble portion is 
filtered off. The solution now contains the rare earths in the 

1 See especially Little's account of the rare earths in vol. iv. of Friend's 
Textbook of Inorganic Chemistry (1917). 


form of chlorides or sulphates, and may also contain salts of 
copper, bismuth, molybdenum, and iron. Thorium also is often 
present. The solution is treated with sulphuretted hydrogen 
to precipitate the metals of the second group. This leaves us 
with the rare earth elements, iron, and thorium in the solution. 
After oxidation of the ferrous iron to ferric with chlorine, we 
add a little hydrochloric acid, and then precipitate the rare earth 
elements and thorium in the form of oxalates. 

Other reagents are sometimes utilized to break up the 
mineral. Fusion with potassium hydrogen sulphate or with 
sodium hydroxide, heating with hydrofluoric acid or with 
sulphur dichloride vapour are processes which have advant ges 
in special cases. 

B. Since the oxalates are insoluble substances, we must 
convert them into some other salts which we can get into solu- 
tion. This is usually done by one of three methods. The 
oxalates may be heated to form oxides, from which any required 
salt can be produced ; or we may simply dissolve the oxalates 
in hot nitric acid, and thus obtain the easily soluble nitrates ; 
or, finally, we may boil the- oxalates for a considerable time with 
caustic potash solution, which gives us the hydroxides, and from 
these we can produce any salt we may need. At this point we 
have to get rid of the thorium which still remains in the solu- 
tion. This is usually done by adding hydrogen peroxide to the 
nitrate solution ; after warming, a precipitate of thorium per- 
oxide separates out and is filtered off. 

C. We have now to separate the rare earth salts in three 
main groups by means of double salt formation with potassium 
sulphate. It has been found that in this way we get the 
following classification : 

1. Elements giving practically insoluble double salts : 

Scandium, cerium, lanthanum, praseodymium, neo- 
dymium, and samarium. 

2. Elements giving soluble double salts : Europium, gado- 

linium, and terbium. 

3. Elements giving easily soluble double salts : Dysprosium, 

holmium, erbium, thulium, yttrium, ytterbium. 
Instead of potassium sulphate, sodium sulphate is sometimes used. 

D. When we come to the actual isolation of one element of 
the rare earth series from its companions, there are two chief 



methods which we may employ. The first of these depends 
upon the basicities of the different metals ; the second upon 
the differences in the solubility of their salts. The former 
method is capable of subdivision ; for we may depend either 
upon a fractional precipitation of the salts by means of bases 
of different strengths, or we may rely upon a differential 
decomposition of the nitrates by means of heat. We must now 
give an outline of these methods 

la. The principle underlying the method of fractional pre- 
cipitation is almost self-evident. If we take two substances X 
and Y in solution, and add to them a third substance Z which 
is capable of precipitating both X and Y, then if we add excess 
of Z, both X and Y will be completely deposited from the 
solution. If, on the other hand, we add a quantity of Z which 
is insufficient to precipitate the total quantity of X and Y 
present, it is clear that the amounts of X and Y precipitated 
will depend to a great extent upon their chemical behaviour 
with respect to Z. Suppose that Z is a base which is very 
much stronger than A, but only very little stronger than B. 
If we have one molecule of each body present, it is clear that 
Z will displace A from its salts rather than B. Consequently 
we shall have A precipitated and B left in solution. In the 
case of the rare earths, this method is employed, using such 
reagents as ammonia, magnesia, caustic potash, caustic soda, or 
organic bases. In the first precipitation the substance is 
separated into two parts : precipitate and mother liquor. The 
precipitate is then redissolved, and the process is repeated. 
The crystals resulting from the second fractional precipitation 
are once more dissolved and fractionally precipitated, and the 
process is continued until a pure substance is obtained. 

15. The second method, depending upon the difference in 
basicity of the various members of the rare earth group, is 
carried out in the following manner. In the first place, the 
mixed oxalates are converted into nitrates in the usual way, 
and to this mixture is added some alkali nitrate in order to 
lower the melting-point of the whole mass. Thereafter the 
mixture is fused. It is then found that the nitrates decom- 
pose, and usually the most negative oxide separates out first. 
This is removed from the mass by solution, and the process 
repeated again and again as in the last case. 


2. The method depending upon the solubilities of the salts 
of the rare earths is simply a process of fractional crystalliza- 
tion. The mixture of salts is dissolved in water, and the 
solution is then evaporated until about half the solid separates 
out. This is filtered off and dissolved in water, and the same 
process is repeated. At the same time the mother liquor is 
again evaporated till half its solute separates out, and the pro- 
cess is repeated a sufficient number of times. Where there 
are more than two elements present, it is usual to mix the 
fractions in the following manner : After the first crystalliza- 
tion, we have crystals (A) and mother liquor (B). In the 
second set of crystallizations, each of these gives rise to a set 
of crystals and a mother liquor. The crystals derived from 
the second crystallization of the mother liquor (B) are then 
mixed with the mother liquor derived from the recrystalliza- 
tion of (A), and in this way a third set of fractions is produced 
which grow more and more rich in the salt whose solubilities 
lie between the two extremes. The scheme below will make 
the point clear 

Original solution 

More soluble Less soluble 

(mother liquor) (crystals) 

More soluble Less soluble More soluble Less soluble 

More soluble Less soluble More soluble Less soluble More soluble Less soluble 

__ ___ _ __ __ __ 

! II [.I'll ! I II I 

(i.) (ii.) (in.) 

The brackets indicate that two fractions have been mixed 
together before recrystallization. 

A glance at the scheme will show that we are accumulating 
the more soluble salts at (I.), the salts of intermediate solu- 
bility at (II.), and the least soluble salts at (III.). 

The salts which have been found most useful in this 
method of separation are the chromate, sulphate, nitrate, 
oxalate, and formate. The metallic derivatives of acetyl- 
acetone have also been utilized. James x has discovered 

1 James, Chem. News, 1907, 95, 181 ; 1908, 97, 61, 205 ; J. Amer. Chem. 
Soc., 1907, 29, 495. See also ibid., 1908, 30, 979, for his general scheme of 


two methods by means of which rapid separations of the 
yttrium earths can be carried out. In the first of these, the 
rare earth oxalates are dissolved by warming them with a 
saturated solution of ammonium carbonate in dilute ammonia ; 
and the fractionation is carried out by simply boiling the 
solution until a precipitate is thrown down. If five fractions 
of approximately equal weight be obtained in this manner, 
the first contains yttrium, the second is mainly holmium and 
dysprosium, the third is a mixture of various elements, while 
the last two contain erbium. James' second method depends 
upon the fractionation of the bromates of the rare earths ; he 
obtains these from the corresponding oxalates by treating the 
latter with sulphuric acid, and then with barium bromate. 

Now, let us suppose that we have completed a series of 
fractional precipitations or crystallizations. The question at 
once arises : Have we carried the process to its end and 
secured the pure product we set out to obtain ? In order to 
answer this question, we must know whether or not we are 
altering the composition of our substance in the course of the 
fractionation. Thus quite early in the series of fractionations 
it becomes advisable to find out what change each successive 
operation produces in the substance under treatment. 

There are three methods by means of which we may settle 
the question. If we are gradually purifying a crude substance, 
it is obvious that if, after each operation, we determine the 
equivalent of the metal (or the average equivalent of the 
mixture in the first case) we shall get a gradual approximation 
to the equivalent of the pure metal, as the fractions contain 
less and less impurity the further the process is continued. 
For example, if we start with a mixture of cerium and 
lanthanum, the equivalent of the mixture (if the two metals 
be present in equal proportions) will lie half-way between 
those of cerium and lanthanum. As our fractions become 
richer and richer in cerium, the equivalent of the fraction will 
draw closer and closer to the cerium equivalent ; and when 
the fraction contains nothing but cerium salt, we shall, of 
course, find the equivalent of cerium. This method is very 
rough at the best; and when applied to the case of the rare 

separation. For a full account of the subject, see Little's chapters in 
Friend's Textbook of Inorganic Chemistry, vol. iv., especially p. 322 ff . 


earths it is unsatisfactory, owing to comparatively small 
differences between the equivalents of the various metals in 
the group. 

The second method, which is much more refined, consists 
in applying the spectroscope to the problem. Here we have 
a considerable choice of procedure, for either absorption or 
emission spectra may be used ; and in the latter class there are 
three different types : spark spectra, arc spectra, and phos- 
phorescence spectra. The spectrum of each set of fractions is 
examined and purity is counted as attained 'when repeated 
fractionation fails to alter the spectrum. 

The third method depends upon the fact that the rare 
earth elements differ considerably from one another in their 
magnetic susceptibilities; so that the determination of this 
property suffices as a guide to the purity of the various 
fractions. 1 

5. Industrial Applications of the Rare Earth Elements. 

The main employment of rare earth elements is to be found 
in the incandescent gas-mantle industry, which utilizes cerium 
and " didymiurn " * salts to a certain extent, as will be seen in 
the next chapter. The quantities required by the trade, how- 
ever, are comparatively small; and owing to the fact that 
large quantities of cerium are isolated as by-products during 
the extraction of thorium, in the mantle industry, it has been 
worth while to look for other uses for cerium. 

Cerium, in the form of alloys, has been utilized in various 
ways. An alloy of cerium with lanthanum, praseodymium 
and neodymium is known technically as mischmetall ; and it 
is utilized as a reducing agent. The well-known cerium tin- 
ders, which are used in gas-lighters or cigarette-lighters, are 
alloys of cerium with 30 per cent, of iron, nickel, cobalt, or 
other metals ; and the use of these is extending rapidly. 

Cerium is also used for impregnating carbon electrodes for 

1 Urbain, Compt. rend., 1908, 146, 406, 922 ; 1909, 149, 37 ; 1910, 150, 913 ; 
1911, 152, 141 ; Urbain and Jantsch, ibid., 1908, 147, 1286 ; Blumenfeld and 
Urbain, ibid., 1914, 159, 323. 

* The so-called ." didymium " is a mixture of neodymium and praseody- 


arc lamps ; and it is claimed that this treatment yields a more 
brilliant and steady arc than ordinary carbon. 

Cerium glass has the power of cutting off about 30 per 
cent, of the heat rays and all the ultra-violet portion of the 
spectrum ; so that for certain purposes it is very useful.- 

Other uses for cerium have been suggested in photography, 
dyeing, medicine, and catalytic processes ; but the field appears 
to be limited. 



1. Introductory. 

THE fact that the introduction of solids into a non-luminous 
flame may cause the emission of light is common knowledge 
to any one who has ever heated a platinum wire in the jet of 
a Bunsen burner; but between this experiment and the pro- 
duction of an efficient system of illumination on a commercial 
scale lie a long series of steps ; and it is only within the pre- 
sent generation that the process of perfecting the incandescent 
gas mantle has been completed. 

In 1829, Berzelius observed that oxides like thoria or 
zirconia emitted a brilliant light when placed in non-luminous 
flames ; but the first practical application of the idea appears 
to have been made when lime-light came into use. Lime 
heated in an oxy -hydrogen flame gives out a brilliant light; 
but for ordinary illuminating purposes it was useless, owing 
to the fact that ordinary gas-flames are too low in temperature 
to produce a satisfactory radiation from lime. 

The incandescent mantle industry, 1 in its modern form, 
dates from 1884, when Auer (later Auer von Welsbach) took 
out a patent protecting the use for illuminating purposes of 
fibrous materials impregnated with oxides of certain rare 
elements. By 1891 the incandescent mantle had become a 
practical possibility; but at this point the fortunes of the 
new industry hung in the balance. Though the supply of 
rare earths was then amply sufficient for the purposes of 
scientific investigation, it seemed doubtful whether enough 

1 Fuller details of the industry are to be found in Martin's Industrial 
Inorganic Chemistry, and similar works. 



material could be obtained to cope with the requirements of 
a commercial process. At the time when the new mode of 
illumination was placed on the market, the minerals upon 
which it depended cost over a sovereign an ounce, and, what 
was even more grave, they could be obtained only in small 
amounts, twenty or thirty pounds being a very fair find at 
one time in the Norwegian deposits. 

The demand for incandescent mantles soon made it clear 
that fresh sources of supply must be found, if the process was 
to keep the position which it had gained; and a search for 
other deposits was begun. It was recalled that thorium had 
been detected in certain abandoned gold claims in Carolina; 
and soon the monazite sand of that district was found to be 
a new mine of rare earths. Indeed, so great were the quan- 
tities discovered there and in Brazil that the epithet " rare " 
ceased to bear its primitive meaning. Ceylon has also yielded 
a material called thorianite, which is rich in thorium oxide. 
Within ten years, owing to the discovery of these new sources, 
the price of the rare earths dropped to 2 or 3 per cent, of its 
original amount ; and the permanency of the incandescent gas- 
mantle industry was assured, so far as this factor was con- 

2. The Treatment 'of Monazite Sand. 

Natural monazite sands, of course, vary in composition 
according to their place of origin; but for the purpose of 
illustration, the Brazilian type will serve our purpose. The 
percentage of the mineral monazite in the sands may vary 
between 2 and 60 per cent.; and the remainder of the sand 
is valueless from the commercial point of view. Even mon- 
azite itself contains a large percentage of worthless material ; 
for its thorium content is only 5 per cent, of the whole mass. 
Now in order to make the material economically valuable for 
export, it is necessary to concentrate the thorium up to at 
least 4 per cent, strength : in other words, a product must be 
prepared which contains about 90 per cent, of pure monazite. 

The raw sand contains, in addition to monazite, quantities 
of quartz mixed with lesser amounts of other minerals ; and 
the exact nature of the purification process depends largely 


upon the kind of impurities which are present. Three main 
methods of segregation are employed: (1) the wet process; 
(2) dry blowing ; and (3) electro-magnetic separation. 

In the wet process, the sand is mixed with water and the 
liquid is run over a vibrating table, by which means the sand 
is sifted out into groups according to the specific gravity of 
its particles. Since quartz has a density of 2*65 whilst mon- 
azite's density is 4'8 to 5 '5, it will be seen that the separation 
of the two is not difficult. The process of dry blowing de- 
pends upon the same principle, but an air-blast is substituted 
for the stream of water. In the electro-magnetic method, 
advantage is taken of the fact that the ingredients of mon- 
azite sand are differently affected by a magnet. The sand is 
fed on to the top of a revolving belt and thrown from this 
against another travelling belt behind which a strong electro- 
magnet is placed. The magnetic attraction retards the velocity 
of certain of the sand particles while leaving others to travel 
forward unaffected ; and by means of a series of slots leading 
to bins, the various constituents of the sand are roughly 
separated from each other. 

The next stage in the concentration takes the form of 
" breaking " which consists in feeding the monazite, obtained 
as above, into cast-iron pans containing twice its weight of 
hot sulphuric acid (specific gravity 1'84). Heating completes 
the process and converts the monazite into a pasty mass. This 
mass is run into cold water and any insoluble materials such 
as silica are allowed to settle to the bottom. The solution, 
which contains the rare earth phosphates dissolved in dilute 
sulphuric acid, is siphoned off from the insoluble matter. In 
this way a material is obtained which contains thoria and the 
rare earths in the proportions of 1 : 12. 

Further concentration of the thoria is effected by taking 
advantage of the fact that thorium is more basic than its com- 
panions and separates before them when the solution is largely 
diluted or is neutralized with ammonia or magnesite. The 
thorium precipitated in this way is filtered off, dissolved in 
the minimum quantity of acid, and reprecipitated by a repeti- 
tion of the process. 

This leaves us with a concentrated preparation of thorium 
phosphate and the phosphates of some rare earth elements, from 


which phosphoric acid is removed by methods which are trade 

Thereafter, the purification of the individual constituents 
of the mixture is carried out by methods which were outlined 
in the preceding chapter. 

3. Fabrics and their Treatment. 

Three fabrics have been utilized in the manufacture of 
mantles : cotton, ramie fibre, and artificial silk. Of these, 
cotton is the least satisfactory and artificial silk the best, 
though its cost hinders its general employment. Kamie fibre 
is prepared from China grass ; whilst artificial silk may be 
obtained either from viscose or from cuprammonium cellulose. 

The first step in the preparation of the mantle is the test- 
ing of the yarn from which the mantle is to be made ; for the 
knitting is done by machinery, and the tension to which the 
yarn is subjected appears to have some effect upon the final 
product. The knitting machine is designed to produce a con- 
tinuous cylindrical fabric which can be cut up into lengths as 

Artificial silk contains no mineral matter ; but both cotton 
and ramie fibre are contaminated with inorganic and fatty im- 
purities which must be removed. Cotton mantles are there- 
fore treated with dilute caustic soda solution and thereafter 
are washed in dilute acid. Kamie fibre fabric is soaked for 
some hours in dilute acid; freed from excess acid by centri- 
fuges ; washed in pure water and dipped in dilute ammonia. 
A final washing in water completes the purification. By 
these methods, the inorganic content of the fabric is reduced 
to about 0*02 per cent. An ash content lower than this is 
found to give poor results in the finished product. The moist 
fabric is dried in special ovens through which a hot-air draught 

The long cylinder of fabric is then placed in a special 
machine which cuts it into pieces of fixed length, ready for 
the next stage in the process. 

4. Impregnation, Fixing, and Branding. 

The " shaping " of the mantle must now be described. In 
the case of the ordinary upright mantle, this is done before 


impregnation ; whilst the inverted mantle is shaped after that 
process is complete. The upright mantle is reinforced at the 
head by a piece of cotton or ramie tulle which is stitched on ; 
and the head so formed is drawn together into the shape seen 
in the ordinary mantle. An asbestos hook is also attached. 
In the case of inverted mantles, the stitching is done with 
impregnated thread. 

The process of impregnation is the main factor in produc- 
ing a good mantle. It is found that mantles soaked in a solu- 
tion of a pure thorium salt do not give results as good as 
those dipped in solutions containing cerium salts in addition 
to thorium. For example, if we take the intensity of the 
light given by a pure thorium mantle as 20, the intensity 
shown by a mantle containing 99 per cent, thorium and 1 per 
cent, cerium is about 140. Further addition of cerium above 
1 per cent, leads to a decrease in the illuminating power of 
the mantle. But this is not the only factor with which we 
are concerned. Since the mantle will be subjected to shocks, 
it is necessary to provide it with a skeleton capable of resist- 
ing vibration ; so about 1 per cent, of other materials such as 
aluminium, calcium, zirconium, beryllium or magnesium is 
added to the impregnating solution. 

The time of impregnation varies according to the material 
of the -man tie; a few minutes suffices in the case of cotton or 
ramie, but artificial silk mantles require longer immersion. 
After the soaking is completed, the mantles are run between 
rollers which remove exactly the proper amount of moisture : 
or, in the case of artificial silk mantles, the excess solution is 
removed in a centrifuge. 

The next stage in the process is termed "fixing." Con- 
sideration will show that the lower end of the upright mantle 
and the upper end of the inverted mantle are the weakest 
points in the structures ; for it is at these places that the 
greatest strain will come in practice. The lower end of the 
upright mantle is exposed to draughts and hence is liable to 
fray ; whilst the upper end of the inverted mantle carries the 
weight of the whole device when in position. These two 
points are therefore reinforced by an additional treatment 
with a solution containing such materials as aluminium 
nitrate, borax, or calcium nitrate; and in this way extra 


mineral matter is deposited at the weak spots, thus forming a 
stronger skeleton. 

Branding the maker's name upon the mantle is accom- 
plished by taking advantage of the fact that " didymium " 
salts decrease the luminosity of the thorium mixture. A 
solution of " didymium " nitrate is employed for the inscription 
which becomes visible as dark lettering on the bright ground 
of the mantle when in use. 

5. Burning, Collodionizing, and Testing. 

The mantles are now ready for the last steps in the pro- 
cess. They are shaped on wooden models and are then sub- 
jected to the action of Bunsen burners which "burn off" all 
the original organic material from the mantle. The heating 
is begun at the closed ends of the mantles, as otherwise the 
contraction of the fibre under the heat would distort the 
shape of the fabric. This first stage in heating leaves the 
mantle a mere fragile structure ; and in order to toughen it, 
high-pressure burners are brought in play which consolidate 
the more fusible oxides in the mantle and render the whole 
mass more resistant. 

It is next necessary to provide some support for the ash 
skeleton so that it may stand the shocks of transport. This 
is provided by the collodionizing process. In groups of forty 
to sixty, the mantles are dipped into a mixture of nitro- 
cellulose and various oils. The oils are added to reduce the 
rapidity of combustion of the nitro-cellulose. The dipping is 
done by machinery ; and the mantles are then dried in ovens 
heated by high-pressure steam. When dry, they are trimmed 
in a machine and are then ready for packing. 

Out of each batch, it is usual to test 1 or 2 per cent, of 
the mantles in order to see that they are up to standard. 
The test is applied by lighting them over a burner in the 
usual manner and then applying to the stand a series of 
shocks by means of a machine-driven hammer. The usual 
test for an upright mantle is 1000 shocks at the rate of 300 
per minute from a 2 oz. hammer ; whilst an inverted mantle 
is expected to remain in good condition after being submitted 
to 3000 shocks from a 3 oz. hammer at the rate of 600 shocks 
per minute. 



1. The Apparatus. 

IN the first section of Chapter VII., it was pointed out that when 
the cathode of a Crookes' tube is pierced with one or more 
apertures, a stream of " canal " or " positive " rays travels from 
the cathode in a direction opposite to that taken by the cathode 
rays. These positive rays are streams of the residual molecules 
of gas in the tube which have acquired a positive charge. 
Like the particles which make up the cathode stream, they 
can be deflected from their normal straight paths by the action 
of either magnetic or electric fields ; and the deviation in each 
case is proportional to the ratio between the mass of the 
particle and the charge which it carries. Sir J. J. Thomson 
has utilized this property in devising a method for the 
recognition of the nature of these residual molecules ; and his 
method has come to be known as positive ray analysis. 1 

In order to employ the positive rays as a means of quali- 
tative gas analysis, three things are required : (1) an apparatus 
for generating positive rays ; (2) means of applying an electric 
and a magnetic field to the rays; and (3) a contrivance for 
registering the deviations thus produced. The essentials of 
the apparatus are shown in the diagram below (Fig. 16). 
A represents part of the glass flask which acts as the Crookes' 
tube. It has a capacity of from 1 to 2 litres. The cathode 
C, of special construction, is placed in the neck, D, of the flask 
as shown ; and protrudes into an ebonite box, UV. The joints 

1 For a full account of the subject, see Thomson, Bays of Positive Elec- 
tricity and their Application to Chemical Analysis (1913) ; a very clear de- 
scription is given in Crowther's Molecular Physics; and a summary is to 
be found in Thomson's Bakerian Lecture, Proc. Roy. Soc., 1914, A, 89, 1. 




are made gas-tight with sealing-wax ; and this is protected 
from heating effects by the water-jacket J. L and M are 
pieces of soft iron which are used as poles for the electric field. 
They have plain faces which are about 3 cm. long and they 
stand about 1*5 mm. apart from each other. They are con- 
nected with the terminals of a battery of storage cells and in 
this way any required difference in potential can be maintained 
between them. P and Q are the poles of an electro-magnet. 
F is a conical glass vessel about 40 cm. long which is fastened 
to a camera containing the photographic plate on which the 
results are recorded. WV are iron plates which are interposed 
in order to screen the positive ray generator from the influence 
of the electro-magnet. 

(From Sir J. J. Thomson's " Rays of Positive Electricity.") 
FIG. 16. 

The shape of the cathode, C, can be seen from the figure. 
The unshaded portion is made of aluminium ; and the rest is 
composed of soft iron. Through the aperture down the centre 
passes a tube of copper with a bore of about 01 mm. and a 
length of about *7 cm. 

One point of great importance must be mentioned. When 
a positively charged molecule enters the conical vessel F, it 
may ionize the residual gas contained therein. The result of 
this will be that its positively charged fellows which follow it 
may find awaiting them negatively charged gas molecules ; 
and discharges may occur between the two sets. As a con- 
sequence, the incoming particles lose their charges and cease 
to be acted on by the electric and magnetic fields ; and com- 
plications are introduced. To avoid this, the pressure in the 


vessel F is kept low by absorbing the gas by means of char- 
coal and liquid air; and since the leak from the generating 
flask is small on account of the narrowness of the connecting 
tube in the cathode it is possible to maintain a difference in 
pressure between the generator A and the receiver P. 

2. The Theory of the Method. 

Let us imagine that a residual molecule of gas becomes 
charged in the vicinity of the anode of the Crookes' tube which 
forms part of the positive ray apparatus. Such a particle will 
be driven towards the cathode by the electrical forces in the 
Crookes' tube; and in this process it will acquire an energy 
which may be represented by D X e, where D is the difference 
of potential between the electrodes and e is the charge on the 
particle. The velocity of the particle when it reaches the 
cathode will therefore be v in the equation 

where m is the mass of the particle in question. 

This particular value of v represents a maximum velocity . 
for if the particle becomes charged at a point nearer the cathode, 
the driving force will have a shorter time to work upon it and 
hence its velocity will be lower than if it started in close 
proximity to the anode. 

Now if neither the electro-magnet nor the static electrical 
field be in operation, the particle will pass through the cathode 
tube unaffected and will strike a point on the photographic 
plate directly opposite to the cathode aperture. Let us repre- 
sent this point by in the figure below. 

Let us next assume that we apply the magnetic field ; and 
that this tends to deviate the next particle from its straight 
path in the direction of M in the figure. Instead of striking 
the plate at 0, the new particle will fall at, say, the point P. 

Further, let us imagine that the magnetic field is switched 
off and that the electric field is brought into operation instead. 
The deviation in this case will be at right angles to that pro- 
duced by the magnetic field ; and a particle under its influence 
will be deflected from in the direction of E. Suppose that 
it strikes the plate at Q. 


Obviously, if both fields are in operation at the same time, 
the particle will be deflected towards M and towards E simul- 
taneously; and it will strike the plate at the point E. We 
have now to determine the position of E in terms of the forces 
in operation. 

Without going into details l it may be said that if M be the 
magnetic field, the particle will be deflected through a distance 

M e 

ki . , where k\ is a constant depending on the apparatus ; 

and if E be the electric field, the deviation due to it will be 
measured by 7c 2 . '-%. Or, in other words, calling the electric 

deflection x and the magnetic one y, we have 

M e 

Magnetic deflection = OP = QE = 


= y 

mv 2 

Electric deflection = OQ = Jc 2 . ^# = x 
From this it follows that y/x 

and y^lx 


Now if the magnetic and electric fields are not varied, it is 
clear that all the quantities within the brackets in the last two 
expressions are constant; and hence 
y/x forms a measure of the velocity of 
the deflected particles whilst y*jx gives 
the ratio of the mass to the charge on a 

So far, we have considered the case 
of a single particle ; and our reasoning 
holds good for all identical particles, 
so that all such particles would strike 
the plate at the point E. In actual 
practice, however, as we have seen, the 
positive ray is made up of a series of 
particles which may be travelling at 

different velocities. The expression yjx will have a different 
value for each particle ; but if the particles all have the same 

FIG. 17. 

See Thomson, Rays of Positive Electricity, p. 7 (1913). 


FIG. 18. FIG. 19. 

{From Sir J. J. Thomson's " Rays of Positive Electricity 


ratio of charge to mass, then the expression y^jx will be con- 
stant for the whole series. In other words, these particles 
will strike the plate at a series of points which will lie ou 
some curve which can be expressed by y^fx = constant. This 
curve is obviously a parabola. 

Let us now put this into a more concrete form. Imagine 
that one particle is a helium atom and that another is an atom 
of argon. Let us suppose that each carries a unit charge. 
Then the helium atom will have a ratio of mass to charge equal 
to 4 : 1, whilst the argon atom will have the ratio 40 : 1. Thus 
the helium atom will strike the plate somewhere on a curve 
represented by y^fx = 4Jc ; and the argon atom will strike the 
plate somewhere on a curve for which y^jx = 40&. Thus the 
two atoms will be sifted away from each other on the plate. 
Further, if we imagine a stream of helium atoms moving at 
different velocities in the ray, they will distribute themselves 
along the parabola y^jx = 4 k ; and a similar stream of argon 
atoms will distribute themselves over the parabola y*/x = 40&. 

As will be seen in the next section, the dimensions of the 
parabolas enable us to calculate the ratio of the charge to the 
mass for those molecules which leave traces upon the photo- 
graphic plate. 

3. The Interpretation of the Photographs. 

The illustrations (Figs. 18 and 19) give some idea of the 
results obtained by the use of Thomson's apparatus.* 

In the first place, it will be noticed that both photographs 
contain a bright spot at the apex of the parabola. This spot 
is the point of impact of the undeviated rays which pass when 
the electric and magnetic fields are not in operation. Secondly, 
it will be seen that two branches of the parabolas are shown, 
one above and the other below the apex. These two branches 
are produced by taking two separate exposures in both of 
which the electric field is kept constant : in the one exposure 

* In the latest form of the instrument a sheet of metal is placed over the 
photographic film and this sheet has a fine parabolic slit cut in it. By vary- 
ing the magnetic field and keeping the electric field constant, the parabolas 
of the various constituents of a mixture under analysis can be successively 
brought into coincidence with the slit. This avoids the " splashed " appear- 
ance shown in the photographs given in Figs. 18 and 19. 





the magnetic field operates in one direction ; whilst in the 
second exposure it is reversed : so that the particles are diverted 
in opposite directions (up the plate in one case and down it 
in the second instance). The parabolic form of the curves is 
easily recognizable in the illustrations. It must be strongly 
emphasized that these tracks are not formed by the particles 
grazing along the plate; but that they represent a series of 
"bullet-marks" made by particles striking the plate from 
above, the separate " marks " being so close together that they 
merge into lines. 

It will be noticed that the parabolas stop short and are not 
continued to the apex. The reason for this is obvious. In 
order that any particle should strike 
M the plate in close proximity to the apex, 
the amount of deviation undergone by 
it would need to be small; and hence 
its velocity would need to be very great. 
But there is a limit to the velocity of 
the particles owing to the limited differ- 
o ence in potential between the electrodes 
of the Crookes' tube ; and thus it comes 
about that we do not find particles 
travelling with a velocity higher than 
a certain maximum. This maximum 
M ' velocity is sufficient to bring them within 
FIG. 20, a certain distance of the apex, but it is 

not sufficient to over-step this point ; so 

that no particles strike the plate in the neighbourhood of the 

The method of measuring the plates can be seen from the 
diagram above. 

Draw any line AF parallel to OM and cutting to two para- 
bolas at A, B, D, and F and meeting the line OE at C. The 
value of yi (for the outermost parabola) is equal to CA ; the 
value 2/2 (for the innermost parabola) is equal to OB. In both 
cases the value of ~x is the same, being equal to 00. Now 
from the equations in the last section the value of the fraction 

c ar ^ e for any particle is given by y^jx. Therefore, if mi and 

ei be the mass and charge for particles on the outermost 


parabola and ra 2 and e z be the corresponding factors for particles 
on the second parabola, we have 


and if e 1 = e% then 

AC 2 
BC 2 

m l 


Thus if all the particles carry the same charge, the ratio of 
the masses of the two sets of particles is given by the ratio 
AC 2 : BC 2 , or from the more easily measurable ratio AF 2 : BD 2 , 
since AF=2AC and BD = 2BC. In actual practice it is found 
that this condition is fulfilled in most cases ; and when any 
deviation occurs owing to a particle carrying more than one 
charge the matter is not serious, for the number of charges 
carried by any particle is always small. 

The following table * gives an idea of the manner in which 
the results obtained from the photographs are interpreted. 
The first column gives the distance d, of the different para- 
bolas from the axis OE, measured along a common ordinate. 
The second column gives the ratio of mass to charge, assuming 
that all the particles carry the same charge. The third column 
gives an interpretation of this when modified by the assump- 
tion of varying charges. 

d mm. 


Probable Cause of Line. 




Mercury atom with single charge 



Hg+ + 

Mercury atom with two charges 



Hg++ + 

Mercury atom with three charges 



G0 2 + 

Molecule of carbon dioxide with single charge 



A + 

Argon atom with single charge 



N 2 + 

Nitrogen molecule with single charge 



Ne + 

Neon atom with single charge 




Oxygen atom with single charge 




Nitrogen atom with single charge 




Carbon atom with single charge 

38 '7 


N+ + 

Nitrogen atom with two charges 

For instance, the heaviest atom likely to exist in the gas 
under examination is a mercury one with an atomic weight of 
200. Since the ratio of the mass to charge is 200 : 1, it is clear 
that this atom carries a single charge. Again, another para- 
bola corresponds to the ratio of mass to charge 100 : 1. This 

1 Crowther, Molecular Physics, p. 52 (1914). 


would represent an element of atomic weight of 100 with a 
single charge on the atom or an element of atomic weight of 
200 with two charges on the atom. We have therefore to 
choose between the probability that the line is produced by a 
mercury atom carrying two charges or by an atom of, say, 
ruthenium, with one charge. Obviously the mercury hypo- 
thesis is the more probable one. Then again, in the case of 
the third line, the atomic weight of the element must be 
approximately 67 or 134 or 201. Neither zinc nor caesium is 
likely to be present under the experimental conditions, so we 
assume that the "67" line is due to an atom of mercury 
carrying three charges. 

The only point about which there might be a doubt in the 
above interpretation is in the case marked with an asterisk ; 
for the value 20 might be obtained either from a neon atom 
with one charge or from an argon atom with two charges. 
The chance of its being due to a mercury atom with ten 
charges is ruled out by the fact that it is rare that mercury 
carries even as many as eight charges, to judge from the 
general run of results. 

From the above it will be clear that in the stream of 
positive rays we may have : (1) positively electrified atoms 
with one charge ; (2) positively electrified molecules with one 
charge; and (3) positively electrified atoms with multiple 
charges. In addition to these we may also find tracks due to 
negatively electrified atoms and molecules which have become 
neutralized on passing through the cathode and have picked 
up a cathode particle en route. We need not, however, enter 
into this question here. 1 From the chemical point of view it 
is interesting to note that among the carbon compounds the 
negatively charged molecule is not detected in the case of 
methane, carbon dioxide, phosgene, etc., wherein carbon is 
linked only to other elements; but negatively charged mole- 
cules are found when the compound examined contains two 
directly linked carbon atoms as in the cases of acetylene, 
ethylene, or ethane. In the examination of benzene vapour, 
negatively electrified triplets of carbon atoms were observed, 
and there seems to be a possibility that quartets may have 
been present. 

i Thomson, Proc, Boy Soc., 1914, A, 89, 1. 


4. The Substance X 3 . 

When solids are bombarded with cathode rays, various 
gases seem to be given off; and the examination of these by 
the positive ray method yielded facts of some interest. The 
results point to the presence in the gas mixture of neon, helium, 
and a third substance to which the name X 3 has been given. 

With regard to the sources of X 3 , it has been found that 
it is produced by the action of cathode rays upon a very varied 
series of substances. Platinum, palladium, aluminium, copper, 
zinc, iron, nickel, silver, gold, lead, graphite, diamond dust> 
lithium chloride, and other metallic salts as well as some 
meteorites have been. found to liberate the gas. The presence 
of mercury vapour in the bombardment tube diminishes the 
intensity of the line due to X 3 ; from which Thomson deduces 
that X 3 combines with mercury vapour under the influence of 
the electric discharge. 

If we assume that the maximum number of charges which 
can be carried by a particle is limited, there appear to be only 
two possible explanations for the X 3 line. It must be produced 
by something in which the ratio of mass to charge is three 
times that found in the case of a hydrogen atom. This can 
be accounted for by assuming that X 3 is either (1) a carbon 
atom carrying four electrical charges; or (2) a molecule con- 
taining three hydrogen atoms and carrying a unit charge. 
Let us examine these possibilities in turn. 

In the first place, the gas giving the "3" line can be 
stored in the bombardment vessel for days and yet it will pro- 
duce the characteristic line when analysed after this period. 
If the line be due to carbon atoms carrying four charges, it 
must be because some volatile carbon compound is produced 
during the bombardment which, when introduced into the dis- 
charge tube gives, on the passage of the discharge through it, 
carbon atoms with four charges. Experiments have been made 
by introducing directly into the discharge tube such substances 
as methane, carbon dioxide, carbon monoxide, ethylene, ethane, 
acetylene, carbonyl chloride, and carbon tetrachloride ; but 
none of these gives results * analogous to X 3 . Other evidence 

* C/., however, Collie and H. S. Patterson's observation that carbon 
compounds are retained under conditions which might normally be expected 
to preclude their appearance (Proc., 1913, 29, 217). 


is furnished by the nature of the tracks on the plate, but into 
this we need not enter here. 1 

With regard to the possibility that X 3 is a singly-charged 
molecule containing three hydrogen atoms, we have the follow- 
ing evidence. Whenever large amounts of X 3 are produced, 
spectroscopic examination detects the presence of a consider- 
able quantity of hydrogen in the gas liberated by the cathode 
bombardment. When various salts were bombarded after 
recrystallization, it was found that those like potassium iodide, 
lithium carbonate, and potassium chloride gave smaller yields 
of X 3 after recrystallization ; whereas in the case of potassium 
hydroxide, lithium hydroxide, lithium chloride, and calcium 
chloride recrystallization had no effect upon the output of X 3 . 
It will be noticed that the first class have no hydrogen in their 
composition; whereas the second group either contain hydro- 
gen or are deliquescent and thus absorb hydrogen in water- 
molecules. This is regarded by Thomson as a proof of the 
connection between hydrogen and the X 3 line. 

Turning now to the properties of X 3 as far as they are 
known, the following seem important : X 3 when mixed with 
hydrogen, is not markedly photo- sensiti ve ; but if a mixture 
of X 3 and oxygen be exposed to a magnesium flash, the X 3 is 
destroyed. The passage of an electric spark has an analogous 
effect : hydrogen and X 3 will stand a good deal of sparking 
without any noticeable result ; but if a spark be passed through 
a mixture of X 3 , oxygen and hydrogen, the X 3 disappears. In 
the absence of oxygen, X 3 can be heated to a high temperature 
without destruction; but it disappears when raised to a red- 
heat in presence of copper oxide. Like hydrogen, X 3 has the 
power of passing through a palladium wall; but its rate of 
transfusion is much lower than that of hydrogen. 

On these grounds, Thomson assumes that X 3 is really a 
triatomic molecule of hydrogen, H 3 ; and he considers it to be 
the hydrogen analogue of ozone. It is evidently more stable 
than ozone, as is seen from its resistance to high temperatures. 
No particular spectrum has been observed for X 3 ; for a mix- 
ture of it and hydrogen exhibits only the normal hydrogen 

Against this may be urged the evidence brought to light 

1 See Thomson, Bays of Positive Electricity, p. 117 (1913). 


by Collie and H. S. Patterson in the course of work in a 
different field. 1 They found that when a heavy discharge is 
passed through a vacuum tube, quite considerable quantities 
of hydrogen can be made to disappear. For example, in one 
experiment, as much as 3*6 c.c. of hydrogen apparently vanished. 
A gas is produced * which gives a carbon spectrum ; and this 
gas, like X 3 , disappears when sparked with mercury vapour. 
Further, it is not easily condensed by the use of liquid air. 

It is a well-known fact that certain carbon compounds are 
extremely difficult to remove from the surface of glass; and 
a reference to the description of Thomson's apparatus which 
was given above will suffice to recall the fact that parts of it 
are stuck together with sealing-wax which would furnish a 
source of abundant carbon. Tap-grease also might yield cer- 
tain carbon derivatives. In the case of ordinary measurements, 
factors like these might be disregarded ; but the extraordinary 
sensitiveness claimed for the positive ray method unfortunately 
renders it all the more liable to difficulties arising from such 
minor sources of impurities. 

The fairest course in the matter appears to be to regard the 
problem of X 3 as still unsolved. It may be merely a carbon 
atom carrying four electrical charges; or it may be an allo- 
tropic modification of hydrogen. Some of the evidence points 
in one direction, some in another. It is too early yet to decide 
definitely in favour of either hypothesis. 

Some work has been carried out which appears to throw 
further light upon the nature of X 3 . Duane and Wendt 2 
found that when hydrogen is exposed to the bombardment of 
a -particles from niton, a contraction is observed. Part of this 
contraction is possibly due to the hydrogen being driven into 
the glass of the containing vessel. The remaining hydrogen 
appears to acquire properties differentiating it from the ordinary 
form of the element. It reduces sulphur to hydrogen sulphide, 
phosphorus to phosphine, arsenic to arsine and in a neutral 
solution of potassium permanganate it throws down manganese 

1 Collie and H. S. Patterson, Proc., 1913, 29, 217. 

2 Duane and Wendt, Phys. Rev., 1917, 10, 116 ; Lindt, J. Amer. Chem. 
Soc., 1919, 41, 545. 

* This gas was trapped in a jacket surrounding the discharged tube ; and 
it may have originated from molecules driven through the inner tube by the 
force of the discharge. 


dioxide. It is unstable and reverts in about one minute to the 
ordinary bi-atomic hydrogen. The activity of this new variety 
of hydrogen is not due to the formation of charged hydrogen 
ions, as the application of a strong electrostatic field seems to 
be insufficient to remove the active variety, which it would do 
were ions of hydrogen the only factors concerned. 

Wendt and Landauer 1 have examined the matter further. 
They find that by discharging an induction coil through a tube 
of hydrogen under 2-8 cms. pressure the same active form 
of hydrogen can be produced. The new gas is unaffected by 
passage through spirals immersed in boiling water or in a freezing 
mixture composed of ice and calcium chloride ; but it appears 
to be condensed by cooling with liquid air. The active form 
was also prepared by passing hydrogen through an ozonizer fed 
with a current of 20,000 volts. The gas formed in this manner 
combined directly with nitrogen to give ammonia. Attempts 
to prepare the active form of hydrogen by an application of the 
Schumann light were not successful. 

It is suggested as yet there is no proof that the new 
form of hydrogen is triatomic and may correspond to Thomson's 

X 3 . 

5. Conclusion. 

It may be desirable at this point to sum up the advantages 
offered by the positive ray method of analysis. In the first 
place, it permits the analysis of extremely small quantities of 
a gas mixture. Thus a quantity of helium which does not 
exceed 4 x 10~ 6 c.c. can be detected in a c.c. of air. 

Secondly, it carries our knowledge further than spectrum 
analysis can do. If we have only a trace of an element mixed 
with large quantities of other gases, the spectrum of the mix- 
ture may fail to reveal the presence of the trace owing to its 
characteristic lines being swamped by the spectra of its com- 
panions ; whereas in the case of the positive ray method, each 
constituent is sifted out from the others. 

Thirdly, in the case of a new element,' an examination of 
the spectrum tells us only that new lines are present ; but 
with positive ray analysis it may be possible to go much 

1 Wendt and Landauer, J". Amer. Chem. Soc., 1920, 42, 930. 


further. For example, if we find two parabolas characteristic 
of the new element, one must arise from the atom and the 
other from the diatomic molecule. If the substance is mon- 
atomic, there will be only one parabola, or, if there be more 
than one produced (owing to the atom taking up more than a 
single charge), we can detect the nature of the charged body. 

These facts prove that in the positive ray method we have 
gained a new and formidable weapon for the chemical armoury ; 
and since the subject is still new, many further developments 
may yet be looked for in its application to chemical problems. 
In a later chapter * some of these will be examined. 

* See Chapter XVII. 


1. Historical. 

IN 1895 Kontgen showed that when cathode rays impinged 
upon the end of a vacuum tube they gave rise to a green 
luminescent patch, from which was projected a series of rays 
the X-rays. From the phosphorescence of the Crookes' tube 
to the green phosphorescence of certain minerals is only a 
short step, and in 1896 Becquerel began an investigation of 
the latter phenomenon. He found that crystals of potassium 
uranium sulphate had the property of affecting a sensitive 
photographic plate (wrapped in black paper) in exactly the 
same way as it would have been affected by an X-ray discharge. 
The experiment led to the discovery of what are now called 
the Becquerel rays. 

These radiations are invisible to the eye, just as the X-rays 
are. They are given off by metallic uranium, and also by 
uranium salts. Like the X-rays, they can pass through thin 
sheets of glass or metal; and a further resemblance is to be 
found in the fact that neither set of rays can be refracted by 
ordinary means. Again, the Kontgen rays have the property 
of ionizing gases through which they are passed ; and it has 
been found that in this instance also, the Becquerel rays re- 
semble the others, though their action is much more feeble. 
If we charge the leaves of a well-insulated electroscope they 
will remain separated for a very considerable time, owing to the 
fact that ordinary dry air is a poor conductor of electricity ; 
but if we bring a piece of uranium near the electroscope, the 
Becquerel rays which are given off from the metal at once ionize 
the air, making it a better conductor of electricity, and thus 



the leaves of the electroscope fall together much more rapidly 
than they did before the uranium was brought near them. We 
thus have a method of determining the activity of any particular 
sample of uranium ; for we need only measure the rapidity 
with which an electroscope is discharged when the sample is 
placed in its vicinity. If the uranium is sending out many 
Becquerel rays, the electroscope leaves will soon fall together; 
if the Becquerel rays are few, the instrument will be very 
slowly discharged. 

It would naturally be concluded, from the above evidence, 
that the Becquerel rays and the Kontgen rays were identical. 
As a matter of fact, however, they are not so. We need not 
go into details here, as the matter will be dealt with in a later 

Soon after Becquerel' s work on uranium and its salts, 
Schmidt discovered that thorium also had radioactive pro- 
perties ; and since that time various other elements have been 
described which belong to the same class. Of these the most 
important is radium. 

We need not continue the history of the subject further 
in this section. The present chapter will be devoted to a 
discussion of the various radioactive metals. In the following 
chapter, some account will be given of the changes which these 
bodies spontaneously undergo. The chapter after that will 
contain a description of the gases which are evolved during 
radioactive change. Of course, in the space at our disposal, 
no attempt can be made to treat the subject in great detail. 
The properties of radium and its derivatives will be taken as 
typical and described at some length, while the allied sub- 
stances will be dealt with very briefly. 

2. Radium. 

In this and the succeeding sections we shall attempt to 
give an outline of the properties of the various radioactive sub- 
stances. To enter fully into the physical side of the question 
would carry us beyond the limits of the present volume : so, as 
far as possible, the subject will be treated from the point of 
view of chemistry, and the physical part of the subject will be 


dealt with only in so far as it aids the comprehension of the 
purely chemical side. 1 

In the introductory section of this chapter, attention was 
called to the fact that uranium salts throw off radiations which 
affect photographic plates just as the X-rays do. Now, uran- 
ium compounds are found naturally in the mineral pitchblende, 
which occurs usually (like the rare earths) in igneous rocks 
such as granite. Madame Curie, 2 in the course of an ex- 
amination of a great number of naturally occurring substances, 
discovered that in some cases the natural ore was much more 
radioactive than the amount of uranium salt contained in it 
would lead us to expect. For example, pitchblendes are about 
four times as active as metallic uranium ; chalcolite (a double 
phosphate of uranium and copper) is twice as active as uranium. 
Further, when Madame Curie prepared artificial chalcolite, 
she found that instead of being more active than metallic 
uranium, it was two and a half times less active. This proves 
conclusively that the activity of the natural chalcolite was not 
due entirely to the uranium contained in it, but must be 
attributed to the presence in the natural ore of some substance 
not found in the synthetic product. 

The presence of this new body in pitchblende and other 
minerals having been established, the question of extraction 
then arose. It was easy to carry out an ordinary analysis, 
and to determine by measurements with an electroscope 
whether the activity was a property of the nitrate or of the 
precipitate. We need not enter into any details with regard 
to the method of analysis employed. 3 A series of operations 
must be gone through before we can obtain a mixture con- 
sisting of radium bromide and barium bromide. 

After the isolation of the bromides of radium and barium 
from the mixture has been accomplished, it is necessary to 
separate them from each other. This is done by fractional 

1 For an account of the practical methods employed in radioactivity 
measurements the reader is referred to Makower and Geiger's Practical 
Measurements in Radioactivity. 

2 Thesis presented to the Faculte des Sciences, Paris ; see Chem. News, 
1903, 88, 85, 97, 134, 145, 159, 169, 175, 187, 199, 211, 223, 235, 247, 259, 271. 

3 Ebler and Bender, Zeitsch. anorgan. Chem., 1914, 88, 255 ; U.S.A. 
Bureau of Mines Bulletin 104 (1915) ;, Schlundt, J. Physical Chem., 1916, 20, 


crystallization, either from aqueous solution or from a solution 
of the salts in water acidified with hydrobromic acid. The 
latter is the better method of the two. It is possible to employ 
the method of fractional precipitation instead of crystallization, 
the mixed salts being thrown down from an aqueous solution 
by means of alcohol. In either case, the radium bromide is less 
soluble than the barium salt. 

The radium salts which are obtained in this w^ay resemble 
the corresponding barium compounds in many respects. 
Kadium chloride is isomorphous with barium chloride ; and the 
two salts, when they are freshly prepared, are similar to each 
other in appearance. It has been found that if the radium 
salt stands for a time it gradually becomes coloured, the tint 
ranging from yellow to rose-pink. The coloration becomes 
much more marked if a trace of impurity (such as a barium 
salt) be present in the crystals. 

Another property which is noticeably altered by lapse of 
time is the blue luminescence exhibited by freshly prepared 
radium salts or their solutions. The presence of barium salts 
in this case also appears to have some influence, for the blue 
light is more strongly exhibited by an impure sample of 
radium salt than it is by a pure one. 

An aqueous solution of a radium salt has been found to 
evolve a mixture of oxygen and hydrogen, and there appears 
to be no cessation of this process, which offers such a peculiar 
parallel to electrolytic action. 1 A still more extraordinary 
property of these salts, however, remains to be described. 
Curie and Laborde 2 observed that the temperature of a radium 
salt is always a little higher than the temperature of the air 
about it; in other words, the radium salts are continually 
giving out heat. Curie and Laborde showed that the quantity 
of heat disengaged in one hour by one gramme of pure radium 
.would amount to about a hundred gramme-calories. 

When a radium salt is placed in the flame of a Bunsen 
burner it gives rise to an intense carmine-red coloration, which 
is very characteristic. The spectrum of the element has been 
examined by several workers, and found to be quite different 
from any known spectrum. 

1 See also Eamsay, Monatsh., 1908, 29, 1013. 

2 Curie and Laborde, Compt. rend;, 1903, 136, 673. 


The atomic weight of radium was at one time the subject 
of considerable controversy, 1 but it is now taken as 226'0. 
This value places radium below barium in the alkaline earth 
column of the Periodic Table ; and the resemblance between 
the two elements fully justifies the position. 

Metallic radium has been isolated by Madame Curie and 
Debierne. 2 A solution containing about a decigramme of 
radium chloride was electrolysed, the cathode being mercury 
and the anode being platinum-iridium. After the electrolysis 
was completed, the amalgam of radium at the cathode was 
placed in an iron boat and heated in a current of hydrogen 
which had previously been purified by passage through the 
walls of a platinum tube heated in an electric furnace. At 
approximately 700 C., the mercury had all been driven off and 
the boat contained a brilliant white metal which fused sharply 
in the neighbourhood of 700 C. and began to attack the quartz 
tube in which the experiment was carried out. On exposure 
to air, a black film formed upon the metal, this being probably 
radium nitride. Eadium attacks water violently, forming a 
soluble oxide. A trace of the metal having fallen upon a sheet 
of paper, it was found that the paper was blackened and 
apparently carbonized. It appears that metallic radium is 
more volatile than barium. 

3. The Becquerel Rays. 

In the previous section we mentioned that the salts of 
radium and other radioactive elements emitted a series of radi- 
ations which have been termed the Becquerel rays, from the 
name of their discoverer. Since a very considerable part of 
the peculiar activity of radioactive bodies is closely connected 
with these rays, it is necessary at this point to enter into some 
consideration of the question. The first problem which we 
have to solve is that of the nature of the rays in question. 
Are they simple, or are they made up of a series of different 
types of vibration, such as a mixture of light and X-rays ? 

1 Madame Curie, Chem. News, 1903, 88, 159 ; Compt. rend., 1907, 145, 422 ; 
Runge and Precht, Physikal. Zeitsch., 1908, 4, 285 ; Watts, Phil. Mag., 1903, 
6, 64; Thorpe, Proc. Roy. Soc., 1908, A, 80, 298; Honigschmid, Monatsh., 
1912, 33, 253; Whytlaw-Gray and Ramsay, Proc. Roy. Soc., 1912, A, 86, 270. 

2 Madame Curie and Debierne, Compt. rend., 1910, 151, 523. 


To settle this point, we have two methods at our disposal. 
In the first place, we may interpose in the path of the rays 
some sort of filter which will give us a chance of separating 
one component from the rest ; or, secondly, we may pass the 
rays through a magnetic field, and determine whether they are 
uniformly deflected as a whole or whether they can be resolved 
into a series of vibrations having different deflections. 

We need not enter into any great detail with regard to the 
first of these methods. An experiment mentioned by Strutt x 
will make the application clear. Suppose that we charge an 
electroscope so that the leaves of it diverge from one another. 
If we now bring near the electroscope a small quantity of a 
radium salt, we shall find that the leaves fall together, say, in 
ten seconds. Let us next wrap the radium salt in a sheet of 
tin-foil and again place it in position near the electroscope ; we 
shall find that the leaves take longer to fall, say, a hundred 
seconds. We have thus filtered off some part of the rays. Now, 
if all the rays were of the same type, we should expect to in- 
crease the time of discharge to a thousand seconds by doubling 
the sheet of tin-foil. This, however, is not found to be the case ; 
but, instead, the rate of discharge hardly falls at all with the 
superposition of the second thickness of tin-foil. It is thus 
made clear that we have filtered off one set of rays by means 
of the tin-foil, but that there still remain other rays which the 
tin-foil sheets do not arrest. We have in this way established 
the presence of two types of vibration, one set stopped by 
tin-foil, the second not. We have now to discover whether or 
not we can further sift the rays to which tin-foil is transparent. 
To do this, we substitute for the tin-foil a thin casing of lead ; 
and here we find a similar behaviour. The addition of one 
sheet of lead causes a perceptible lengthening of the period of 
the electroscope's discharge ; but the interposition of a second 
slip of lead does not further diminish the velocity of the dis- 
charge to any marked extent. Thus again we have reached a 
border-line between two sets of rays, one of which will pass 
through the lead, while the other cannot do so. 

This method, however, is only a very rough one, and does 
not lend itself to measurements of such great accuracy as are 
obtainable by the second method, in which a magnetic field 

1 Strutt, The Becquerel Rays and the Properties of Radium, 1904, p. 51. 



is used. The actual details of the experimental methods 
employed need not be described in this place, 1 but we may 
give the results which have been obtained by the application of 
the electro-magnet to the problem. 

Suppose that we have a magnetic field applied at right 
angles to the plane of the paper and directed towards the 
paper ; the state of affairs may be represented by Fig. 2 1. 2 

FIG. 21. 

Here L is a piece of lead in which a hollow has been drilled. 
At the bottom of the hollow is placed some radium salt, E. 
Under the influence of the magnetic field, the radiations from 
the radium salt are split up into three groups, which have 
been termed by Eutherford the a-, /3-, and y-rays.* The y-rays 
are not deflected by the magnetic field, and are thus comparable 
to ordinary X-rays ; but their penetrating power is very much 

1 See Rutherford, Radioactivity, 1905, chaps, iv. and v. 

2 Madame Curie, Chem. News, 1903, 88, 169. 

* As will be seen later, the |8-rays do not arise from radium itself but come 
from certain other substances which appear in specimens of radium salts after 


greater than that of the X-rays. The /3-rays resemble the 
cathode rays of a high vacuum tube ; and like these they are 
deflected by the magnetic field in the manner shown in the 
figure. They appear to be streams of particles carrying a 
negative electric charge. If we place a sensitive photographic 
plate, PP, under the leaden cell containing the radium salt, we 
shall find that the spots where the j3-rays fall are affected by 
them. Finally, we come to the third set of components of the 
Becquerel rays. These, the a-rays, are deflected from the 
straight path of the discharge by the electric field ; but instead 
of being bent in the same way as the j3-rays, they are curved in 
the opposite direction. They are therefore positively charged 
particles and are now known to be helium atoms. We 
must now discuss the properties of these different rays in 

The y-rays, as we have mentioned, are not deflected to any 
extent by a magnetic field, even when this is very powerful. 
Their penetrating power also is very marked. For instance, 
Strutt l has observed that even 8 centimetres of sheet lead will 
not suffice to arrest these radiations. The origin of the j -rays 
can easily be surmised. They resemble X-rays so closely as to 
leave little doubt as to the identity of the two types; 2 and bearing 
in mind the fact that the j3-rays of radium and the cathode 
rays of a Crookes' tube are very closely allied, we must conclude 
that the -y-rays of radium salts are produced by the bombard- 
ment of the solid salt by the particles of the j3-rays produced 
within itself. 

The nature of the j3-rays seems to be beyond doubt. They 
are material particles projected from the surface of the radium 
salt, just as particles are driven out from the cathode of a high 
vacuum tube. Naturally, they do not all travel with the same 
velocity ; some sets of particles move more slowly than others, 
and it has been found possible to analyse the swarm of 
particles by deflecting them in magnetic fields. It has been 
found that the slowest-moving particles are most deflected 
as can be foreseen and it has been proved, further, that the 
)3-rays as a whole are less deflected than the cathode rays from 
Crookes' tubes. Hence it is clear that the particles forming 

1 Strutt, The Becquerel Rays, p. 83. 

2 J. A. Gray, Proc. Boy. Soc., 1912, A, 87, 489. 



the j3-rays of radium are moving at a much higher velocity 
than those which are shot out by the cathodes of vacuum 
tubes. It has been estimated that the velocities of light, the 
/3-rays, and the ordinary cathode rays are approximately the 
following : 

Light . . . 80 x 10* kilometres per second 

0-rays . . . (6 x 10 4 ) to (28 x 10 4 ) 
Cathode rays . (2 x 10 4 ) to (10 x 10 4 ) 

It appears, then, that the velocity of the electrons in the /3-rays 
is greater than that of any other known natural body. From 
the fact that the radioactive bodies are losing negative electri- 
city at a considerable rate owing to the departure of the 
electrons of the |3-rays, it is clear that the remaining portion 
of the salt must gradually acquire a positive charge, and that 
if loss of this charge be prevented by insulation it will eventu- 
ally become quite measurable. In point of fact, one experi- 
menter l noticed that when a sealed glass tube containing 
radium was opened after several months, a bright electric 
spark was produced. 

Turning now to the a-rays, we enter quite a different field. 
Here we have to deal with a series of helium atoms, travelling 
at high velocities (probably one- tenth that of light). These 
atoms are positively charged, and are deflected by magnetic 
fields in the direction opposite to that taken by the /3-particles. 
The extent of the deflection, however, is by no means so great 
as that observed in the case of the electrons of the )3-rays. 
The a-rays appear to be similar in character to the positive 
rays which can be produced in vacuum tubes. It seems 
certain that the heat generated by radium is due to the impacts 
of these particles. Crookes has devised an instrument, the 
spinthariscope, by means of which we can actually observe the 
effects of the collision between a-particles and ordinary matter. 
The spinthariscope consists of a zinc sulphide screen, above 
which a tiny fragment of a radioactive substance is suspended 
on the end of a steel pointer. When the screen is examined 
through a lens, it is found to be covered with tiny points of light 
which appear and vanish almost at once. These light-spots are 
apparently due to the phenomenon known as triboluminescence, 

1 Dorn, Physical. Zeitsch. 1903, 4, 507 


in which the breaking of a crystal causes a flash of light ; and 
it is supposed that the flashes in the spinthariscope are caused 
by the fracture of the zinc sulphide crystals under the impact 
of helium atoms discharged from the radioactive substance on 
the pointer. The triboluminescence phenomena on a larger 
scale can be observed by any one who rubs two pieces of sugar 
together in the dark. 

In the foregoing paragraphs we have discussed the proper- 
ties of the a-, )3-, and y-rays seriatim ; but we must bear in mind 
that the effects produced by radioactive substances in general 
are not caused by the separate action of each type or ray, but 
are really due to the united action of all three kinds. We shall 
now give a few instances of the changes which radium produces 
upon various substances when it is placed in their vicinity. 

The first effects which are recognizable are those due to 
fluorescence or luminescence of one kind or another. Suppose 
that we bring an X-ray fluorescent screen near a radium com- 
pound; the screen will become strongly lit up, just as it is 
when acted upon by the Eontgen rays. Bary 1 has found that 
many of the alkali metals and alkaline earths are also fluores- 
cent when brought into the neighbourhood of a radium pre- 
paration. Madame Curie 2 showed that paper, cotton, and 
other substances could exhibit the same phenomenon. Bec- 
querel 3 observed that while a ruby will fluoresce under the 
action of light-rays, it remains inert with regard to radium. 
Again, a diamond which shines in presence of the rays from 
radium does not light up when X-rays are thrown upon it ; 
and calcium sulphide appears to behave similarly. Many other 
instances of this action of the Becquerel rays might be quoted. 

If we expose a fluorescent substance of the type mentioned 
above to the continued action of the radium rays, it is found 
that the power of shining which it possesses does not remain 
constant, but slowly diminishes while at the same time the 
physical appearance of the fluorescent body changes. Barium 
platinocyanide on long exposure to the ray& grows darker in 
colour, finally becoming quite brown ; its luminescence also 
dies away gradually. Both the original colour and the power 

1 Bary, Compt. rend., 1900, 130, 776. 

2 Madame Curie, Chem. News, 1903, 88, 212. 

3 Becquerel, Compt. rend., 1899, 129, 912. 


of fluorescing can be regenerated by exposing the salt to light 
for a time. 

When some varieties of glass are exposed to the 'action of 
the Becquerel rays they become fluorescent, and at the same 
time become tinted brown or violet according to the alkali 
metal contained in them. If we warm the glass after the 
process has gone on for a time, we shall find that it loses its 
tint and becomes again transparent, regaining at the same 
time its original power of fluorescence. Not only so, but 
when heated it now possesses the power of spontaneous fluores- 
cence. The same is true for such minerals as fluorspar. 

Owing to the presence of the y-rays in the Becquerel rays, 
the latter have the power of ionizing gases, just as the X-rays 
do. We need not enter into details with regard to this pro- 
perty ; it is obvious that the electroscopic method of determin- 
ing activity depends upon it. 

In many cases the action of the Becquerel rays upon 
ordinary chemical elements or compounds is well marked. For 
instance, the Curies have shown that by their influence we can 
convert oxygen into ozone. 1 Again, if we submit yellow phos- 
phorus to the action of the radiations, it becomes changed into 
the red allotropic modifications. 2 On the other hand, it has been 
shown by Sudborough 3 that the presence of radium salts has 
no influence upon some geometrically isomeric substances which 
are transmuted into the stable form by the action of light. 
If we allow the Becquerel rays to act upon a solution of iodo- 
form in chloroform, we shall find that it becomes purple owing 
to the separation of iodine. 4 Ammonia 5 is decomposed by the 
Becquerel rays ; and water 6 is broken up, yielding oxygen 
and hydrogen. If a mixture of dry oxygen and hydrogen 7 
be allowed to remain in contact with radium bromide, 
however, no combination results. Hydrogen sulphide and 
sulphur dioxide are both 7 decomposed by the rays. When 
air is left in contact with a radium salt over mercury, nitrous 

Compt. rend., 1899, 129, 823. 

Becquerel, Compt. rend., 1901, 188, 709. 

Sudborough, Proc., 1904, 20, 166. 

Hardy and Willcock, Proc. Boy. Soc., 1903, 72, 200. 

Perman, Trans., 1911, 99, 132. 

Usher, Jahrb. Radioaktiv. Elektronik, 1911, 8, 323. 

Baker, Royal Institution Lecture, March 11, 1911. 


oxide is produced in comparatively large amounts. 1 The 
action of the Becquerel rays upon hydrogen peroxide has been 
studied; 2 but in this case it is difficult to know exactly how 
much must be ascribed to the direct action of the rays and 
how much to the influence of alterations upon the glass of the 
containing vessel, which, in turn, will affect the rate of de- 
composition of the peroxide. Other examples 3 of the chemical 
effects of the Becquerel rays have been observed, but it is not 
necessary to deal with them here. 

The physiological effects of the radium radiations are very 
striking. If a radium salt be allowed to remain in contact 
with the skin for even a few minutes, it is apt to produce ex- 
tremely painful eruptions ; and these are the more troublesome 
owing to the fact that they take a considerable time to make 
their appearance. Thus even by bringing radium near the 
skin, without actual contact, we may produce very deep-seated 
changes in the tissues without having any outward sign that 
we have injured them. The action of the radiations upon 
the tissues is apparently due to the break-down of lecithine, 
which makes up a considerable part of the epidermis. This 
destructive character of the radium salts, however, is not 
without its uses, for in cases of rodent ulcer it has been 
successfully employed to cure the disease ; and at the present 
time we seem to be on the verge of considerable advances in 
this direction. Hardy 4 has found that the coagulation of 
globulin may be brought about by the influence of the 
Becquerel rays, and in this case the action appears to be due 
to the positive charge carried by the a-particles. 

1 Soddy, Annual Reports, 1911, VIII., 299. 

2 Korosy, Pfluger's Archiv., 1910, 137, 123; Kailan, Monatsh., 1911, 32, 

3 Lind, Monatsh, 1912, 33, 295; Kailan, ibid., 71; Sitzungsber., K. Akad. 
Wiss. Wien., 1912, 121, 1329. 

4 Hardy, Proc. Physiol. Soc., 1908, XXIX. 



1. The Disintegration Theory. 

AT the time of their discovery, the radioactive properties of the 
radio-elements seemed capable of explanation on either of two 
lines. According to purely physical ideas, radium might be 
regarded as a kind of transformer which took up energy from 
its surroundings and subsequently emitted the same energy 
in other forms, just as a fluorescent substance absorbs certain 
wave-lengths of light and emits the energy thus acquired in 
the form of a characteristic fluorescence. This hypothesis has 
now been shown to be invalid. The second view assumes that 
radium is an unstable material which undergoes decomposition ; 
and that this decomposition is accompanied by the emission 
of matter and energy. 1 

The disintegration theory, as it is termed, can be very 
briefly explained. It assumes that the radio-elements are 
unstable and spontaneously undergo change. During the 
change, Becquerel rays are emitted and the original radio- 
active atom is converted into a new atom with a nature 
different from the parent. This new atom may in its turn 
be radioactive and emit rays, changing in the process into a 
third atom, different from its two predecessors. Thus radium 
emits an a-ray and changes to the inert gas niton; niton 
in its turn emits a new a-ray and becomes converted into 
radium-A ; the expulsion of yet another a-ray from radium-A 
leads to the production of radium -B ; and after this a /3-ray is 
emitted and the radium-B changes into radium-C, and so on. 

1 Rutherford and Soddy, Phil. Mag., 1903, 5, 576. 
1 66 


Thus the ejection of an a- or of a /3-ray from a radioactive 
element leads to the disintegration of the original element and 
the formation of a new element which may or may not have 
radioactive properties. 

This theory can be established by evidence of the following 
nature. Suppose that we have a radioactive element A, which 
disintegrates into a second element B, by radioactive change. 
In any ordinary specimen, we shall have a mixture of A and B, 
since A is always producing B. Now let us suppose that A 
and B can be separated by chemical means. If we remove all 
traces of B, then we shall have a pure specimen of A; but 
after this has stood for a time it will become contaminated by 
traces of B, which will be formed by the disintegration of A. 
It was an actual experiment of this type which led to the 
enunciation of the disintegration theory. 

Crookes, 1 in the course of an investigation of the properties 
of uranium, found that if a uranium salt was precipated by 
means of ammonium carbonate and the precipitate treated with 
excess of the reagent, the precipitate almost entirely redissolved, 
leaving behind it only a very slight residue. The redissolved 
uranium was found to be radioactively inert, while all the 
original radioactivity appeared to be concentrated in the small 
residue. The residue Crookes termed uranium X. So far, 
there was nothing out of the common; it appeared that the 
activity of uranium was really due to the presence in it of 
this trace of uranium X. A much more important result was 
obtained later, however, when it was found that the inactive 
uranium, after standing for some months, became once more 

About the same time, Becquerel 2 separated uranium and 
uranium X in another way. He mixed solutions of uranium 
and barium salts, and then, on precipitating the barium as 
sulphate, he found that the barium had acquired radioactive 
properties, while the uranium had lost them. After standing 
for a year, the two products were re-examined, and it was found 

1 Crookes, Proc. Roy. Soc., 1900, 66, 409. 

2 Becquerel, Compt. rend., 1900, 131, 137 ; 1901, 133, 977. 

* The activity determinations were made photographically, and are thus 
concerned only with the j8-rays ; had the o-rays been used as a test, the 
uranium would not have appeared to lose much activity. 


that by that time the barium had lost its activity, while the 
uranium was again as radioactive as it was as the beginning of 
the experiments. 

The only possible conclusion which can be drawn from 
these results is that the activity of uranium is due to some 
substance which is produced spontaneously by uranium. 

Eutherford and Soddy 1 found a similar series of phenomena 
in the case of thorium, and they proceeded to investigate 
quantitatively the rate at which thorium-X lost its activity. 
They found that, starting with inactive thorium and active 
thorium-X, after four days the thorium had regained half 
its original activity, while in the same time the thorium-X 
had lost half its radioactive powers. The rate of decay and 
recovery can be expressed in both cases by exponential 

' For the decay of Th X, I, = T x e~ Kt 

For the recovery of Th activity, I< = I (l e~ xf ) 

in which I represents the initial activity of the thorium or 
the thorium-X,' I t the activity after a time t has elapsed, and e 
the base of the Napierian logarithms. The factor A, it is found, 
is the same for both decay and recovery. Similar results were 
obtained with uranium and uranium-X, except that in this 
case the period during which the uranium-X lost half its 
activity was about twenty-two days, instead of four as in 
the case of thorium. 

The most striking peculiarity of this decay and regenera- 
tion is the fact that it is totally unaffected by changes of tem- 
perature, even a white heat appearing to have no accelerating 
influence. This differentiates it from ordinary chemical re- 
actions, which are all more or less susceptible to changes of 

There is one other point which we must mention. It has 
been shown that the degree of activity of any salt is directly 
proportional to the amount of the radioactive element in the 
salt, and has no connection with the acidic part of the molecule. 
The activity of radium bromide is different from that of radium 
carbonate, and depends purely, in each case, upon the percentage 
of radium metal in the salt. 

1 Rutherford and Soddy, Trans., 1901, 81, 321, 837. 


From the foregoing evidence, we can draw certain con- 
clusions. In the first place, the fact that the percentage of 
radioactive element present in a salt is the measure of its activity 
proves conclusively that radioactivity is a property of the atom, 
and not of the molecule. Secondly, the fact that radioactive 
change is independent of temperature proves that we are not 
dealing with an ordinary molecular decomposition. We must 
therefore have to do with some new atomic property. Finally, 
since the a-particles which are driven out in the Becquerel rays 
are material bodies, we are obviously witnessing the breakdown 
of some material system ; and from what has gone before it is 
obvious that this system cannot be a molecular one. It must 
therefore be atomic. 

Another matter may conveniently be mentioned at this 
point. Let us assume that we have a radioactive substance 
X which is disintegrating and yielding a second element Y ; 
and that this element in its turn is disintegrating into a third 
one Z. If we start with a pure specimen of X, we shall soon find 
it contaminated with a certain amount of Y ; and the quantity 
of Y present would gradually increase, provided no further 
change were to take place. Actually, however, Y in its turn 
is breaking down ; so that obviously a time will come when 
X is producing Y at the same rate as Y is decomposing. The 
ratio of the amounts of X and Y then present will remain con- 
stant after this point has been reached; and X and Y are 
then said to be in radioactive equilibrium. 

2. Multiple Disintegration. 

It will be recalled that the a-rays of radioactive elements 
are made up of positively charged helium atoms, whilst the 
j3-rays are streams of electrons. If an element emits a-rays, 
its disintegration involves a loss of matter from its atom in the 
form of helium atoms ; and 'consequently its disintegration 
product must have an atomic weight four units lower than 
that of the parent substance. On the other hand, the ]3-ray 
change involves only the displacement of an electrical charge 
and entails no loss of matter ; so that the product will have 
the same atomic weight as the parent. 

From this it follows that if a radioactive element is 


susceptible to disintegration along both lines simultaneously it 
is possible that it may give rise to two different products, one 
of which will have the same atomic weight as the parent (|3 -ray 
change) whilst the second product has an atomic weight four 
units lower than that of the parent (a-ray change). 

Several cases of this type have been observed among the 
radioactive elements. In such cases it is usually found that 
one mode of disintegration predominates over the other. 

3. The Radioactive Series. 

From what has already been said it will be clear that among 
the radio-elements we can trace certain family relationships. 
A given radio-element disintegrates and gives rise to a product ; 
this in its turn produces a third substance, and so on. All 
these elements stand in a genetic relation to one another and 
may be regarded as a " radioactive series." 

In the earlier days of radioactive research three of these 
series were recognized and were named, after their parent 
substances, the uranium, thorium, and actinium series. It has 
now been shown that the uranium series includes as one of its 
branches the actinium series also ; so that at the present time 
we have only two series : the uranium series and the thorium 

To avoid cumbrousness, however, it will be convenient to 
treat the material in the older form, indicating merely the point 
of junction of the uranium and actinium series. 

The thorium series is the simplest, and we may therefore 
begin with it. In the following table l (see opposite page) the 
main facts are given as to the atomic weights, average life, type 
of rays emitted by the element, and also the group in the 
Periodic Table in which its properties enable it to be placed. 

At thorium-C multiple disintegration occurs, and it is found 
that thorium-C gives rise to two disintegration products. One 
of these, thorium-C', is the product of a /3-ray change. Its 
atomic weight is therefore the same as that of its parent, 
viz. 212. The second disintegration product of thorium-C 

1 The constants in these tables are taken from Soddy, Chemistry of the 
Radio-elements (1914). 



is termed thorium-D. It results from an a-ray change in 
thorium-C and its atomic weight is therefore 208. The dis- 
integration of thorium-C results in the change of 65 per cent, 
of it into thorium-C' and 35 per cent, of it into thorium-D. 









2-6 X 10 l years 







7*9 years 
8'9 hours 





91 years 





5-25 days 



Thorium Emanation 


78 seconds 






0-2 seconds 
15-4 hours 



Thorium-C . 


87 minutes 

a and 


Both thorium C' and thorium-D break down to end-products 
having the atomic weight 208. The constants for the two 
substances may be given here. 





Thorium-C' . 


10 " seconds 



Thorium-D . 


4-5 minutes 


End-products . 




We must now turn to the uranium series and, leaving out 
of account for the present the branching in the chain which 

* This radiation is so soft that it has not yet been experimentally detected. 
Its existence, however, can be deduced from Soddy's Law (see Chapter XVIII.). 



introduces the actinium series, deal simply with the successive 
disintegrations which give rise to the radium group. 









8 X 10 9 years 



Uranium-Xi . 


35-5 days 


Uranium-X 2 . 


1-65 minutes 





3 X 10 6 years (?) 



Ionium . 


2 x 10 s years (?) 



Radium . 


2440 years 



Niton . 


5-55 days 




4-3 minutes 



Radium- B 


38-5 minutes 





28*1 minutes 

a and ft 


At this point, as in the case of the thorium series, a 
branched chain begins. Eadium-C disintegrates simultaneously 
into radium-C' and radium-C 2 . We may take the further 
disintegration of these in succession. 








10~ fl seconds 



Radium- D 


24 years (?) 





7-20 days 





196 days 



End-product . 




Eadium-0 2 yields the end-product of its series directly 






Radium-C 2 


1*9 minutes 




End-product . 




It is now necessary to deal with the earliest branching in 
the series of uranium disintegration products and to show the 
connection between the uranium and actinium series. The 
matter is one which is not yet * cleared up completely ; and 
further research will probably advance our knowledge of the 
subject. At present it is only possible to give a provisional 
account of the problem. 1 

Our present knowledge of the subject can be brought into 
agreement with either of two alternative schemes for the dis- 
integration of the uranium series 

Of these, the former is considered by Soddy to be the more 
probable. It will be seen that in either scheme the parent of 
actinium is the element uranium-Z, which occupies the position 
of eka-tantalum in the Periodic Table. 

The table on next page shows the constants for the mem- 
bers of the actinium series as far as they have been de- 
termined. The atomic weight column is omitted, as definite 

* 1920. 

1 See Soddy and Cranston, Proc. Roy. Soc., 1918, A, 94, 384; compare 
Hahn and Meitner, Physikal Zeitsch., 1918, 19, 208. 



results have not yet been obtained. Uranium Y may have an 
atomic weight of either 230 or 234, so that doubt is left as to 
the remainder of the weights. 

It will be noticed that in the following scheme no branch- 
ing of the chain is shown at the bottom, though branching 
occurs in both the thorium and uranium series as the end- 
products are approached. It is possible, however, that there 
may be a multiple disintegration of actinium -C. 



Average Life. 



Uranium-Z (Eka-tantalum) 
Actinium .... 
Kadioactinium . 
, emanation . 

2-2 days 
5000 years l 
28-1 days 
5-6 seconds 
0*003 second 




1 ; 

52-1 minutes 







End-product .... 


4. Potassium and Rubidium. 

The numerous investigations 2 which have been made of the 
radioactive properties of potassium seem to show that this 
element possesses distinct activity. Eubidium also appears to 
have radioactive powers. 

5. Radioactive Recoil. 

We have already seen that, when an a-ray change takes 
place, a helium atom is ejected by the disintegrating atom. 
Now since action and reaction are equal, it is clear that the 
residue of the original atom the disintegration product will 

1 Soddy and Cranston, Proc. Roy. Soc., 1918, A, 94, 384. 

McLennan and Kennedy. Phil Mag., 1908 (vi.), 16, 377 ; Levin and 
Ruer, Physikal. Zeitsch., 1908, 9, 248 ; Strong, Physical Review, 1909, 29, 170; 
Henriot, Compt. rend., 1909, 148, 910; Henriot and Vavon, ibid., 149, 30; 
Biichner, Proc. K. Akad. Wetensch. Amsterdam, 1909, 12, 154; Campbell, 
Proc. Camb. Phil. Soc., 1909, 15, 11 ; Ebler, Chem. Zeit., 1908, 32, 812. 


move in a direction opposite to that taken by the helium atom, 
as a rifle recoils when the bullet is fired from it. And just as 
in the case of the rifle, the momentum of the helium atom and 
the momentum of the disintegration product will be equal to 
one another. 

It may be well to choose a concrete case so that the matter 
may be readily grasped. When radium-A disintegrates, it 
gives rise to a helium atom with atomic weight 4 and an atom 
of radiurn-B with an atomic weight 214. The helium atom is 
ejected with a velocity of 1/77 X 10 9 cm. per second, so that its 
momentum is 4 X 1'77 X 10 9 in atomic units. The momentum 
of the atom of radium-B is clearly 214 x v t where v is the 
velocity of that atom. Whence we have, since the two 
momenta are equal in numerical value 

4 x 1-77 x 109 
v = - . - = 3-3 x 10 7 cm. per sec. 

The atom of radium-B formed during the disintegration will 
thus be moving at a velocity of 3*3 X 10 7 cm. per second; 
and this velocity may be quite sufficient to carry it away from 
the mass of radium-A in which it was formed. This pheno- 
menon is termed radioactive recoil ; and in some cases it forms 
a means of separating a disintegration product from its parent 
element, e.g. radium-B and radium-A. 

Now if the original radium-A be deposited in a thin layer 
upon the surface of a plate, it will be evident that, since recoils 
will be taking place in all directions, half of the freshly-formed 
radium-B will recoil into contact with the plate whilst the re- 
mainder will fly upwards and escape from the plate's surface. 
In actual practice, some of the particles, even when they are 
ejected in an upward direction, fail to get clear of the plate. 
The ratio of the number of atoms of radium-B actually escap- 
ing from the plate to the number of atoms which theoretically 
should escape is termed the efficiency of the recoil. 

6. The G-eiger-Nuttall Relation. 

A most interesting relationship has been detected by Geiger 
and Nuttall 1 which connects together the period of life of a 

1 Geiger and Nuttall, Phil. Mag., 1911, 22, 619 ; 1912, 23, 439 ; 24, 647. 


radio-element and the velocity with which it ejects an a-particle 
during disintegration. 

The range of a particle is the extreme distance over which 
that particle can travel through a gas before coming to rest. 
This range is proportional to the cube root of the initial velo- 
city of the particle. Thus any relation which holds good for 
the range must also hold good for the velocity of the particle. 

Now if the logarithms of the ranges of a-particles from 
a number of different radio-elements be plotted as abscissae 
whilst the logarithms of the life periods of the same elements 
are plotted as ordinates, it is found that a series of straight 
lines is obtained. Each of the three disintegration series : 
radium, thorium, and actinium, is found to lie on one straight 
line ; and the three lines lie parallel and quite close together. 

It will be seen that this relationship, though entirely 
empirical, gives us a means of calculating the life period of 
any radio-element provided that we can measure the range 
of the a-particles emitted by it. Thus it is quite possible to 
calculate the life period of elements whose existence is so 
transitory that direct measurement would be difficult. 

As to the physical meaning of the Geiger-Nuttall relation, 
we can suggest nothing definite ; 1 but it tends to show that 
some common link unites all the members of a given series 
and differentiates them from elements belonging to a different 
series. As Soddy puts it : 2 " The atoms in their successive 
disintegrations preserve a feature distinctive of their origin.'' 
But now that the radium and actinium series have been 
traced to their common origin in uranium, the Geiger-Nuttall 
relation appears, if anything, more mysterious than ever. 

1 See Chapter XIX. for a suggestion as to its meaning and compare 
Lindemann, Phil. Mag., 1915, 30, 560. Cf. Wolff, Physikal. Zeitsch.,1920, 21, 

2 Soddy, Chemistry of the Radio-elements, Part I., p. 29 (1914). 



1. The Emanations of Thorium and Actinium. 

IN a volume of this size, no attempt can be made to deal in de- 
tail with the properties of the thirty and more radio-elements ; 
and for a full account of these reference must be made to 
Soddy's Chemistry of the Radio-elements. Niton, however, 
has played such a distinctive part in the history of the subject 
that it deserves description in some detail; and the present 
chapter will therefore be devoted to an account of the three 
radioactive gases : thorium emanation, actinium emanation, 
and niton. 

In 1899 Owens * was engaged in an examination of the 
radiation effects of various thorium derivatives, and in the 
course of the work he observed that the radiations were by 
no means constant when measured electrically. Further in- 
vestigation showed that the inconstancy was due to air-currents 
about the apparatus; for when the experiments were carried 
out in closed vessels, the ionization of the air reached a maxi- 
mum and then remained constant. Now, if radioactive effects 
can be influenced simply by passing air across the radioactive 
substance which forms the subject of the experiment, it seems 
probable that part, at least, of the radioactive influence is due 
to some material which can be mechanically blown away by 
the current of air. This material might be either a gas or a 
cloud of tiny particles of thorium, which had been loosened 
from the main body of the radioactive mass. 

Rutherford 2 pursued this line of research, and was able to 

1 Owens, Phil. Mag. (6), 1899, 48, 360. 

2 Rutherford, Phil. Mag. (6), 1900, 49, 1. 

177 N 


prove that thorium did actually liberate something which we 
may term an emanation, and that this emanation had radio- 
active properties. He showed that the properties of the 
emanation very closely approximated to those of a gas. For 
instance, it can be blown about by air-currents; it can be 
bubbled through liquids ; a mica stopper will prevent it escap- 
ing ; and, finally, it causes no deposition of water-globules in 
the dust-counter. This last experiment proves that the emana- 
tion is not a cloud of thorium particles, for these would of 
course form nuclei for the condensation of water; while the 
molecules of a gas are boo small to produce this effect. 

The activity of the emanation was found to decay very 
rapidly. To measure this decrease in activity, a very ingenious 
method was employed, the outline of which is as follows. Pure 
air was blown across some thorium salt, and in this way be- 
came mixed with the emanation from the thorium ; the mixed 
gases were then passed into a long brass cylinder through 
whose walls projected three insulated electrodes. The brass 
cylinder was insulated and connected with a battery ; and the 
current through the gas was measured by means of an electro- 
meter in the usual way. It was found that the current 
diminished progressively along the cylinder ; and by passing 
the stream of mixed gases at different rates through the tube, 
it was possible to determine the rate of decay of the emana- 
tion. Le Eossignol and Gimingham 1 found that the activity 
of the emanation fell to half value in fifty-one seconds. 
Thorium emanation condenses at -120 C., and the process, 
which is gradual, is complete at 155 C. 

In the light of this work, it seemed probable that the other 
radioactive elements also might give off analogous emanations. 
Debierne 2 showed that this was true in the case of actinium, 
which gives off a gas having properties similar to those of 
thorium emanation, though its time of decay is very much 
shorter about four seconds. It condenses between - 100 C., 
and 143 C. This substance does not call for detailed treat- 
ment, as the general properties of thorium and actinium emana- 
tions resemble those of radium emanation, with which we 
shall now deal in some detail in the next section. 

1 Le Eossignol and Gimingham, Phil. Mag. (6), 1904, 8, 107. 

2 Debierne, Compt. rend., 1903, 136, 146. 


2. Niton* 

The discovery of thorium emanation speedily led to that 
of the emanation from radium, which was detected in 1900 by 
Dorn. 1 It resembles the thorium derivative in most respects, 
but is much more durable. 

When we consider the amount of emanation which is 
evolved from a given quantity of radium, the first thing which 
forces itself upon our notice is the difficulty which would be 
experienced in handling the minute quantity of gas which can 
be obtained. It has been found that the amount of emana- 
tion from 1 gramme of radium in radioactive equilibrium has 
a volume of about one-tenth of a cubic millimetre. 2 Now, the 
usual quantity of a radium salt used by experimenters varies 
from 20 to 60 milligrammes, and it must be remembered that 
only a part of this salt is radium. From these figures it will be 
seen that the volume of emanation obtainable at any time 
will be very minute. Consequently, it is necessary to devise 
some method by means of which we can transfer these tiny 
bubbles of gas from one vessel to another. It has been found 
that this is best accomplished by mixing the emanation with a 
large quantity of neutral gas. The mixed gases can then be 
transferred from vessel to vessel without any appreciable loss 
of emanation. One of the simplest ways of obtaining the 
emanation is to pump off the gas from a solution of radium 
bromide in water. Pure niton can then be obtained by ex- 
ploding the mixture of hydrogen and oxygen thus pro- 
duced along with the niton, condensing the niton by cooling 
with liquid air and pumping off any excess hydrogen and 
residual helium. 

Niton behaves as an ordinary gas. It obeys Boyle's Law, 8 
and diffuses like other gases when placed in a vessel. Experi- 
ments have been made by various workers, 4 with a view to 

* This name has been proposed by Ramsay and Gray as a convenient 
substitute for " radium emanation." 

1 Dorn, Abh. d. Naturforscher Gesellsch. Halle, 1900. 

2 Gray and Eamsay, Trans., 1909, 95, 1073. 

3 Ramsay, Compt. rend., 1904, 138, 1388. 

4 Rutherford and Brooks, Trans. Eoy. Soc. Can. (2), 1902, 7, 21 ; Chem. 
News, 1902, 85, 196 ; Curie and Danne, Compt. rend., 1903, 136, 1314 ; Bum- 
stead and Wheeler, Amer. J. Sci. (4), 1904, 17, 97 ; Makower, Phil. Mag. (6), 
1905, 9, 56. 


determine the molecular weight of niton from its rate of diffu- 
sion, but the results are, of course, extremely inaccurate. It 
must be borne in mind that in these diffusion experiments we 
are dealing with a very minute quantity of niton mixed with 
a very large volume of some indifferent gas, so that the results 
are affected by many factors which do not come into view in 
ordinary diffusion experiments. All that we can safely con- 
clude from these investigations is that the molecular weight of 
niton, to judge from its density, must be very high, probably 
over a hundred. As we shall see later, this estimate is under 
the mark. 

Niton has been liquefied and its critical constants have 
been determined by Earn say and Gray. 1 The liquid emanation 
is colourless and transparent by transmitted light. It is 
phosphorescent and shines with a colour which varies with the 
nature of the glass of the tube in which it is enclosed, the 
usual tints being green to lilac. When highly compressed, 
the tint resembles that of a cyanogen flame, being slightly 
bluish-pink. The solid emanation is not transparent. It melts 
at -71 C. Like the liquid, the solid phase of the emanation 
is phosphorescent ; but the colour of the phosphorescence is 
much more brilliant, and varies with the temperature. On 
cooling below the melting-point, the tint of the emitted light 
is steel-blue; further cooling changes it to yellow, and it 
finally becomes orange-red. On warming the tube containing 
it, the colours reappear in inverted sequence. The red phos- 
phorescence disappears at 118 C., while at 59 C. or 
60 0. the liquid is dull bluish-green. The critical tempera- 
ture is 377*5 absolute, and the vapour-pressure at this tem- 
perature is 47,450 mm. 

The atomic weight of niton was determined by Gray and 
Eamsay, 2 whose method consisted in weighing a given volume 
of the gas. When it is remembered that the total volume 
which they were able to obtain at any one time was less than 
0*1, some idea of the difficulties of the research will be 

Evidence with which we shall deal later in this chapter 
proves conclusively that niton belongs to the inactive gas 

1 Gray and Ramsay, Trans., 1909, 95, 1073. 

2 Gray and Ramsay, PTQC. Boy. Soc., 1911, A, 84, 536. 


group of elements ; and making the usual assumptions with 
regard to the periodic arrangement of the elements, we can 
predict that niton will lie somewhere above xenon in the series. 
Its atomic weight might therefore be either 176 or 222 ap- 
proximately, as the following figures show, since the difference 
between two successive atomic weights in the lower section of 
this group is generally about 45 : 

Helium. Neon. Argon. Krypton. Xenon. I. II. 

4 20 40 83 130 176 222 

If the figure 222 were correct, then O'l of niton would 
weigh less than 1/1400 milligramme ; and in order to weigh 
this small mass with sufficient accuracy it was necessary to 
devise a balance which would turn with a load of not more 
than a hundred- thousandth of a milligramme. The production 
of this balance is certainly the high- water mark of modern 
physico-chemical technique. 1 We cannot enter into details of 
the construction of this apparatus, but must confine ourselves 
to a brief description of its essentials. The beam of the balance 
is formed from threads of quartz fibre, and carries a small 
mirror of platinized silica. Instead of weights, a counterpoise 
is used which consists of a small quantity of air sealed up in a 
quartz bulb. The whole balance is surrounded by an air-tight 
case in which the pressure can be varied by means of a pump. 
When the air-pressure in the balance is the same as that in the 
bulb, the apparent weight of the bulb's contents is nil. In a 
vacuum, the sealed-up air exerts its full weight, as it is 
not counterpoised by the buoyancy of the air in the case. 
At any pressure intermediate between ordinary pressure and 
a vacuum, the apparent weight of the air in the bulb can be 
calculated. In this way, by varying the pressure of air in the 
case, we can bring the balance into equipoise ; and this can be 
determined by throwing a beam of light from the platinized 
mirror on to a scale some feet away from the apparatus. Many 
corrections are necessary in the course of a weighing, but for 
an account of these the reader is referred to the original paper. 
The results obtained by Kamsay and Gray give a mean atomic 
weight of niton equal to 223 ; so that niton should lie two 
places above xenon in the Periodic Table. 

1 Steele and Grant, Proc. Roy. Soc., 1919, A, 82, 580. 


The spectrum of niton has been examined by Eamsay and 
Collie, 1 who have found that it closely resembles in general 
characteristics the spectra of the inactive gases. The spectrum 
fades very soon, and is replaced by the hydrogen spectrum. 
There is one bright line at 5595 ; Kamsay and Collie suggest 
that this may be identical with the line in the spectrum of 
lightning 2 which does not seem to have been identified with 
that given by any known gas. 

Like radium itself, radium emanation spontaneously gives 
out a very considerable quantity of heat. The maximum value 
of heat liberated per hour from the emanation generated by 
one gramme of radium is given by Rutherford 3 as 75 calories ; 
this includes the heat emitted by the disintegration products 
of the emanation. Mton gives out only a-rays. 4 

It has been shown by Curie and Debierne 5 that the amount 
of emanation evolved by radium is independent of the pressure 
to which the radium is subjected ; and it has been found that 
changes of temperature also appear to be without influence 
upon the rate of formation. 

We must now turn to the chemical nature of niton. It 
has been shown by Eutherford and Soddy 6 that the emana- 
tions of thorium and of radium both showed an extraordinary 
inertness even when submitted to the action of strong chemical 
reagents. For example, no change could be detected in the 
gas after passing it over red-hot platinum black, or tinely 
divided palladium, lead chromate, magnesium powder, or zinc 
dust. Such inertness can be paralleled only by the elements 
of the argon group ; and it seems evident that niton must be 
reckoned as belonging to that class of bodies. Ramsay and 
Soddy 7 have made even more stringent tests, by sparking 
niton with oxygen in presence of alkali a process which brings 
even nitrogen into combination as well as by passing a mixture 
of air and niton over a highly heated mixture of magnesium 

1 Eamsay and CoUie, Proc. Roy. Soc., 1904, 73, 470; Rutherford and 
Royds, Phil. Mag., 1908, 16, 313 ; Watson, Proc. Roy. Soc., 1909, A, 83, 50. 
Pickering, Astrophys. J., 1901, 14, 368. 
Rutherford, Radioactivity, p. 431 (1913). 
Rutherford and Soddy, Phil. Mag. (6), 1903, 5, 445. 
Curie and Debierne, Compt. rend., 1901, 133, 931. 
Rutherford and Soddy, Phil. Mag. (6), 1902, 4, 580 ; 1903, 6, 457. 
Ramsay and Soddy, Proc. Roy. Soc , 1902, 72, 204. 


powder and lime. In the latter case the mixture of gases was 
passed for three hours across the magnesium-lime mixture which 
was heated to a bright redness ; the measurements of the radio- 
activity of the niton made before and after the experiment 
gave exactly the same result. In the course of these experi- 
ments it was found, further, that niton is unattacked even by 
phosphorus burning in oxygen. 

From the results quoted in the foregoing paragraphs, we 
can now assign to niton a place in the Periodic Table. From 
the chemical evidence, it is clear that niton is one of the 
inactive gases ; and the evidence of its physical constants 
makes it practically certain that it must lie below xenon in 
the Table. 

If niton per se is inert, it displays a very great influence 
when brought into contact with other substances. This in- 
fluence has nothing whatever to do with the chemical reactivity 
of the gas, but is due purely to its radioactive powers. We 
must now mention one or two experiments which have been 
carried out in this field. 

Giesel 1 noticed that when a solution of radium bromide 
was allowed to stand, it evolved some gas which investigation 
showed to be chiefly hydrogen. Eamsay and Soddy 2 found 
that the gas mixture contained 29 per cent, of oxygen, the 
rest being hydrogen. The slight excess of hydrogen they 
ascribed to contact between the gas mixture and the grease of 
a tap, which would remove some of the oxygen. 

Kamsay 3 carried out a further series of experiments upon 
the action of niton on water, and found that there was a mean 
excess of 5*51 per cent, of hydrogen over and above the quantity 
required to form water with the oxygen liberated. When the 
gases were stored over mercury, the percentage of hydrogen 
was still greater, owing to some of the oxygen being used up 
in oxidation of the mercury. When mercury is not present, 
various causes of the presence of the excess of hydrogen might 
be suggested. It was shown that the corresponding amount of 
oxygen was not lost by oxidation of the radium bromide to 

1 Giesel, Ber., 1902, 35, 3605. 

2 Ramsay and Soddy, Proc. Roy. Soc., 1903, 72, 294. 

3 Ramsay, Trans., 1907, 91, 931 ; Cameron and Ramsay, ibid., 1908, 93, 
966, 992. 


bromate, nor was there any formation of ozone or hydrogen 
peroxide; no bromine is liberated from the radium bromide. 
In all cases care was taken to prevent the gases coming in 
contact with tap-grease. Ramsay showed, further, that the 
action of niton is a reversible one; for while, on the one 
hand, it decomposes water to produce electrolytic gas, it also 
has the faculty of recombining oxygen and hydrogen to form 
water again. In later experiments it was found that traces 
of hydrogen peroxide may be formed by the action of niton 
upon water. 

Enough has been said to show that niton possesses simul- 
taneously two sets of properties which, before its discovery, it 
would have been hard to believe capable of co-existence. On 
the one hand, it is itself chemically inert ; but, on the other 
hand, it can influence the chemical properties of other sub- 
stances to a very marked extent. There is one other property 
which it exhibits, and this is perhaps the most extraordinary 
of all. Spontaneously, it disintegrates and yields, another 
element, helium. 

Before dealing with the disintegration of niton, however, 
it will be well to mention some experiments which throw 
some light upon the earlier stages of the decompositions which 
it undergoes. Ramsay and Soddy l made an investigation of 
the volume of the emanation, with a view to determining 
whether or not the substance remained constant. The results 
which they obtained are given in the following table : 

Time in days. Volume in c.c. 

Start 0-124 

1 0-027 

3 O'Oll 

4 0-0095 

6 0-0063 

7 0-0050 

9 . 0-0041 

11 . . 0-0020 

12 0-0011 

28 0-0004 

The first number seems very large in comparison with the 
others; this may possibly be due to an uncondensable gas 
being present and forced into the walls of the tube. 

1 Kamsay and Soddy, Zeitsch. physikal. Chem., 1904, 48, 691. 


From these results it is clear that the emanation is gradually 
disappearing. Now, we need not suppose that matter is being 
destroyed, but rather that something akin to the condensation 
of a gas to the liquid state is taking place, which will, of course, 
be accompanied by a contraction in volume. The simplest 
hypothesis is that the gaseous emanation which was derived 
from solid radium is undergoing a further change which is 
reconverting it into a solid substance. 

Here we touch another line of evidence tending to prove 
the same point, and we must turn aside to consider the phe- 
nomena which are classed under the heading excited activity. 

It was shown by M. and Mine. Curie l in the case of radium, 
and by Kutherford independently 2 in the case of thorium, that 
these substances have the faculty, when placed near other 
bodies, of communicating to the latter the power of exhibiting 
the phenomena of radioactivity. Debierne 3 later found that 
actinium had a similar property. 

We need not describe the experiments which have estab- 
lished the laws governing this phenomenon, but we may 
summarize the results which have been obtained. It is found 
that the strength of the excited activity depends, not upon the 
nature of the object upon which it is located, but purely upon 
the strength of the activity of the exciting radioactive prepara- 
tion, and the length of time that it was left in the neighbour- 
hood of the excited object. After the radioactive substance has 
been removed from the neighbourhood of the excited object, 
the latter begins to lose its radioactive properties, and the 
decay of these follows an exponential curve; Further research 
proved that the excited activity is proportional to the amount 
of emanation present. 

Now, Eutherford 4 has shown that if a platinum wire is 
exposed to thorium emanation it becomes endowed with excited 
activity. If we immerse a wire so treated in hot water, very 
little change can be detected in the activity when it is with- 
drawn and dried ; but if we immerse the wire in concentrated 
hydrochloric acid, we shall find that the activity is lost by the 

1 M. and Mme. Curie, Compt. rend., 1899, 129, 714. 

2 Rutherford, Radioactivity, p. 295. 

3 Debierne, Compt. rend., 1904, 138, 411. 

4 Eutherford, Phil. Mag. (6), 1900, 49, 188. 


wire, but is acquired by the solution ; and, further, if we 
evaporate the solution to dry ness, we shall find that the activity 
has been transferred to the dish. This active matter can be 
removed from the wire or the dish by simple scraping. 

All this goes to show that we are dealing now with a solid 
substance, and not with a gas. But if we have a solid sub- 
stance, it might be supposed to be a compound of the emanation 
with the platinum of the wire. This view is quite untenable, 
when we consider that red-hot platinum black will not attack 
the emanation ; so we are driven to conclude that the emana- 
tion has deposited the solid substance upon the platinum. 

Thus the results of measurements of diminution in the 
volume of the emanation, as well as those phenomena which we 
have just described, point alike to the view that the emanations 
of radium and thorium are continually being transformed into 
solid substances. 

Further investigation of the rate of decay in the case of 
excited activity showed that this solid active deposit (as Euther- 
ford terms it) was not a permanent substance, but was one 
which rapidly passed through a series of changes. Instead of 
the decay being expressible by the usual exponential equation 

I, = I X e~" 

it is found that it really follows the exponential law, but that 
the total decay series is made up of different factors. The first 
decay period has one value for A, the second has another value 
for A, and the third yet another. This proves, of course, that 
the primary radioactive deposit is first converted into a second 
one, having a different radioactive capacity; and that this 
second one is in turn changed into a third substance which 
has a decay constant different from that of its predecessors. 

In this way the existence of the radium derivatives A, B, C, 
D, E, and F has been established, and it has been possible to 
compare them with other bodies by a comparison of the decay 
constants of the two substances. The chief point of interest in 
these substances lies in the fact that radium D, E, and F are 
found in radio-lead, which appears to owe its activity to their 
presence ; while polonium appears to have constants agreeing 
with those of radium F. 


3. The Production of Helium. 

Though the measurements of radioactive decay mentioned 
in the last section were sufficient to show that the emanation 
from radium was capable of undergoing transformation into 
other forms of matter, the results were not quite convincing, 
for, to some extent, the proof depended upon certain theoretical 
assumptions which were incapable of rigid demonstration. 
The quantities of the transformation products were extremely 
small ; and no chemical or spectroscopic experiments could be 
made which threw much light upon the differences between 
the various substances. It was not until Ramsay and Soddy * 
began their work upon the emanation that a product was 
actually observed whose chemical individuality was beyond 

Rutherford and Soddy, 2 after finding that niton was an inert 
gas, put forward the view that it belonged to the argon family ; 
and they further pointed out the fact that helium is always 
found in minerals which contain uranium or thorium. The 
question as to whether this association of helium with the 
radioactive minerals had any connection with their activity was 
thus opened. 

Ramsay and Soddy took 20 mg. of radium bromide which 
had been prepared three months previous to their experiments, 
dissolved it in water, and collected the gas which was evolved. 
This gas was for the most part electrolytic gas, which had 
been produced by the action of the radium and radium ema- 
nation upon the water; but it also contained some emanation. 
To separate the latter from the other gases, the mixture was 
passed over a red-hot, partly oxidized copper spiral, and 
the water so formed was removed by means of phosphorus 
pentoxide. After this the gas was passed into a tiny vacuum 
tube, in which, the spectrum was examined and carbon dioxide 
detected. This gas was eliminated by means of liquid air, 
and a re-examination of the spectrum in the small vacuum 
tube showed the presence of helium, the D 3 line being visible. 
Further experiments were made, and practically all the lines 

1 Ramsay and Soddy, Proc. Boy. Soc., 1903, 72, 206 ; 1904, 73, 346. 
8 Rutherford and Soddy, Phil Mag. (6), 1902, 4, 581. 


in the helium spectrum were found. This work has been con- 
firmed by several workers, 1 and it has been shown that actinium 
also gives rise to helium. 2 

In this way, it was proved conclusively that radium emana- 
tion actually gives rise to helium. The objection might be 
made that the helium is present throughout the course of the 
experiments ; but this is shown to be untenable by the fact 
that the helium spectrum is not visible at first. When an 
examination is made at the beginning of the process, it is found 
that a new spectrum is visible which does not contain the 
helium lines : the latter develop slowly and are visible only 
some days after the emanation has been brought into the 
vacuum tube. For instance, in one experiment the emanation 
was led into the vacuum tube on July 17 ; the new spectrum 
probably that of the emanation was then observed, which 
contained no helium lines. After standing until the 21st, the 
helium spectrum was observed, and compared with that of a 
helium vacuum tube. 

It is thus shown that the helium is not present in the 
earlier part of the experiments, nor is it in any way connected 
with the presence of the radium salt in the solution : it is 
derived from the emanation alone, In this way the disintegra- 
tion hypothesis has obtained its strongest support. Previous 
to the work of Eamsay and Soddy, the evidence in 'favour of 
this hypothesis depended, to a great extent, upon postulates 
which could not be experimentally tested ; but by the pro- 
duction of helium from the emanation, the break-down of one 
radioactive substance into a non-active body was conclusively 

These experiments throw light upon another point. It has 
been found that helium exists in many minerals, as well as in 
the waters of several mineral springs ; and for a considerable 
time the presence of this very rare gas under such conditions 
was inexplicable. Kamsay and Soddy's researches have cleared 
up this question also ; for it has been shown that both minerals 
and mineral waters which contain helium have also more or 
less well-marked radioactive properties. They are therefore 

1 Dewar and Curie, Compt. rend., 1904, 138, 190; Meyer and Himstedt, 
Ann. d. Physik., 1904, 15, 184. 

2 Debierne, Compt. rend., 1905, 141, 383; Giesel, Ber., 1907, 40, 3011. 


certain to contain a small proportion of radioactive emanation, 
and it is doubtless from this source that helium is derived. 

4. The Disintegration Theory and the Age of Minerals. 

If a mass of uranium were placed in a hermetically sealed 
vessel and allowed to disintegrate under these conditions, it is 
evident that either the quantity of helium or the amount of 
lead which collected in the vessel would give a measure of the 
period during which the uranium has been enclosed. Calcula- 
tion shows that one gramme of uranium in radioactive equili- 
brium with all its disintegration products will give rise to 
about 11 X 10- 5 cubic millimetres of helium; so that from the 
ratio of the quantities of helium and uranium present we can 
readily ascertain how long the process has been going on. A 
similar calculation will be possible with regard to the uranium - 
lead ratio. 

Now in certain minerals uranium is present; and the 
helium given off by it and its disintegration products remains 
occluded in the solid material instead of being lost to the air. 
By breaking up the mineral, liberating and estimating the 
helium and comparing its mass with that of the uranium 
present, we can arrive at a rough estimate of the period during 
which the uranium has been confined in the rock under these 
conditions in other words, since the rock solidified. Strutt 1 
has investigated the matter and has been able to put forward 
estimates of the duration of the geological periods based on his 
results. It must be borne in mind that these estimates are 
minimum ones, since it is most improbable that all the helium 
generated is retained by the rocks. 

The uranium-lead ratio in minerals is open to even graver 
criticism when it is taken as a guide to geological age. It can 
be calculated that one gramme of uranium produces one gramme 
of lead in 7,500 million years ; but this is a very rough approxi- 
mation. Further, the value of the method depends upon the 
assumption that during that time no lead has been introduced 
into the mineral from external sources, which can never be 
certainly established. 

1 Strutt, Proc. Roy. Soc., A, 1908, 81, 272 ; 1910, 83, 298 ; 1911, 84, 379. 


It will be seen that both methods can lead only to approxima- 
tions to the truth ; but they are none the less valuable, in that 
they give us definite evidence on a period which has hitherto 
been even more vaguely calculated. The following table ] 
gives some of the main results obtained. The figures represent 
millions of years which have elapsed since the strata were 
laid down. Under the heading He/U are given the results 
derived from the helium-uranium ratios ; while the column 
headed Pb/U shows the figures deduced from the lead-uranium 
ratio in minerals. The numbers in the case of the helium 
ratios show the minimum possible age of the rocks ; but the real 
age is probably much greater. 

Strata He/U Pb/U 

Miocene 6'3 30 

Eocene 31 70 

Carboniferous 146 330 

Devonian 145 390? 

Archaean 405-715 940-1580 

It will be seen that the values differ considerably among 
themselves, as is only to be expected in view of the difficulties 
of the problem; but they provide us with a much more 
accurate gauge than do the previous estimates based on time 
required for the deposition of the strata. 

1 A fuller table is given by Holmes, Discovery, 1920, 1, 108. 



1. Introductory. 

THE problem of the transmutation of the elements is as old as 
the days of the alchemists ; and from the modern standpoint 
it presents a certain interest of its own. Two distinct questions 
are involved : (1) Is the conversion of one element into 
another possible? and (2) can this change be accomplished 
voluntarily in the laboratory with the means at our disposal ? 

The progress of radiochemistry has given us the answer to 
the first question ; for the disintegration of the radio-elements 
has been studied so fully that no doubt can now remain on 
the point. Certain elements do break up spontaneously and 
in their disruption they give rise to other materials the ele- 
mental nature of which cannot be denied. But since the 
radioactive processes are entirely beyond our control, the second 
question still remains unsettled. 

From the practical standpoint, the transmutation question 
is approachable on two sides : the synthetic and the analytic. 
We may attempt to build up complicated atoms from two or 
more simpler ones ; or we may endeavour to copy the radio- 
active processes and break up complex atoms into simpler 
materials. In the following sections, an account will be given 
of the researches with appear to bear upon the subject ; but it 
must be borne in mind that the question is still an open one. 

2. The Work of Eamsay and Cameron. 

It has already been mentioned that either radium or radium 
emanation, when brought into contact with water, decomposes 



the latter just as an electric current does. Apparently this 
action is confined to the liquid state, for steam does not appear 
to be affected in the same manner. 1 Hydrochloric acid is broken 
down into hydrogen and chlorine under the same circumstances. 
This parallel between the actions of the emanation and the 
electric current led Eamsay 2 to try the effect of niton upon 
a solution of copper sulphate, from which, by analogy, he 
expected to get copper deposited. When this experiment was 
carried out, Kamsay detected no helium in his apparatus, but 
observed the presence of neon and argon. Traces of lithium 
were found in the copper salt solution after the experiment was 
completed, though no lithium was found at the end of a blank 
experiment in which no niton was employed. 

Further results were published later. 3 In every case, 
lithium was found when niton had been employed, whilst no 
lithium was detected in exactly similar experiments wherein 
no niton was brought into contact with the solution of the 
copper salt. 

Again, 4 when a large quantity of thorium nitrate was left 
sealed up in a bulb for several months, it was found that 
a certain amount of carbon dioxide was detected among the 
gases evolved. The action of niton upon thorium nitrate 
solution resulted in the detection of carbon dioxide in this case 
also ; and similar results were observed in the cases of zirconium 
nitrate and hydrofluosilicic acid. 

Eamsay suggested that the action of the niton had broken 
down the complex atom of copper into lithium, which is the 
lightest element in the same family ; and that in a similar 
manner the atoms of thorium, zirconium, and silicon had been 
degraded into carbon. 

A repetition 5 of the lithium-copper experiment by Mme. 
Curie and Mdlle. Gleditsch led to negative results ; and 
Eutherf ord and Eoyds 6 failed to detect the presence of neon. 

Cameron and Kamsay, Trans., 1908, 93, 966. 
Ramsay, Nature, 1907, 76, 269. 

Cameron and Ramsay, Trans., 1907, 91, 1593 ; 1908, 93, 992. 
Ramsay, Trans., 1909, 95, 624 ; Ramsay and Usher, Ber., 1909, 42, 2930. 
Curie and Gleditsch, Compt. rend., 1908, 147, 345; cf. Perman, Trans., 
1908, 93, 1775. 

6 Rutherford and Royds, Phil. Mag., 1908, 16, 812. 


3. The Investigations of Collie and Patterson. 

An account of experiments in transmutation would be 
incomplete without some mention of the following results ; 
but it must be noted that the experimenters themselves have 
never described their results as due to the transmuting of 
hydrogen into other elements. They have confined themselves 
in their papers to a description of their experimental work and 
have put forward no theory on the subject. 

Collie and H. S. Patterson 1 observed that when pure cal- 
cium fluoride was bombarded with cathode rays a mixture 
of gases was evolved. This mixture consisted chiefly of 
hydrogen, oxygen, and carbon monoxide ; but when sparks 
were passed through it and the residue was condensed by 
means of charcoal and liquid air, a small quantity of neon was 
detected. The same result was obtained when carefully puri- 
fied glass wool was substituted for the calcium fluoride. 

There are obviously only two possible sources of this neon. 
It must either have leaked into the tube from outside or it 
must have originated in the interior of the tube. 

With regard to the leakage possibility, various tests were 
tried. Of these, the crucial one was as follows. The dis- 
charge tube was sealed inside another glass tube which was 
filled with neon at about half an atmosphere pressure ; hydro- 
gen at low pressure was then admitted into the inner tube, 
through which the cathode discharge was passed.* After the 
experiment, the hydrogen was pumped out of the inner tube 
and on examination it was found admixed with just the same 
quantity of neon as in previous cases. Had there been any 
leak in the discharge tube ; or had the glass of the discharge 
tube become permeable to neon under the action of the cathode 
ray, then a much greater quantity of neon might have been 
expected to make its way through the intervening wall in the 
second experiment. In the first experiment the pressure of 
neon outside the vessel is about 1/55,000 of an atmosphere, 
whilst in the second experiment with the enclosed tube the 
neon surrounding the discharge tube was at a pressure more 
than twenty thousand times as great. If a leak were in 

1 Collie and H. S. Patterson, Trans., 1913, 103, 264. 
* No calcium fluoride or glass wool was used in this experiment. 


question, the amount of neon found in the tube during the 
second experiment would be far greater than that observed 
with the unjacketed tube; and the practical coincidence of 
the results in the two cases appears to exclude leakage as a 
possible explanation of the phenomena. Further, if neon had 
leaked in from the air during the first experiment, a corre- 
sponding quantity of argon would also have found its way 
into the discharge tube ; and as there are about 500 parts of 
argon to 1 part of neon in the atmosphere, the argon spec- 
trum could not possibly have escaped detection had any air 
penetrated into the cathode tube. 

It must therefore be taken as proved that the neon 
originates within the discharge tube ; and the problem of its 
origin is narrowed down. Only three sources can be suggested : 
the glass of the tube ; the metal of the electrodes ; and the 
hydrogen which the discharge tube contains. 

It was proved that when the glass was finely powdered, 
washed with chromic acid, and heated in an exhausted tube 
until it fused, no neon was evolved. A similar negative result 
was obtained when the aluminium electrodes were fused in 
a vacuum : hydrogen was evolved from the metal but no neon 
made its appearance. As to the hydrogen, it contained no neon 
before the experiment. 

A further series of experiments led to even more surprising 
results. The jacketed tube was employed ; and in this case 
the external tube was exhausted. After the experiment, the 
usual amount of neon was found in the inner tube. When 
oxygen was admitted into the outer jacket after the experi- 
ment and pumped out again it was found that sparking of 
this oxygen produced a slight explosion, pointing to the 
presence of hydrogen which must have been driven through 
the intervening wall between the two tubes by the violence of 
the cathode discharge. The mixture from the outer jacket 
was then treated with charcoal and liquid air in the usual 
way; and it was found that the uncondensable residue con- 
tained chiefly helium mixed with a small quantity of neon. 
When the discharge experiment was repeated with oxygen in 
the jacket, the usual quantity of neon was detected in the 
inner tube, at the end of the experiment; but the oxygen 
was then found to contain neon also, mixed with a little helium. 


Eepetitions of the experiment under varying conditions showed 
that when the outer tube was completely exhausted before the 
discharge passed in the inner tube, helium predominated over 
neon in the outer tube after the discharge; whereas when 
oxygen at pressures up to 15 mm. was present in the outer tube, 
neon was the main new product of the discharge and helium 
was subsidiary. 

Finally, 1 Collie and Patterson showed that the presence of 
electrodes in the tube is not necessary for the production of 
the phenomena ; for similar results are obtained when a power- 
ful oscillating discharge is passed through a coil of wire 
wrapped round the outside of an electrodeless tube containing 
hydrogen at low pressure. Considerable quantities of hydro- 
gen disappeared from the interior of the tube when the current 
was running. A gas made its appearance in the tube and 
showed the carbon spectrum when examined. It may possibly 
be X 3 .* 

Soddy and Mackenzie, 2 Strutt, 3 Merton, 4 and Egerton 5 were 
unable to obtain the results of Collie and Patterson; though 
Masson 6 and Thomson 7 confirmed them. Collie, 8 using Mer- 
ton's actual apparatus, bombarded uranium powder in an 
atmosphere of hydrogen, and showed that both helium and 
neon were produced. The production of the former is perhaps 
not astonishing ; though no helium was observed when the 
uranium was raised to a red-heat in vacuo for an hour : but 
the production of the neon cannot be ascribed to any ordi- 
nary radioactive decomposition of the uranium. 

One factor which appears to exert considerable influence is 
the nature of the interrupter used on the coil generating the 
current. With a 10-inch coil and platinum break, positive 
results were obtained; whilst only negative results were 

* See p. 149. 

1 Collie and H. S. Patterson, Proc., 1913, 29, 217. 

2 Soddy and Mackenzie, Proc. Roy. Soc., 1908, A, 80, 92. (This paper, as 
can be seen from the date, is antecedent to the work of Collie and Patterson.) 

3 Strutt, Proc. Roy. Soc., 1914, A, 89, 499. 

4 Merton, Proc. Roy. /Soc., 1914, A, 90, 549. 

5 Egerton, Proc. Roy. Soc., 1915, A, 91, 180. 
Masson, Proc., 1913, 29, 233. 

7 Thomson, Nature, 1913, 90, 645 ; Rays of Positive Electricity t p. 122 

8 Collie, Proc. Roy. /Soc., 1914, A, 90, 554. 


observed during the use of a larger coil with a mercury 

A summary of the evidence is to be found in a paper by 
Collie, Patterson, and Masson. 1 

4. Rutherford's Experiments. 

When a-particles from a radioactive source are allowed to pass 
through an atmosphere of hydrogen it is found that scintilla- 
tions can be observed on a fluorescent screen which is placed 
beyond the normal extreme range of the a-particles. This im- 
plies that under these conditions something is striking the 
screen; and since the screen and the radioactive source are 
separated by a distance which no a-particle can span, it is 
evident that this " something " cannot be an a -particle. The 
explanation of the phenomenon 2 appears to be as follows. To 
take an extreme case, let us imagine that an a-particle meets a 
hydrogen atom " end-on " and that a collision occurs. Since 
the hydrogen atom is four times lighter than the a-particle, it 
is evident that it will recoil after the collision at a speed much 
greater than 'that possessed by the a-particle ; and being 
endowed with this greater velocity it will be able to travel 
through a longer distance than the original a-particle would 
have done before coming to rest at the end of its normal range. 
It is therefore assumed that the scintillations on the screen are 
produced by these recoiling hydrogen atoms, which are termed 

Eutherford has examined the phenomenon in the case of 
gases other than hydrogen, and in the case of nitrogen 3 he 
obtained an anomalous effect. When a stream of a-particles 
was passed through chemically pure nitrogen, it was found that 
the range of the recoiling atoms, instead of being small as 
might have been expected when the relative masses of the 
a-particle and the nitrogen atom are compared was compar- 
able to that which would have been produced if hydrogen had 
been present instead of nitrogen. In fact, the penetrating 

1 Collie, Patterson, and Masson, Proc. Roy. Soc., 1914, A, 91, 30. 

2 Marsden, Phil. Mag., 1914, 27, 824; Marsden and Lantsberry, ibid., 
1915, 30, 240. 

3 Rutherford, Phil. Mag., 1919, 37, 581. Cf. Hinsberg, Chem. Zeit., 1920, 
44, 294. 


power of these new particles was found to be even greater than 
that of the H-particles from hydrogen. 

If these particles should eventually be proved to be H- 
particles, there seems to be no escape from the idea that the 
nitrogen atom has been shattered by the impact of the a-particle 
upon it ; and that hydrogen is one of the disintegration-products. 
In support of this view, it is urged that the atomic weight 
of nitrogen corresponds almost exactly to the united weights of 
three helium and two hydrogen atoms ; so that the atom of 
nitrogen may be built up from three helium nuclei and two 
hydrogen nuclei. 

If such a disruption of the nitrogen atom takes place 
under the action of a-particles, it might be suggested that traces 
of hydrogen should be detectable in a vessel filled with a 
mixture of niton and nitrogen or in the gases evolved from 
thorium nitrate. 

Should the Eutherford particles eventually prove to be 
hydrogen atoms, it will establish the correctness of Eamsay's 
view (which was rejected at the time by more than one expert 
in radioactive problems), that the energy of radioactive bom- 
bardment might suffice to disintegrate atoms which normally 
are stable. 



1. Soddy' s Theory of Isotopic Elements. 

IN 1910, Soddy 1 called attention to the fact that the evi- 
dence then accumulated proved that certain members of the 
radioactive group of elements exhibited a complete identity of 
chemical behaviour. Thus thorium-X, actinium-X, and radium 
were found to be chemically indistinguishable from one 
another ; ionium, radiothorium, radioactinium, and thorium 
were also endowed with identical properties so far as chemical 
reactions went ; and the three emanations could not be 
distinguished from each other by any chemical means. 

As Soddy indicated, we have here three groups of sub- 
stances in which each element is readily distinguishable from 
the rest by means of its radioactive constants ; but in which 
no member of a group can be isolated from its neighbours by 
the employment of purely chemical methods of separation. 

When the atomic weights of these substances are examined, 
it is found that they are not identical^ That of ionium is 
230, of thorium 232, and of radiothorium 228. It thus became 
clear that there could exist two or more elements with different 
radioactive constants and different atomic weights but with 
identical chemical properties. Soddy hazarded the conjecture 
which has since proved well-founded that these phenomena 
would prove to be the beginning of some embracing general- 
ization which would throw light, not only on radioactive pro- 
cesses, but on the elements in general and the Periodic Law. 

In 1911, Soddy carried his views to a further stage. 

1 Soddy, Chemical Society's Annual Reports, 1910, VII., 285; see also 
Soddy, The Chemistry of the Badio-ekments, 1st edition, I., 29 (1911). 



Chemical analysis, as consideration will show, is not neces- 
sarily a method of separating matter into homogeneous 
elements ; but is, instead, a mode of isolating from each other 
certain types of matter which differ in behaviour when treated 
with a given reagent. Hitherto it has been assumed that, 
when chemical analysis has done its best, the resulting sub- 
stances are entirely homogeneous or, as we say, elemental 
forms of matter. But are we justified in this conclusion? 
Suppose that we had a mixture of thorium and ionium. No 
chemical process with which we are acquainted at present will 
separate these two bodies from one another ; yet when we apply 
a totally new method of analysis radioactive analysis we 
find that we are not dealing with one substance but with a 
mixture of two. 

Chemical analysis, then, permits us to separate matter into 
a number of "types" which are homogeneous in chemical 
behaviour. Each " type " is recognizably different from other 
" types " by means of its chemical reactions ; but may or may 
not be homogeneous as regards other properties such as atomic 
weight or stability. 

To express this idea, Soddy coined the word isotopes. A 
group of two or more elements occupying the same place in the 
Periodic Table and being chemically identical and non-separable 
is defined as a group of isotopes ; and within a group the 
separate members are said to be isotopic with one another. 
Thus ionium, thorium, and radiothorium are isotopes; the 
three emanations are isotopic with one another; and meso- 
thorium-1 is an isotope of radium. 

This conception of isotopy is so alien to all the old-fashioned 
views of the elements that it deserves further elaboration here. 
It is natural to inquire whether isotopes are really inseparable 
and indistinguishable from one another ; and whether, if they 
cannot be entirely separated, they are not partially separable 
i.e. whether a series of mixtures of them in unequal quanti- 
ties cannot be obtained by the use of particular methods of 

The work of Paneth and von Hevesy * on radium-D and 
lead leads to the conclusion that ordinary modes of separation 

1 Paneth and von Hevesy, Sitzungsber. K. Akad. Wiss. Wien, 1913, 122, 
993 ; Monatsh., 1913, 34, 1393. 


are quite incapable of producing the desired result. Precipita- 
tion, distillation, adsorption, electrolysis of aqueous solutions 
or of the fused material, diffusion and dialysis all failed to 
yield any detectable result; and when the delicacy of radio- 
active measurements is considered, this seems to prove that 
very little doubt can remain on the point. 

The most severe test to which the theory of isotopes could 
be subjected was applied by Paneth and von Hevesy 1 in the 
following manner. In the first place, they showed that the 
decomposition potential at which radium-E is deposited on 
the cathode is altered in the same direction and to the 
same extent by the addition of the isotopic element bis- 
muth as it would be by the addition of the same number of 
ions of radiurn-E. Similar results were obtained in the case of 
thorium-B and its isotope, lead. In the second place, in the 
deposition of radium-E and of thorium-B at potentials below 
the decomposition voltage, the addition of isotopic elements 
prevents this decomposition though other elements are with- 
out influence. Finally, 2 when a galvanic chain was formed in 
which radium-D was one of the members, the potential of the 
chain was found to be practically the same as that of a similar 
chain wherein ordinary lead. was substituted for the radium-D. 
The voltage in the former case was - 0'884 volts ; whilst in the 
latter it was - 0'888 volts. 

With regard to the spectra of isotopes, it is too early to 
decide whether or not they are identical, though the evidence 
already at our disposal tends to show that no difference can be 
detected between them. 3 Thus an ionium-thorium mixture 
containing at least 10 per cent, of ionium showed not a single 
line which was not already known in the thorium spectrum. 
It is regrettable that the life periods of the emanations from 
thorium and actinium are so short, as they would yield decisive 
evidence when their spectra were compared with that of niton. 
It might at first sight be supposed that the isotopes furnish 
a case parallel to that of the elements of the rare earths ; but 
this is not so. In the rare earths the difficulty encountered 

1 Paneth and von Hevesy, Physikal Zeitsch., 1914, 15, 797. 

2 See also Klemensiewicz, Compt. rend., 1914, 158, 1889. 

3 Exner and Haschek, Sitzungsber. K. Akad. Wiss. Wien, 1912, 121, Ha.. 
1075 ; Kussell and Kossi, Proc. Roy. Soc., 1912, A, 87, 478. 


does not lie in separating one element from another but rather 
in discovering whether or not a separation has been attained. 
In the series of isotopes it would be very simple to ascertain 
the reality of a separation could it be achieved ; but up to the 
present no separation has been carried out. 

2. The, a-Ray Change. 

To Soddy also we owe the next step in advance in our 
knowledge of the rationale of radioactive change, 1 his publica- 
tion on the subject dating from 1911. 

When a helium atom is expelled by a radioactive element, 
it carries with it two positive charges ; and the disintegra- 
tion product is therefore characterized by an electrical charge 
two units less positive or more negative than the original 
element. At the same time, owing to the loss of four units of 
matter in the form of the helium atom, the atomic weight of 
the disintegration product is four less than that of the parent 

Now the loss of one positive charge by an atom may be 
regarded as equivalent to a diminution of one positive valency 
of that atom; which is the same as a shift of one column to 
the left in the Periodic Table. The loss of two positive charges 
would therefore correspond to shifting an element into the 
second column to the left in the Table. 

Taking the two changes together, Soddy enunciated the 
rule that the expulsion of an a-ray by an element results in 
the formation of a new element which has an atomic weight 
four units less than that of its parent and which is situated 
two columns to the left in the Periodic Arrangement. 

For example, radium has an atomic weight of 226 and it 
lies in Group II. of the Table. As a result of the a-ray change 
it passes into niton which has an atomic weight of 222 and 
which lies in Group two columns to the left of Group II. 
Again, niton in its turn ejects an a-particle and becomes con- 
verted into radium- A which has an atomic weight of 218 
and which occupies a position in Group VI.* of the Periodic 

1 Soddy, Chemistry of the Badio-elements, Part I., 30 (1911). 

* This example has been chosen to illustrate a particular point ; but it 
must be understood that our knowledge of the chemical character of radium-A 
came at a later date from Fleck's investigations, of which an account is given 
in the next section. 


Arrangement. This last change may be rather difficult to 
grasp, but possibly the following reasoning may make it clearer. 
Niton loses two positive charges (which are equivalent to two 
positive valencies) during the a-ray change. But niton starts 
by having no valency at all. The loss of two positive valencies 
is therefore simply equivalent to the gain of two negative 
valencies; so that the new element is characterized by two 
negative valencies ; and this brings it into the sulphur column 
of the Table. 

3. Fleck's Investigations of the Chemistry of the Radio- 

The next advance in our knowledge of the subject is due to 
the work of Fleck 1 which was carried out at the suggestion 
of Soddy. The object of the investigation was to find the 
common element which each radio-element most resembled 
and then to see whether, after the two had been mixed, any 
separation of the constituents was possible by chemical means. 
At the time when this investigation was begun, very little was 
known about the chemistry of the "active deposit" group; 
but Fleck's researches established the relationships between 
some sixteen radio-elements and also showed their kinship with 
certain non-radioactive elements. 

The difficulties of the research can hardly be over-estimated. 
In the first place, some of the substances examined could be 
obtained only in minute traces. Secondly, the average life of 
the elements was, in some cases, extremely brief. Thirdly, 
owing to the rapidity with which the substances became con- 
taminated with disintegration products, radioactivity measure- 
ments were complicated by subsidiary factors. 

It is impossible to give here a detailed account of all the 
various examinations made by Fleck ; but for the present pur- 
pose it will be sufficient to describe a single example, which 
will give some idea of the difficulties which had to be over- 
come. The case of thorium-B and thorium-C may be chosen. 

First as to the manner in which the two products can be 
identified. Thorium-B emits only /3-rays, whilst Thorium-C 
yields a-rays. If a complete separation of the two compounds 

1 Fleck, Trans., 1913, 103, 391, 1052, 


is effected, then the thorium-B will be inactive at first as far 
as a -ray measurements go but will show a gradually increasing 
a-radiation owing to the formation of thorium-C by disinte- 
gration. On the other hand, thorium-C will show a gradual 
falling off in a-radiation owing to its disintegration. 

Fleck found that when a small quantity of a lead salt was 
added to a solution containing thorium-B and thorium-C in 
radioactive equilibrium, complete precipitation of the lead as 
sulphide carried down both of the radio-elements with it. 
Incomplete precipitation led to a preponderance of thoriuin-C 
in the sulphide precipitate. When, in addition to the lead, 
another metal was present which gave a sulphide soluble in 
ammonium sulphide, no activity was found in the ammonium 
sulphide solution. This proves that neither thorium-B nor 
thorium-C is allied to the noble metals. 

A mixture of the two radioactive materials with salts of 
lead and copper was then tested. It was found that when the 
solution of the mixture was precipitated with sodium carbonate 
and the precipitate was digested with potassium cyanide solu- 
tion, no activity was found in the solution. This demonstrates 
that the radio-elements in question must be akin to either lead 
or bismuth, since these are the only two metals of the second 
analytical group which behave in this manner. 

A mixture of lead and bismuth salts with salts of the two 
radio-elements was then treated with sulphuric acid and alco- 
hol. The precipitated sulphate was removed and the bismuth 
was precipitated by means of sulphuretted hydrogen. Kadio- 
activity measurements proved that the lead sulphate precipitate 
contained thorium-B whilst the bismuth sulphide preciptate 
contained the thorium-C sulphide. This established the re- 
semblance between thorium-B and lead on the one hand and 
between thorium-C and bismuth on the other. 

Many other methods were tested in the attempt to separate 
thorium-B from admixed lead and thorium-C from admixed 
bismuth; and all led to the same conclusion. Among them 
may be mentioned the following : (1) Precipitation of lead as 
sulphate and of bismuth as sulphide ; (2) Precipitation of lead 
as sulphate and of bismuth as hydroxide ; (3) Precipitation of 
lead as chloride ; (4) Precipitation of lead as sulphide and of 
bismuth as oxychloride ; (5) Precipitation of lead as lead 


hydroxide and of bismuth by means of m-nitrobenzoic acid; 
(6) Precipitation of bismuth with alcohol; (7) Electrolytic 
separation; (8) Fractionation of lead by means of sulphuric 
acid and alcohol. 

It will be seen that these reactions prove conclusively that 
no separation of lead from thorium-B or of bismuth from 
thorium-C can be attained by the ordinary methods. 

This series of investigations by Fleck established the cor- 
rectness of Soddy's a-ray rule ; and also led on to the further 
generalization with which we are about to deal. 

4. The fi-Eay Change. 

The credit of throwing light upon the effect of j3-ray changes 
is shared by no less than four investigators who almost 
simultaneously and, possibly, independently had arrived at 
solutions of the problem. 1 It was pointed out by them that 
the expulsion of a )3-ray * from a radioactive element resulted 
in the formation of a disintegration product with a chemical 
character which placed it in the column to the right of that 
which is occupied by the parent substance. 

For instance, uranium-X 2 , which belongs to Group V., ejects 
a j3-ray and is converted into uranium-2, which falls into 
Group VI. ; radiuin-B, in Group IV., throws off a /3-ray and 
yields radium -C, which belongs to Group V. 

Thus the j3-ray change is different from the a-ray change 
both in nature and in degree. The a-ray change, in the case 
of an element in Group N, results in the production of a sub- 
stance belonging to Group (N - 2) ; whilst the /3-ray change 
gives rise to a new element which lies in Group (N + 1). 
Not only so, but the a-ray change results in a loss of four 
units of atomic weight; whilst the /3-ray change yields a new 
element with the same atomic weight as that of its parent. 

1 von Hevesy, Physical. Zettsch., 1913, 14, 49 (January 15) ; Kussell, 
Chem. News, 1913, 107, 49 (January 31) ; Soddy, ibid., 97 (February 28) ; 
Jahrb. Radioakliv. Elektronik, 1913, 10, 188; Fajans, Physikal. Zeitsch., 
1913, 14, 131, 136 (February 15). Of the four, only Soddy and Fajans pro- 
duced correct results. See footnote on p. 205. 

* " Ray less " changes are probably actually -ray changes in which the 
-ray is of a penetrating power so feeble as to defy identification at present. 


5. Soddy s Law* 

We are now in a position to fit together into a connected 
whole the three main facts with regard to the relations between 
the Periodic Table and the course of the disintegration of the 

Soddy's Law may be stated as follows. When a radio- 
element expels an a-particle, the disintegration product takes 
up a position two places from the parent substance in the 
direction of diminishing mass : so that the product is not in 
the next family but in the next but one; and the atomic 
weight of the product is four units lower than that of the 
parent substance. In a /3-ray change, the product retains the 
same atomic weight as the parent, but occupies a position in 
the Periodic Table in the column next the parent substance 
and on the side opposite to that on which the product of an 
a-ray change would lie. 

It is evident that the successive operations of an a-ray 
change and two j3-ray changes, in any order, will produce a 
final disintegration product which lies in the same column 
of the Periodic Table as the original parent did. Thus, for 
example, in the case of uranium we have an element lying in 
Group VI. It throws off an a-particle, producing uranium- X lt 
which belongs to Group IV. A ]3-ray change yields uranium- 
X 2 , a Group V. element ; and a further j3-ray change gives rise 
to uraniuin-2 which belongs to Group VI. Thus, starting with 
an element in Group VI., we have ended with a new element 
in the same Group though with an atomic weight four units 
less than that of the original uranium. Again, take the case 
of actinium-B. This element lies in the lead Group. It ejects 

* Hitherto there has been no name to distinguish this law and it seems 
desirable that there should be some concise mode of referring to it. The main 
credit in the matter is due to Soddy, who first grasped the meaning of the 
isotopes and the a-ray change. To him also is due the initiation of Fleck's 
researches which definitely established the validity of the law. Russell, when 
enunciating the rule for the -ray change, combined it with Soddy's previous 
a-ray rule ; but owing to the fact that he failed to regard the Periodic Table as 
continuous from Group I. through Group to Group VII., his statement of the 
law was somewhat imperfect. Soddy and Fajans practically simultaneously 
published the complete generalization ; but from the foregoing it will be seen 
that the lion's share of the credit belongs to the former investigator. Fleck's 
results, though unpublished, seem to have been known to those interested in 
the subject at the time. 


a /3-particle and yields actinium-C which belongs to the bis- 
muth Group. Actinium-C now expels an a-particle and pro- 
duces actinium-D, an element in the thallium Group. Finally, 
actinium-D throws out a j3-particle and passes into the end- 
product of the series. This end-product lies in the lead Group 
from which we started. 

An examination of Fig. 26, shown on p. 207, will make clear 
the effects of the two types of radioactive change. The Table 
is the original one drawn up by Soddy 1 in 1913 and therefore 
does not show * the connection between the uranium and actinium 
series which has already been discussed on a previous page. 
It will be noticed that Soddy's diagram expresses the fact that 
there is a unit difference of charge between successive places in 
the Table, thus anticipating Moseley's independent deduction 
drawn from the X-ray spectra of the elements. 

Inspection of the diagram will show that in certain columns 
two or more elements are to be found. For example, at the 
right-hand side in Group Via. we find uranium- 1 and urani- 
um-2, whilst in some of the other columns the numbers are even 
greater, reaching a maximum in Group IVb. The result of 
Fleck's investigations proved conclusively that elements which 
occupy the same place in the Periodic Table cannot be separated 
from one another by ordinary chemical means. This holds good 
whether these elements belong to the same radioactive series 
or not. For example, radio-actinium, thorium, and ionium, 
though each belongs to a different series, are chemically in- 
separable from one another. They belong, as has already been 
indicated, to the class named by Soddy the isotopes. If we 
mix thorium and ionium together we shall not be able to 
separate them again by any process of ordinary chemical 
analysis, though by applying radioactivity tests we could 
determine that more than one element was present. 

Soddy 2 has drawn attention to some of the consequences 
which follow from the recognition of isotopy. In the first 
place, the position occupied by an atom in the Periodic Table 
is not a mere function of its mass but depends mainly upon 

1 Soddy, Paper in Section B, of the British Association, 1913. 

2 Soddy, Chemistry of the Radio-elements, Part II., 6 (1914). 

* Soddy's latest diagram is to be found in his lecture to the Chemical 
Society (Trans., 1919, 115, 16). 



the electrical content of the atom and is only to a subsidiary 
extent influenced by atomic mass. From this it follows that 
chemical elements are not necessarily homogeneous but may be, 


as Crookes l suggested as far back at 1888, composed of atoms 

1 Crookes, Trans., 1888, 53, 487 ; see also Trans., 1889, 55, 257. The reader 
is advised to consult these papers. A mixture of two or more isotopes would 
correspond to Crookes' conception of a meta-element. 


of different atomic weights; and what we term the "atomic 
weight" may be merely an average value and not a true 
physical constant.* 

Secondly, Soddy pointed out that when we go further up 
the Table in the vertical columns we come to the ordinary 
families of elements. In these families the differences in mass 
between similar elements are much greater than in the case of 
the isotopes ; and instead of identity we get analogy in chemical 
character between them. From this comparison, Soddy draws 
the conclusion that the chemical nature of matter is a 
function of two variables : mass and electrical character. The 
electrical content is the essential variable in the horizontal 
rows: whilst in the vertical columns the mass plays the 
preponderant part. Mass also enters into the problem of 
the stability of elements, and it may in some cases play a pre- 
ponderant part from this standpoint. 

6. fi-Bays and Ionic Charges. 

When a radio-element undergoes two successive j3-ray 
changes, the ' net result is an increase of two units in its 
positive valency. Now an apparently similar process takes 
place when a divalent metallic ion becomes converted into a 
quadrivalent one ; and it is natural to inquire whether the 
two processes are parallel or identical in nature. At Soddy's 
suggestion l this problem was investigated by Fleck ; 2 and it 
was shown that the two cases are not identical but are merely 

An examination of the diagram on p. 207 will show that 
the quadrivalent element uranium-Xi loses two /3-particles and 
is converted into the hexavalent element uranium-2. Parallel 
to this we have the conversion of quadrivalent uranous salt, 
such as UC1 4 , into the hexavalent uranyl compound U0 2 C1 2 . 
In both cases the parent substance gains two positive charges 
as a net result. 

* It is noteworthy that in the same paper Crookes suggested that elec- 
tricity might be one of the ultimate elements, a view subsequently modernized 
by Ramsay in his paper on "The Electron as an Element" (Trans., 1908, 
93, 774). 

1 Soddy, Nature, 1913, 92, 399. 

2 Fleck, Trans., 1914, 100, 247. 


Now uranium-2 is isotopic with uranium-1 and is there- 
fore chemically inseparable from the latter. If the /3-ray 
change and the ordinary gain of two positive charges by an 
atom are identical processes, then the two starting substances 
also should be isotopic and inseparable by chemical means. 
In other words, a uranous salt and a salt of uranium-Xi ought 
to be chemically inseparable. 

It is not necessary to select uranium-Xi itself as the subject 
of the experiment ; any of the five isotopes shown in the table 
will do equally well ; and in practice thorium was chosen, since 
it is easily procurable. 

Great precaution had to be taken that the uranous salt 
was not converted into the uranyl derivative during the process 
of separation; but by the use of a special air-free apparatus 
this was accomplished. 

The results proved that when a mixture of uranous and 
thorium salts was fractionally precipitated, either by means 
of oxalic acid or by boiling off the excess of ammonium carbonate 
which was used to hold the salts in solution, a separation of 
the thorium from the uranium was readily obtained. 

This result shows that uranous uranium and thorium are 
not isotopic with one another; and hence the conversion of 
uranium-Xi into uranium-2 by the loss of two j3-particles is 
merely parallel to and not identical with the change which 
occurs when a uranous compound is converted into a uranyl 
derivative. Soddy deduces from these results that the electrons 
which are ejected during /3-ray changes come from the nucleus 
of the atom, whilst mere alterations in valency affect the outer 
ring of electrons; and further, that there is no interchange 
of electrons between the nucleus and the shell in chemical or 
electrochemical changes. 

7. The Atomic, Weight of Lead. ' 

An inspection of the chart on p. 207 will show that the 
end-products of the radioactive series, marked as black discs, 
all fall into Group IVb. of the Periodic System; and, since 
they only occupy one space in that Group, they are isotopic 
with one another. Closer inspection will reveal a still more 
interesting fact. 



If we take the atomic weight of radium as 226,* then the 
atomic weight of the product of radioactive change (passing 
through radium F) will be 206, since this series of changes 
involves the expulsion of five a-particles each of which has 
an atomic weight of four units. The end-product in the 
branch through radium-C 2 will have an atomic weight of 210, 
since only four a-particles are expelled during the changes. 

Again, starting with thorium, which has an atomic weight 
of 2 32 '4, we find a succession of six a-ray changes in both 
branches of the chain, so that in each case the atomic weight 
of the end-product must be 208*4. 

With regard to actinium we are not on quite such sure 
ground ; but the atomic weight of the end-product here must 
be either 208 or 212. 

Now the atomic weight of lead itself, as given in the 
International Table of Atomic "Weights, is 207;1 ; and the ques- 
tion at once arises : Is ordinary lead a homogeneous element 
or is it a mixture of two isotopes ? A mixture of the end- 
products from uranium (through radium-F) and from thorium 
in equal proportions would have exactly the atomic weight 
which is required; and since the material is not radioactive 
and its two constituents would be inseparable from each other 
by chemical means, the results of atomic weight determination 
would work out exactly as in practice. 

Soddy l threw out a hint that the examination of the atomic 
weight of lead derived from minerals containing uranium and 
no thorium might differ from the atomic weight of lead found 
in other minerals containing thorium but no uranium ; though 
he was careful to point out that the results might be obscured 
by other factors. 

This line of research was taken up by several investigators. 
The first results were published by Soddy and Hyman 2 who 
examined the lead derived from Ceylon thorite, a mineral 
containing 55 per cent, of thorium and 1 to 2 per cent, of 
uranium. The percentage of lead in the mineral is so small 
(0*4 per cent.) that it may be entirely of radioactive origin. 
Since the rate of change of uranium is two-and-a-half to three 

* Honigschmid's value is 225'95 (Monatsh., 1912, 33, 253). 
1 Soddy, Annual Reports, 1913, X., 269-70. 
Soddy and Hyman, Trans., 1914, 105, 1402, 


times that of thorium, the lead, if derived from both sources, 
should contain about ten parts of thorium-lead to one part of 
uranium-lead. On this calculation the atomic weight of lead 
from thorite ought to be rather higher than that of ordinary 
lead; and when the atomic weight was actually determined, 
it was found to be 207*694 * as compared with 207'2 for ordinary 
lead. 1 Photographs of the spectra of the two varieties of lead 
were found to be identical with the exception of one line, 
47601, which was much stronger in the spectrum of ordinary 
lead than in that of thorite lead. 2 

Eichards and Lembert 3 then published a series of deter- 
minations of the atomic weights of lead from five different 
radioactive minerals and from two commercial, non-radioactive 
specimens of lead. They found that lead occurring in urani- 
nite derived from North Carolina had an atomic weight of 
206*40 whilst common lead had an atomic weight of 207*15. 
The leads from other minerals took up positions intermediate 
between these extremes. Kichards and Wadsworth obtained 
the value of 206*08 for radioactive lead from Norwegian cleveite. 

Another series of determinations is due to M. Curie. 4 Lead 
from carnotite was found to have an atomic weight of 206'36, 
whilst the metal from galena gave an atomic weight of 

Finally, Honigschmid 5 determined the atomic weight of 
lead from Joachimsthal pitchblende and found it to be 

Taking the whole of the results, 6 it will be seen that the 
lowest figure is 206*08 and the highest is 207*694, common 

* This is the corrected figure ; see Soddy, Annual Reports, 1916, XIII., 247. 

1 Baxter and Grover, J. Amer. Chem. Soc. t 1915, 36, 356. 

2 See also Honigschmid and Horovitz, Monatsh., 1915, 36, 353; Merton, 
Proc. Boy. Soc., 1915, A, 91, 198; Siegbahn and Stenstrom, Compt. rend., 
1917, 165, 428 ; Harkins and Aronberg, Proc. Nat. Acad. Sci., 1917, 3, 710, for 
other data with regard to the identity of the spectra of isotopes. 

3 Kichards and Lembert, J. Amer. Chem. /Soc., 1914, 36, 1329; see also 
Eichards and Wadsworth, J. Amer. Chem. Soc., 1916, 38, 2613. 

4 Maurice Curie, Compt. rend., 1914, 158, 1676. 

5 Honigschmid and Mile. Horowitz, Compt. rend,, 1914, 158, 1796; 
Monatsh., 1915, 36, 353. 

6 See also Soddy, Nature, 1915, 94, 615 ; Engineering, 1915, May 28 and 
October 1; Richards and Wadsworth, J. Amer. Chem. Soc., 1916, 38, 221, 
1658, 2613. 


lead being 207*2. Since some of the results were obtained by 
investigators who have specialized in atomic weight determina- 
tions, no doubt can be entertained as to their accuracy; and 
it must be taken as : proved that the "atomic weight of 
lead " is not a physical constant, but differs according to the 
specimen chosen for examination. 

Density determinations of various specimens of lead 
yielded further results of interest. Soddy and Hyman, work- 
ing with the thorite lead mentioned above, found the density 
to be D 4 20 = 11*376 whereas the value for ordinary lead is 
11 '341 5. Eichards and Wads worth found the density of lead 
from Norwegian cleveite to be 11 '273. Other results were 
obtained by the same investigators. On calculating the atomic 
volumes for the various isotopic forms of lead from these 
measurements, ifc was found that all the isotopes had the same 
atomic volume. The melting-points of isotopes appear to be 
identical. 1 

The importance of the foregoing results requires but little 
emphasis. Clearly they strike at the very root of our ordinary 
chemical conceptions ; and, since they are found in the case of 
non-radioactive materials, they cannot be brushed aside as 
forming part of the abnormalities which characterize the 
radio-elements. They form the most conclusive evidence of 
the correctness of Soddy's theory of the isotopes from the 
purely chemical side, as the proof of the existence of different 
forms of lead can now be based entirely upon chemical in- 
vestigation, independent entirely of radioactivity measure- 
ments ; and with this advance we enter upon the field of the 
normal elements themselves. 

8. Other Cases of Isotopy. 

If it be assumed that isotopic elements are identical in 
chemical properties, but different in atomic mass, it is evident 
that the only method of separating one isotope from another 
will depend upon the employment of some process which is 
influenced by the mass factor. Positive ray analysis appears 
to supply the needed implement; since it is essentially a 
method of sifting heavy atoms from lighter companions. 

1 Lembert, Zeitsch, EleUrochem., 1920, 26, 59, 


When the positive ray method was applied to the fraction 
of liquid air which contains the heavier gases, parabolas were 
produced corresponding to neon, argon, krypton, and xenon ; 
which was just what might have been expected. The photo- 
graph taken with the lighter constituents, on the other hand, 
yielded something quite new. 1 In addition to the lines 
corresponding to helium and to neon with one and two charges, 
there was a line which pointed to the presence of an element 
having the atomic weight 22. This might be ascribed to a 
doubly-charged molecule of carbon dioxide (44 -f- 2) ; but when 
all the carbon dioxide had apparently* been removed from 
the gas this line still persisted. The line is fainter than the 
neon line, showing that the percentage of the " 22 " gas in the 
atmosphere must be less than that of neon. 

There is another possible explanation of the occurrence of 
the line. If neon combined with two atoms of hydrogen to 
form a compound H 2 Ne, the results would be explicable. 
There is, apparently, evidence which weakens this suggestion. 
Another line corresponding to the value 11 has been detected ; 
and this implies the presence of the same substance with a 
double charge. Now doubly-charged atoms are known ; but 
up to the present no molecule has been found to carry more 
than a single charge. Thus we have to choose between accept- 
ing an atom with a weight of 22 or a molecule of H^Ne which 
behaves differently from all other molecules hitherto examined. 

Thomson suggests that neon has a close ally, termed 
metaneon, with an atomic weight of 22. Attempts have been 
made to separate the two gases from one another by diffusion, 
and it is stated that the result was a separation of atmospheric 
neon into two fractions; but this has been contradicted by 
another investigator. 

The spectra of two diffusion fractions showed no abnor- 
mality ; but this may be explicable either on the assumption 
that the spectra of neon and metaneon are identical, or by 
assuming that the quantity of metaneon present was so small 
that its spectrum was swamped by the brilliant neon spectrum. 

1 Thomson, Rays of Positive Electricity, p. 112 (1913). 

* Compare, however, the observation of Collie and H. S. Patterson (Proc., 
1913, 29, 217) that carbon compounds are retained in vacuum tubes under 
conditions which normally might be supposed to preclude their existence. 


When the positive ray method is applied to the study of 
other gaseous elements, 1 equally interesting results are obtained. 
Helium and argon give parabolas which accord with the atomic 
weights 4 and 40; and no trace of subsidiary parabolas was 
obtained. Krypton, on the other hand, seems to yield no less than 
six parabolas which are ascribed to constituents having the atomic 
weights : 78, 80, 82, 83, 84, and 86. Xenon produces five para- 
bolas, which would correspond to the atomic weights : 128, 130, 
131, 133, and 135. Nitrogen appears to be composed of homo- 
geneous atoms ; but chlorine leaves parabolic traces on the plate 
which accord with the atomic weights : 35 and 37. The results 
in the case of mercury are said to support the view that this 
element contains atoms with weights of 202 and 204 as well as 
material which lies between 197 and 200. 

Thus, if the positive ray method be accepted, it seems 
evident that isotopy is not confined to the direct environment 
of the radioactive elements, but is a general phenomenon ; and 
these results appear to favour a return to a modified form of 
Prout's hypothesis, since they point to the possibility that all 
elements are mixtures of atoms from each other in atomic 
weight by complete units. 

9. Conclusion. 

Before closing this chapter, some account must be given of 
the bearing of the isotopic theory upon the general question 
of the elements. It is not claiming too much to state that 
the isotopes furnish us with a means of defining the difference 
between one element and another to a degree which ten years 
ago might have been thought impossible. Until the discovery 
of isotopy, it was assumed by the majority of chemists that 
chemical analysis did actually segregate matter into homo- 
geneous classes if it were carried far enough ; and those who 
held Crookes' ideas on the possibility of meta-elements were 
regarded as somewhat eccentric. It is true that in the field 
to which Crookes originally applied the meta-element view 
the rare earths the conception has not been substantiated ; 

1 Aston, Nature, 1919, 104, 393; 1920, 105, 8; Phil. Mag., 1920, 39, 449. 


but the isotopes are exactly in agreement with the meta-ele- 
ment definition. 

We now know that chemical methods of separation are 
at the best superficial. To use a crude analogy, the local 
surveyor may tell us that three houses are absolutely identical 
in every way ; but when we call in the local registrar he will 
be able to state that one house is inhabited by a bachelor, the 
next one by an old maid, whilst the third contains a family of 
seven. Chemical analysis is represented by the surveyor, 
whilst radioactivity measurements play the part of the regis- 
trar and tell us what the internal economy of the house 
may be. 

According to Soddy, 1 it may now be taken as proved that 
so long as the net charge on the atomic nucleus is the same, 
the element will show the definite chemical and physico- 
chemical character associated with one or other of the ninety- 
two places of the Periodic Table, quite independently of the 
nature and constitution of the nucleus. Its light spectrum 
and its X-ray spectrum will be identical with that of any 
other element fulfilling the same conditions of net charge. 
But since the net charge on the nucleus depends not on the 
absolute but only on the relative numbers of positive and 
negative charges associated with it, it is clear that two atoms 
may possess identical chemical properties and yet differ in 
mass. For example, one atom may contain three negative 
and five positive charges ; whilst another contains one negative 
and three positive charges. In both cases the net charge will 
be two positive ; and the chemical characters of the two atoms 
will be identical ; but their masses will be different, since it 
appears to be a function of the number of positive charges 

Even this does not entirely state the case ; for it is possible 
that two atoms may contain the same net nuclear charge and 
the same absolute number of opposite charges and yet these 
two atoms may differ from one another in the structure of their 
nuclei. Among the radio-elements cases of this kind actually 
occur. For example, in the following scheme, it will be seen 
that radium-C may disintegrate in either of two ways : 

1 Soddy, Annual Reports, 1916, XIII., 245. 


P At. Wt = 214\ * 

>^ x -^ 

Kadium-C' Kadium-C 2 

At. Wt. = 214 At. Wt. = 210 

^a 40 

Eadium-D End-product 

At. Wt. = 210 At. Wt. = 210 

In the one case, an a-particle is ejected first and then a /3-ray 
is subsequently thrown off; in the second case, the order is 
inverted and the expulsion of the j3-ray precedes that of the 
a-particle. The two end-products obviously have resulted from 
the same substance, radium-C, by the loss of exactly the same 
number of charges in each case ; but the two end-products are 
not identical : for radium-D continues to disintegrate, yielding 
radium-E, whilst the other element with atomic weight 210 
has apparently reached the end of its active disintegration. 

This example illustrates what might be called a finer degree 
of isotopy, since the two elements are alike in chemical cha- 
racter, spectrum, and atomic mass, and yet, as the difference in 
radioactivity shows, they are not identical. 

Mass and chemical properties alone, then, cannot be taken 
as criteria of elemental homogeneity. It is true that we cannot 
voluntarily disintegrate " chemical " elements ; but even at 
present we can " manufacture " them. If we mix together equi- 
molecular quantities of two forms of lead, one having an 
atomic weight of 20 6 '4 and the other with an atomic weight 
of 207*7, any chemist to whom we gave the mixture for analysis 
would find that the element lead was present and that its 
atomic weight was 207*05. Thus the mixture is a " chemical " 
element with an atomic weight which characterizes it ; and it 
cannot be separated by chemical means into anything simpler. 

Paneth 1 suggests that the name " element " should be con- 
fined to those substances which cannot be separated by chem- 
ical processes. All elements iso topic with one another would 
then be regarded as one " element." The light spectra and the 
X-ray spectra would thus still remain as criteria of chemical 
elements. When an element is composed of only one kind of 

1 Paneth, Zeitsch. physikal. Chem., 1916, 91, 171 ; ibid., 1918, 93, 86, 677 ; 
Wegscheider, ibid., 92, 741. 


atom, it would be termed a " pure " element ; whilst elements 
containing more than one type of atom would be regarded as 
" mixed " elements. Time alone can show whether this system 
of nomenclature is sufficiently elastic to cope with the addi- 
tions to our knowledge which are being made in the field of 

Some attempts 1 have been made to,devise methods whereby 
two isotopes might be separated from each other : but with the 
exception of a trial of centrifuging 2 these suggestions are 
theoretical only, and positive ray analysis is at present our only 
experimental means of sifting homogeneous atoms from a 
heterogeneous mixture. 

1 Lindemann and Aston, Phil Mag., 1919, 37, 523 ; Fleck, Nature, 1920, 
104, 565 ; Merton and Hartley, ibid., 105, 104 ; Chapman, ibid., 487 ; Soddy, 
ibid., 516. 

2 Joly and Poole, Phil. Mag., 1920, 39, 372. 



1. Isotopes and Isobar es. 

INVESTIGATIONS in the field of radioactivity have broadened 
our outlook upon many of the basic problems of chemistry; 
but, at the same time, it must be confessed that they have 
increased our difficulties in certain directions, especially in 
respect to the conception of elements. 

The following definition of an element appears to cover 
practically all the ground: "An element is a form of matter 
which has (1) a definite atomic weight ; (2) a definite spectrum, 
whether it be an emission or an absorption spectrum ; (3) a 
definite valency or electrical charge; (4) a definite series of 
chemical properties ; (5) a resistance to any ordinary chemical 
reaction such that no decomposition of the material into any- 
thing simpler is possible ; and, in certain cases, (6) a definite 
series of radioactive properties." If any form of matter agrees 
with all these criteria, we are justified in assuming that it is 
chemically elemental in character. 

Now let us return for a moment to some of the facts which 
were mentioned in the last chapter. It was there made clear 
that change in the radioactive elements may take place in 
either of two ways: (1) by the expulsion of an a-particle; 
or (2) by the ejection of a j3-ray. The alteration in properties 
which ensues when an a-particle is expelled results from two 
causes: a loss of atomic weight of four units and a lessening 
of the positive charge on the atom by two units ; so that the 
net result is to bring the product of the change into a position 
in the Periodic Table two places to the left of the parent 
element. The new substance thus produced conforms with all 
the criteria of an element which have been given above. Let 



us turn now to the j3-ray change. It results in the removal 
of one negative charge from the parent atom; so that the 
product, while retaining the same atomic weight, is shifted 
into the next column to the right in the Periodic Table. Here 
again the new form of matter produced is elemental in type. 
As we have already seen, an a-ray change followed by two 
successive /3-ray changes results in the production of an ele- 
ment whose atomic weight is four units less than that of the 
parent substance; and this new element occupies a position 
in the same column of the Periodic Table as its parent did. 

Let us look a little more closely into the matter; and 
choose as a concrete example the transformations which begin 
with thorium and end with radiothorium 

Group Ila. Group Ilia. Group IVa. 

esothorium-1 > Mesothorium-2 > Kadiothoriun 

change change 

228 ^.^ 228 228 

, Thorium 

An examination of the scheme above will show that we can 
distinguish between two classes of radioactive elements. 

In the first class are elements such as thorium and radio- 
thorium which can be distinguished from one another by 
their different radioactive properties, but which are inseparable 
from one another by purely chemical means. They differ from 
one another in atomic weight; but owing to the identity of 
their chemical properties they must be placed in the same 
space in the Periodic Table. On this account Soddy named 
them isotopes (from isos, equal, and topos, a place). 

In the second class, to which special attention must now 
be drawn, we may place elements such as mesothorium-1, 
mesothorium-2 and radiothorium. These elements are identical 
in atomic weight but differ from each other in chemical 
character; so that they are separable from one another by 
chemical reagents. Since their atomic weights are identical, 
the name isobares has been suggested for them (from isos equal, 
and baros, weight). Isobaric elements may be defined as 



elements which possess the same atomic weight but which 
differ from each other in chemical properties. 1 

Thus, as Soddy 2 has pointed out, the atoms of the elements 
can now be grouped under four heads 



Places in 


Heterobaric heterotopes . 



Lithium and chlorine 

Heterobaric isotopes 



Thorium and ionium 

Isobaric heterotopes 



Mesothorium-1 andmeso- 


Isobaric isotopes . 



Kadium-d, and radium-C 2 

So far, we have concerned ourselves principally with the 
radio-elements ; but the question now suggests itself : Are the 
isobaric elements confined to the radioactive group; or do we 
find any example of this class outside the last line in the 
Periodic Table ? If isobares exist apart from the radio-elements, 
they must fufil all the conditions which were laid down 
earlier in this chapter as tests for elemental character. 

Substances which fulfil the first five * conditions are actu- 
ally known ; and it is probably only ingrained conservatism 
which has hindered a recognition of them at an earlier stage. 
Ever since the days of alchemy, the transmutation of the ele- 
ments has been the most entrancing object of research ; but 
at the end of last century it had gradually become one of the 
dogmas of conservative chemists that transmutation was im- 
possible. Then came the discovery of radioactive changes ; 
and transmutation was actually found to be occurring continu- 
ally and spontaneously in one particular series of substances. 
Still, however, the idea of transmutation was confined to a 
process whereby the atomic weight of an element was altered. 
The old attempts to convert lead into gold had left their legacy 
behind them ; and the only change recognized as " transmuta- 
tion " was one in which an alteration in atomic weight oc- 

The researches which led to the formulation of Soddy's 

1 A. W. Stewart, Phil. Mag., 1918, 36, 326. 

2 Soddy, Trans., 1919, 115, 1. 

* Since they are not radioactive bodies the sixth criterion is not applicable 
in their case. 


Law have altered our ideas in this field. The conversion of, say, 
mesothorium-1 into raesothorium-2 is just as much a case of 
" transmutation " as the change of radium into niton is. It is 
now impossible to deny that " transmutation " may take either 
of two forms : (1) a change in atomic weight and (2) a change 
in chemical character unaccompanied by any alteration in 
atomic weight. Mesothorium-1 and mesothorium-2 are quite 
obviously different elements, yet each has the same atomic 
weight : 228. The fact that this type of transmutation among 
the radio-elements is spontaneous and irreversible has ap- 
parently blinded chemical investigators to some obvious 
deductions which can be drawn from the behaviour of better- 
known materials ; for no attempt has hitherto been made to 
apply analogous reasoning to elements outside the radioactive 

Let us select as an illustration the case of. ferrous and 
ferric iron ; and let us place the properties of the two " irons " 
in parallel columns, using the elemental criteria given above in 
each case 

Ferrous Iron. 

Ferric Iron. 

Atomic weight . 
Absorption spectra . 
Valency .... 
Chemical properties . 
Resistance to decomposition 

Different from ferric 
Resembles magnesium 

Different from ferrous 
Resembles aluminium 

An examination of the above table shows that ferrous and 
ferric iron fulfil all the criteria which we imposed in order to 
distinguish one element from its isobare. The atomic weights 
of the two bodies are identical, so they are isobaric with one 
another. The absorption spectra of their salts are different 
even to the naked eye. (It is impossible to determine whether 
or not their emission spectra are different, owing to the possi- 
bility of transmutation taking place under the abnormal con- 
ditions necessary to produce emission spectra.) Their valencies 
are different from each other, since ferrous iron unites with 
two chlorine atoms whilst ferric iron combines with three. In 
their chemical properties there is little resemblance between 
them : for ferrous iron in its reactions closely resembles 


magnesium ; whilst ferric iron is much more akin to aluminium. 
Neither can be broken down by ordinary chemical reactions 
into anything simpler. 

It is worth while to examine the parallelism which exists 
between ferrous and ferric iron on the one hand and, say, 
mesothorium-1 and mesothorium-2 on the other ; though it 
must be carefully borne in mind that the two cases are not 
identical.* None the less the closeness of the parallelism 
suggests that, although the origins of the phenomena are 
different, there is marked kinship between the operations of 
the two transmutational processes which are at work. 

Mesothorium-1 and ferrous iron both exhibit properties 
which are shown by elements in Group II. of the Periodic 
Table. Both form salts of the type EC1 2 , ES0 4 , etc. Both 
are convertible into a form of matter which possesses an extra 
positive charge mesothorium-2 in the one case and ferric iron 
in the other. On the other hand, mesothorium-2 and ferric 
iron are both trivalent ; and each of them forms salts of the 
type KC1 3 . 

The obvious difference between the two pairs lies in the 
fact that we can convert ferrous iron into ferric iron or vice 
versa at will ; whilst the conversion of mesothorium-1 into 
mesothorium-2 proceeds spontaneously and irreversibly as far 
as our present means go. In both cases, negative electric charges 
are removed from the divalent atoms ; but in the one case the 
charge is taken from a certain part of the atomic structure 
whilst in the other case it comes from a different portion of the 

Even the spontaneity of the /3-ray change finds its parallel 
among certain of the non-radioactive elements. Thus when 
the chloride of monovalent indium is dissolved in water it is 
spontaneously converted into metallic indium and the chloride 
of trivalent indium 

3InCl = InClg + 2In 

Eeduced to its essentials, this change corresponds to the loss 
of two negative charges from each of the monovalent indium 
atoms ; and no external forces are required in order to bring 

* For the evidence of this, see p. 208. 


about the phenomenon. The case of indium is by no means an 
isolated one ; for other elements exhibit a similar behaviour. * 

Yet a further parallelism between the /3-ray change and 
the conversion of an ion into one of a higher valency may be 
adduced here. In several cases, elements are found to exist in 
monovalent and trivalent forms, or in the divalent and quad- 
rivalent condition only, instead of forming a complete series of 
mono-, di-, tri-, and quadrivalent varieties. Thus thallium 
forms the chlorides T1C1 and T1C1 3 , but does not yield the 
intermediate T1C1 2 . Similarly, germanium forms GeCl 2 and 
GeCl 4 , but not GeCl 3 . It may be asked why these intermediate 
forms are not isolated as we remove electric charges from the 
substances of lower valency. 

An examination of the state of affairs in the case of the 
radio-elements may throw some light upon this point. The 
conversion of GeCl 2 into GeCl 4 is paralleled by two successive 
/3-ray changes in the radio-elements; and in the following 
table the results of such successive changes are given. The 
cases have been selected in which no disturbing factor in the 
shape of an alternative a-ray change occurs. The figures give 
the average life of the elements. 

-ray -ray 

Group x > Group (x + 1) > Group (x -f 2) 

change change 

Uranium Xi Uranium X 2 Uranium II 

35'5 days T65 minutes 3 x 10 6 years 

Mesothorium-1 Mesothorium-2 Eadiothorium 

7*9 years 8'9 hours 2'01 years 

Radium D Eadium E Eadium F 

24 years 7'20 days 196 days 

It will be seen that the intermediate product in the double 
j3-ray change has an average life very much shorter than those 
of the parent or the disintegration product. If the same 
reasoning holds good in the case of the salts of, say, thallium, 
we should expect to find that when monovalent thallium loses 
an electric charge and passes into divalent thallium, the latter 

* I am indebted to Dr. Smiles for calling my attention to the fact that 
a reaction of this type is the most general one in chemistry. It finds its 
parallel in organic chemistry in such reactions as the conversion of benzalde- 
hyde into benzoic acid and benzyl alcohol by treatment with alkali. 


substance loses a second electric charge readily and passes 
almost immediately into trivalent thallium, just as meso- 
thorium-2 soon breaks down to radiothorium. The parallel 
between the two cases is thus closer than appears at first sight. 

The question of isobarism may be considered from another 
point of view. Hittorf observed from E.M.F. measurements 
that chromium l existed in no less than three forms ; and he 
stated that " in the three conditions the metal exhibits differences 
of properties such as have been found only in different metals." 
In the inactive state, chromium behaves like a noble metal 
and stands at the end of the electrochemical series beside 
platinum. In the second state it ranks immediately behind 
zinc in the series. The third state brings it into a position 
intermediate between the two others. In the inactive con- 
dition it refuses to unite with iodine even when the latter 
element is nascent ; whilst in the active condition it is capable 
of displacing hydrogen from hydriodic acid. 2 

Again, just as two isotopes can be mixed in order to pro- 
duce a new substance having an atomic weight dependent upon 
the proportions in which the two components are present, so 
two forms of chromium can co-exist and give rise to a material 
which has an E.M.F. intermediate between those of the two 
forms when alone. 3 

For the sake of clearness, it may be well to summarize the 
matter in a few words. It is now proved by experimental 
evidence that two elements having totally different chemical 
and radioactive properties may have the same atomic weight. 
This is termed isobarism. Parallel to this but not identical 
with it, is the case of an element exhibiting two or more 
degrees of valency. This might be termed pseudo-isobarism. 
Between isobarism and pseudo-isobarism there are certain 
points of resemblance which appear to point to a parallelism 
between the two phenomena in their ultimate origins. 

1 Hittorf, Zeitsch. physical. Chem., 1898, 25, 729 ; 1899, 30, 481 ; 1900, 43, 
885. Compare also the cases of iron, Hittorf, ibid., 1900, 43, 385 ; Miiller and 
Koenigsberger, Physikal. Zeitsch., 1904,14,413,797; Miiller, Zeitsch. physical. 
Chem., 1904, 48, 585; Finkelstein, ibid., 1902, 39, 91; cobalt and nickel, 
Hittorf, ibid. t 1900, 43, 385. 

2 Hittorf, Zeitsch. physikal Chem., 1898, 25, 748. 

3 Hittorf, ibid., 1899, 30, 505. 


2. The Structure of the Atom. 

The problem of atomic structure has frequently been 
examined in recent years; and in this place only the barest 
outline of the subject can be given. For further information 
the reader is referred to other works. 1 

Kelvin 2 suggested a model atom in which negative par- 
ticles were embedded in a uniform sphere of positive electricity 
having the volume of the atom. This view has been further 
developed in detail by J. J. Thomson. 3 From the analogy 
of Mayer's Magnets,* Thomson has deduced that under the 
conditions of the hypothetical atom, the electrons would group 
themselves in certain ring-systems, the nature of the rings being 
governed by the number of electrons present. The periodicity 
of properties among the elements is explained on this basis by 
the periodic occurrence of similar rings in the arrangement as 
the system grows more complex. 

This view served to account for a certain number of the 
known facts with regard to chemical valency ; but some of 
Thomson's deductions failed to satisfy the ordinary chemist. 

Experimental work by Eutherford brought the matter into 
a new stage. When a stream of a-particles is allowed to 
pass through a thin sheet of some solid substance, it is 
found that some of them are stopped entirely or deflected 
through an angle from their original paths. This is called 
" scattering." Now calculation shows that the chance of the 
particles passing through the interstices between the molecules 
of the sheet is a very small one ; whence it must be deduced 
that these a-particles in many cases actually pass through the 
atoms which make up the sheet. On this view, the deflections 

1 J. J. Thomson, The Corpuscular Theory of Matter (1907) ; Eutherford, 
Radioactive Substances and their Radiations (1913); Kichardson, The Elec- 
tron Theory of Matter (1914) ; Millikan, The Electron (1917). 

2 Kelvin, Phil. Mag., 1902, 3, 257. 

3 J. J. Thomson, Phil Mag., 1904, 7, 237, and The Corpuscular Theory of 

* Mayer's Magnets are magnetized needles with cork floats which are so 
arranged that when the magnets are placed in water all the like poles float 
above the surface. When a large magnet is placed with its opposite pole over 
the centre of the vessel in which the needles float, the latter arrange them- 
selves into a series of fixed groupings, according to the number of magnets 
afloat at one time. 


observed experimentally must be brought about not by atomic 
collisions but by the action of electrical charges within the 
atoms of the sheet. 

Thomson l believes that the total deflection observed is 
actually the resultant of a series of small deviations brought 
about by the successive actions of a series of small charges 
within the atom. Kutherford, 2 on the other hand, holds that 
certain cases of this " scattering " can be explained only on the 
assumption that the deflections are due to the single operation 
of a large concentrated charge within the atom. The experi- 
mental evidence appears to favour Eutherford's view. 3 

In order to explain this " scattering " of a-particles, Kuther- 
ford 4 suggests that the atom consists of a concentrated positive 
charge surrounded by a zone of electrons sufficient in number 
to render the system as a whole neutral. For example, the 
uranium atom consists of a central positive charge surrounded 
by a number of concentric rings of negative electrons in rapid 
motion, the outer ring being comparable in diameter to the 
diameter of the atom. Practically the whole positive charge 
and mass of the atom is confined within a sphere of radius 
not greater than 10 12 c.m. The positively charged centre of 
the atom is supposed to be a complicated moving system 
consisting in part of charged helium and hydrogen atoms. 
The assumption is made that at very small distances the 
positively charged particles attract instead of repelling one 

Another suggestion has been put forward by Bohr 5 which 
also assumes that negative electrons are circulating in orbits 
around a positive nucleus. The pecularity of the Bohr atom, 

1 J. J. Thomson, Camb. Phil. Soc. Proc., 1910, 15, 465. 

2 Eutherford, Phil. Mag., 1911, 21, 669. 

3 Geiger and Marsden, Proc. Roy. Soc., 1909, A, 82, 495 ; Geiger, ibid., 1910, 
A, 83, 492. 

4 Kutherford, Radioactive Substances, 1913, p. 619 ff. 

5 Bohr, Phil. Mag., 1913, 26, 1, 476, 857; 1915, 29, 332; 30, 394. 

* This assumption might form the basis of an interesting speculation. 
Mass is always regarded as a " signless " quantity i.e. we do not take the 
possibility of negative mass into consideration. But under the abnormal con- 
ditions assumed by Rutherford, the possibility of negative mass suggests itself, 
and the atomic weight might be regarded as the algebraic sum of positive and 
negative masses within the atom. If the positive mass exactly equalled the 
negative mass, then something akin to the luminiferous ether would result. 


however, lies in the fact that the electrons are assumed to 
" jump " from one orbit to another ; and that during this 
"jump," radiation takes place. Bohr's calculations agree 
very well with spectral results in the case of hydrogen ; but 
they are unsatisfactory when applied to more complicated 
atoms. 1 

Bohr's atom depends upon a fusion of Newtonian dynamics 
with the quantum theory ; but Nicholson has avoided this by 
assuming that the vibrations causing spectral lines occur in 
a plane perpendicular to that in which the electrons revolve 
about the centre of the atom. This theory has received strong 
support from stellar spectroscopic evidence ; for lines predicted 
by Nicholson have been detected in the Lick Observatory 
photographs when special attention had been called to their 
position as a result of Nicholson's work, though they had 
escaped notice in previous examinations owing to their faintness. 

Nicholson's work has brought again into prominence the 
possibility of elements existing which are unknown upon the 
Earth and for which there are no places in our present Periodic 
Table unless, as van den Broek 2 suggests, they are isotopic 
with hydrogen. If this view be correct, then the isotopes of 
hydrogen would be the " proto -elements " : coroniuin, nebulium, 
and proto-fluorine. But if the spectra of two isotopes are 
identical, as seems to be the case, then " protofluorine " would 
have the same spectrum as fluorine, which is, of course, not 

Other attempts have been made to construct atoms by the 
combination of lighter elements ; 3 but they cannot be dealt 
with here. 

Most of the foregoing views have been based chiefly upon 
physical evidence ; but other attempts have been made which 
treat of the chemical side. 

According to G. N. Lewis, 4 every atom consists of a kernel 
and a shell. The kernel remains unaffected by chemical changes 

1 Nicholson, Phil. Mag., 1914, 27, 541 ; 28, 90. 

2 van den Broek, Phil. Mag., 1914, 28, 630. 

3 Nicholson, Phil. Mag., 1911, 22, 864 ; van den Broek, Physikal. Zeitsch., 
1911, 12, 490 ; 1916, 17, 260 ; Harkins and Wilson, /. Amer. Chem. Soc., 1915, 
37, 1396; Phil. Mag., 1915, 30, 723. 

4 G. N. Lewis, J. Amer. Chem. Soc., 1916, 38, 762 ; Science, 1917, 46, 297 ; 
cf. J. Amer. Chem. Soc., 1913,35,1448; Bray and Branch, ibid., 1913,35, 1440. 


and possesses an excess of positive charges numerically equal 

to the ordinal number of the Group in the Periodic Table to 

which the atom belongs. The number of negative electrons in 

the shell may vary during chemical change between and 8. 

The atom tends to hold an even number of electrons in its 

shell ; and especially to hold eight electrons, which are normally 

arranged at the eight corners of a cube. Lewis assumes that 

two atomic shells are mutually interpenetrable ; and that the 

electrons may pass with readiness from one position of the 

outer shell to another. The crux of his hypothesis is that one 

electron may simultaneously occupy the corners of two shells, 

if the corners are brought into superposition. By a further 

process of superposition a second pair of corners may come into 

coincidence ; and thus a second electron will become common to 

the two atoms represented by the cubes. In this way, single 

and double bonds are represented. Further, it is possible 

to distinguish between different types of bonds ; and thus 

differences in reactivity may be expressed. Finally, if it be 

assumed that the electrons in the shell are free to move along 

the edges of the cube towards the middle of the edge, an 

arrangement can be obtained which corresponds to the van't 

Hoff tetrahedric arrangement of the carbon atom. 

Noyes l has put forward a dynamic conception of the atom. 
He assumes that atoms are built up from two different sets of 
electrons: (1) those attached to a positive nucleus, and (2) 
valency electrons which are free to wander from atom to atom. 
When two atoms capable of interacting are brought together, 
Noyes suggests that a valency electron which is rotating round 
the positive nucleus of the first atom may find a positive 
nucleus of the second atom so close to it that it may proceed 
to describe an orbit round the positive nuclei of both atoms. 
During the portion of the orbit which lies within the second 
atom, that atom would become wholly negative whilst the first 
atom would be positive. During the other portion of the orbit 
each atom would be electrically neutral and the atoms might 
fall apart. In ionization, the electron would revolve entirely 
about the nucleus of the negative atom, leaving the other atom 
positive. If it be assumed that there are four, or eight, 
positive nuclei in the carbon atom, the hypothesis serves to 
1 Noyes, J. Amer. Chem. Soc., 1917, 39, 879. 


account for the tetrahedric grouping which is found in optically 
active carbon compounds. 

A more detailed hypothesis has been suggested by Stewart. 1 
On his view, the atom is assumed to be made up of three 
separate regions : (1) a core of negative electrons ; (2) an 
intermediate zone occupied normally by positive electrons but 
containing also, in the case of the radio-elements, certain 
negative electrons ; and, finally, (3) an external region occupied 
by negative electrons. The orbits of the electrons in the two 
inner zones are assumed to be approximately circular ; whilst 
those of the external electrons are supposed to be extremely 
elongated ellipses, similar to the paths of comets in the Solar 

The negative electrons in the core of the atom are assumed 
to be travelling at high velocities; and since charges moving 
at high speeds are difficult to deviate from their normal orbit 
by external forces, this serves to account for the fact that 
ordinary reactions fail to affect the intimate chemical structure 
of atoms. 

In the radio-elements, the core is assumed to be the point 
of origin of the /3-rays ; whilst the intermediate zone generates 
the d-particles. The atomic number corresponds to the sur- 
plus of positive over negative charges in the two inner zones 
together. 2 The radius of the positive zone is limited, by 
Eutherford's work, to approximately 10~ 12 cm. 

With regard to the conditions of attraction in the system, 
the negative charge in the core forms a centre around which 
the positive electrons can revolve ; whilst the two inner zones 
together, having a surplus positive charge, will serve to retain 
the " cometary " electrons in their orbits. 

The expulsion of the a-particle is assumed to result from 
the following process. In the intermediate positive zone of 
the radio-elements there are, in addition to the positive 
electrons, certain negative electrons ; and the two types of 
charge are linked into " sun-and-planet " systems of one 

1 A. W. Stewart, Phil. Mag., 1918, 36, 326. 

2 Cf. Soddy, The Chemistry of the Badio-elements, ii., 40 (1914). 

* It has been shown mathematically by Jackson, Phil. Mag., 1919, 38, 
256, that this system is unstable on the Newtonian assumptions. The exist- 
ence of Rutherford's positive nucleus would be equally impossible on the same 


negative and two positive charges which revolve as a whole 
around the central negative core of the atom. In the case of the 
uranium atom, which ejects eight a-particles in the course of 
its disintegrations, there will be at least sixteen such systems. 
When, by a crossing of orbits or by mutual disturbance, two 
such systems collide within the atom, a mass of four positive 
and two negative charges will be formed i.e. a helium atom 
with two extra positive charges ; and this mass will be ejected 
from the atom as a result of the collision, thus forming the 
a-particle. In this way the theory provides for the production 
of helium from the radio-elements without the necessity of 
postulating the actual presence of helium atoms, as such, 
within the atomic structure.* 

The chief feature of Stewart's atom, however, is the con- 
ception of the "cometary" orbits of the external electrons. 
These electrons are assumed to be the class involved in chemical 
changes. Now during alterations in valency by means of 
ordinary chemical reactions, two phenomena are noticeable : 
first, the comparative ease with which the change is brought 
about in certain cases ; and, second, the marked effect upon the 
chemical and physical character of the element which is pro- 
duced by the alteration in valency. A satisfactory model 
atom must account for both these facts \ and none of the 
" physical " atoms appear to throw much light upon the sub- 
ject. Stewart's atom seems to succeed at this point. 

When the electrons in the " cometary " paths are in 
aphelion to the centre of the atom, they will be travelling 
slowly in their orbits and they will also be only weakly 
attached to that centre. Thus at this particular region in 
their path they can be readily removed from or replaced in 
their orbits ; which accounts for the comparative ease with 
which changes of valency and chemical reactions take place. 
Secondly, when these electrons reach perihelion with regard 
to the centre of the system, they will pass very close to the 
positive zone ; and they will then affect the positive electrons 
in that zone from the outside just as the negative electrons of 
the core affect it from the inside. A change in the number 

* If the neon discovered in certain mineral springs eventually proves 
to be of radioactive origin, an extension of the above hypothesis would be 


of electrons in the " conietary " orbits will therefore influence 
the positive zone just as it would be affected by a change of 
the number of electrons in the core. The two effects are not, 
however, identical : for the influence of the core electrons is 
constant ; whereas that of the " cometary " electrons is tem- 
porary and periodic. Thus a change in the number of the 
" core " electrons produces a permanent effect and alters the 
atomic number ; whilst a change in the number of " cometary " 
electrons tends in the same direction, but lacks permanence. 
The latter influence is therefore sufficient only to produce a 
change in chemical character approaching to but not identical 
with a change in atomic number. For example, the conver- 
sion of ferric iron into ferrous iron yields a substance which 
has many of the properties of magnesium but which is not 
isotopic with that element. Similarly, the conversion of 
hexavalent into quadrivalent uranium produces a material 
very closely resembling thorium but not isotopic with it ; so 
that Stewart's atom furnishes a simple explanation of the work 
of Soddy and Fleck.* 

This model atom also suggests a solution of the rare earth 
problem, though it must be admitted that the matter is purely 
speculative. In order to account for the effects of the a- and 
|3-ray changes upon the atomic number, it is necessary to 
assume that this figure depends upon the algebraic sum of the 
negative and positive charges in the two inner regions of the 
atomic structure. The proper atomic numbers may be arrived 
at in several ways ; but for the present it will be sufficient to 
indicate two of these. In the first place, it may be assumed 
that all the members of a Series in the Periodic System have 
the same negative charge in their cores ; whilst the number of 
charges in the positive zone increases by single units as we 
pass up the series from element to element. A second arrange- 
ment giving the same atomic numbers would result if the core 
charges increase by single units from element to element ; 
whilst the corresponding positive charges increase in steps of 
two units. The charges in the " cometary " orbits will remain 
constant throughout the series if it be taken for granted that 
each unit increase in valency corresponds to the possibility of 
inserting an electron into the " cometary " orbit, an assumption 

* See p. 208. 



which corresponds to Kamsay's conception * of the electron as 
an element. 1 

The following figures (which are intended to be merely 
illustrative, though they may possibly correspond to reality), 
give an idea of the results obtained under the two assumptions 
given above : 

Case I. 








Electrons in core 








positive zone . 








" cometary " orbits . 








Case II. 

Electrons in core 








,, ,, positive zone . 








,, " cometary " orbits . 








Now, for the sake of simplicity, we may assume that the 
atoms of the ordinary elements which find places in the 
Periodic Table are constructed on the model of Case II. ; and 
it may be suggested that at lanthanum a change in atomic 
architecture takes place, so that the rare earth elements are 
constructed on the system indicated in Case I. When tantalum 
is reached, the structure of the atom reverts to the original 
type of Case II. and continues in this form up to uranium .t 

A further point of interest arises here. All the rare earth 
elements are trivalent J ; and this implies that as the positive 
charge in the atoms increases with the rise of the atomic 
number there must be a corresponding increase in the number 
of negative electrons in the " cometary " orbits in order to re- 
establish the proper valency, since the valency of the atom 
is assumed to be the algebraic sum of the total negative and 

* Ramsay regarded hydrogen as " hydrogen electride," metallic magnesium 
as "magnesium di-electride," metallic cobalt as "cobalt tri-electride," and so 

t Or possibly the A and B sub-groups in the Periodic Table are composed 
of differently constructed elements, whilst the rare earth group owes its peculiar 
character to the additional factor dealt with in the next paragraph. 

J Even cerium shows the trivalent character. 

1 Ramsay, Trans., 1908, 93, 774. 


positive charges in the system as a whole. This conception 
of the rare earth group, therefore, leads to the conclusion that 
the "cometary" orbits of the rare earths contain an extra 
set of electrons. From this, two results would follow. In 
the first place, the influence of these extra electrons in the 
" cometary " orbits would produce a " blurring " of properties 
throughout the group of elements, as was made clear above; 
and, in the second place, since these " cometary " electrons 
appear to be closely connected with absorption spectra (as 
can be seen from the cases of ferrous and ferric iron) it 
might be expected that the absorption spectra of the rare 
earths would display an unusual complexity in which again 
theory and practice accord with one another. 

With regard to isobares and pseudo-isobares Stewart's model 
atom furnishes a simple solution of the problem. When the 
change from one isobare to another involves the alteration of 
the atomic number, the electrons are ejected from the core of 
the atom and their loss permanently influences the numerical 
value of the surplus positive charge in the two inner regions. 
On the other hand, when pseudo-isobarism arises from a change 
in the " cometary " zone (as in the case of simple change of 
valency by an atom) then the effect upon the two inner zones 
is only temporary and periodic ; with the result that there is 
a change in general chemical character but no alteration in 
atomic number. It might simplify the matter if such isobares 
were regarded as " pseudo- elements " whilst the isobares 
arising from alterations in the core were treated as "true" 

Finally, Stewart's atom suggests a possible explanation of 
the instability of the radio-elements. The radius of the positive 
zone is not great when compared with the diameter of the 
negative electron ; and if the number of the latter electrons 
in the core be increased beyond a certain point there will be 
insufficient space for them to move freely within the positive 
zone. 1 When the atom reaches this degree of complexity, two 
things may happen to relieve the strain : either negative 
electrons may be forced out of the atom in the form of /3-rays ; 
or some of the negative electrons may be drawn into the 
positive zone to form the " sun-and-planet " arrangements 

1 Compare J. Q. Stewart, Science, 1917, 46, 568. 


which were postulated in order to account for the a-particles. 
Thus radioactivity could only be expected as a general rule 
among the elements of complex atomic structure and high 
atomic weight.* 

3. The, Problem of Atomic, Weight. 

An obvious rough relationship exists between atomic num- 
ber and atomic weight, since the latter is approximately double 
the former figure. Closer examination reveals that the ratio 
between the two values is not constant. Thus calcium has the 
atomic weight 40 and an atomic number 20, so that here the 
ratio is 2 : 1. Strontium has the atomic weight 87*6 and the 
atomic number 38, so that in this case the ratio is 2*31 : 1. In 
barium it rises to 2*45 : 1. Radium yields 2*57 : 1 ; and finally, 
in the case of uranium, the value is 2*59 : 1. 

Now apparently the atomic number corresponds to the 
surplus positive charge in the interior of the atom ; so that the 
foregoing figures really imply that the ratio of mass to surplus 
positive charge increases steadily as the series of elements is 

This can be explained quite easily if it be assumed that the 
mass of the atom is purely electrical. On this view, the atomic 
weight is simply the sum of the electrical masses of the various 
charges within the atomic structure. Now the mass of an 
electrical charge can be made to increase if the velocity with 
which the charge is moving be accelerated to near the velocity 
of light. If it be imagined that the positive charges in the 
calcium atom are revolving comparatively slowly in their orbits 
and that each charge corresponds to two units of atomic weight, 
then the charges in the radium atom, if they also were moving 
slowly, would produce an atomic weight of 88 X 2 = 176. 
Actually, however, the atomic weight of radium is 226, and to 
account for this it is necessary to assume that the charges 
within the radium atom are revolving much more rapidly than 
those within the calcium atom. The approximate speed 
required to produce this change in mass is given by the 

* It must be admitted that the case of potassium requires some further 


where m is the original mass of the charge at low speeds ; m 
is the mass at velocity v ; and c is the velocity of light : and, 
on working this out, the required velocity is found to be 0*6 
of the speed of light. The actual velocities of j3-particles 
ejected from radium have been experimentally found to be 
0*52 and 0*65 of the velocity of light ; which shows that certain 
intra-atomic particles do attain velocities of this order of 

This view of atomic mass can be used to furnish a solution 
of the isotope question. If it be assumed that in two isotopic 
elements those parts of the atom which furnish the mass are 
identical in the two cases but that in each case the intra-atomic 
machine is revolving at a different speed, then obviously the 
masses of the two atoms would be different. At the same time, 
owing to their identity in construction, they would have 
identical chemical properties. 

Further consideration suggests that possibly this " velocity- 
factor" lies at the root of the Geiger-Nuttall relation, since 
that relation obviously rests upon a physical rather than upon 
a chemical basis. 



1. Mendelceff, LotJiar Meyer and Crookes. 

IN its outlines, the history of our views on the elements is the 
narrative of a conflict between two opposed conceptions : con- 
tinuity and discontinuity. The attempts of the alchemists to 
convert lead into gold prove that the idea of continuity had 
taken root in their minds, and that they did not regard the 
metals as totally unrelated materials. With Boyle, discon- 
tinuity found its first scientific expression in the definition 
of elementary forms of matter ; but early in last century Prout 
revived the hypothesis of continuity once more by assuming 
that all the elements were built up from hydrogen. Again 
discontinuity appeared during the period in which the exact 
determination of atomic weights was undertaken ; and for 
many years it seemed to be conclusively established that no 
theory of a homogeneous matter could be produced which would 
account for the known facts. Yet in 1888, Crookes again 
revived the discarded hypothesis, though without diverting the 
main trend of chemical opinion in its favour. At last the 
phenomena of radioactivity turned the scale ; for at the 
present time continuity appears to be regarded as the final 
solution of the problem, and electricity has taken the place of 
Prout's hydrogen and Crookes' protyle. How long this phase 
will last, it is impossible to prophesy ; but judging from history, 
it is unsafe to assert that even now either of the conceptions 
has reached its final stage; and it seems probable that fifty 
years hence a new set of phenomena may have been discovered 
which will bring discontinuity again into favour.* 

* Even at the present day we have the conception of electricity including 
positive and negative varieties, which shows the idea of discontinuity again. 



Bearing the foregoing facts in mind, it is not without interest 
to turn back to the early history of the Periodic Law and to 
examine the ideas which guided Mendeleeff and Lothar Meyer 
in their development of the subject; for, curiously enough, 
their views diverged markedly from each other although their 
results were in harmony. 

If one may judge from his writings, Mendeleeff inclined by 
preference to numerical data ; and since numbers are always 
discontinuous, he was led, possibly unconsciously, to the con- 
ception of the elements as distinct forms of matter isolated 
entirely from each other even though displaying certain general 
resemblances. As he himself explains, 1 he selected as the basis 
of his speculations the atomic weights and the compound- 
forming capacities of the elements, and he deliberately excluded 
from his survey the physical properties of matter (such as 
cohesion, specific gravity, etc.), on the ground that exact 
numerical data with regard to them were lacking at that time. 
In other words, he chose atomic weight and valency as his 
criteria ; and since the law of multiple proportions holds good, 
he was naturally ' inclined to regard the atoms of each element 
as uniform in properties and entirely different from those of 
any other primary material. He objected strongly 2 to the 
employment of graphic methods of expressing the Periodic 
Law, on the ground that such methods did not indicate the 
existence of a limited and definite number of elements in each 

Lothar Meyer, 3 on the other hand, proceeded on the 
assumption that atoms are aggregates formed from one and the 
same type of matter and differ from each other only in 
their masses. For the purpose of tracing the change in pro- 
perties produced by increase in mass, he selected the atomic 
volume as his index. He thus started from a narrower basis 
than Mendeleeff did ; but the results obtained by him were 
much wider in scope than those arrived at by the tabular 
method. The difference in the two lines of thought is most 
easily perceived if it be borne in mind that Mendeleeff's in- 
vestigations led directly to the calculation of accurate values 

1 Mendeleeff, Annalen (SuppL), 1872, 8, 134. 

2 Mendeleeff, Principles of Chemistry (1897), II., 19 ff. 

3 Lothar Meyer, Annalen (SuppL), 1870, 7, 358. 


for the properties of unknown elements, whilst Lothar Meyer's 
generalization brought into prominence a large number of 
physical resemblances between the elements which, up to that 
time, had not been seen as a whole. 

To the acute and far-ranging mind of Crookes we owe the 
conception which enables us to fuse together the apparently 
irreconcileable ideas of continuity and discontinuity. In his 
Presidential Address to the Chemical Society in 1888, he was 
led to speculate l upon the ultimate nature of matter ; and he 
drew attention to the difficulties which would be involved if 
some forms of matter proved to be unclassifiable within the 
Periodic System. " If we suppose the elements reinforced by 
a vast number of bodies slightly differing from each other in 
their properties, and forming, if I may use the expression, 
aggregations of nebulae where we formerly saw, or believed 
we saw, separate stars, the periodic arrangement can no longer 
be definitely grasped. No longer, that is, if we retain our 
usual conception of an element. Let us then modify this 
conception. For 'element' read 'elementary group,' such 
elementary groups taking the place of the old elements in the 
periodic scheme and the difficulty falls away." 

For almost twenty years Crookes' views received little more 
than contempt from the average chemist ; but with the dis- 
covery of the isotopes, Crookes' conception of the meta-elements 
was found to be justified. What we term " lead " is an " ele- 
mentary group " the atoms of which differ from one another in 
the most fundamental property of all, viz., mass. 

With the coming of this fresh knowledge, the old clean-cut 
separation of element from element on a pure atomic weight 
basis has passed away. Chemical properties and X-ray spectra 
are now our chief instruments in classification, and it seems 
possible that soon we must either increase the delicacy of our 
chemical methods of separation or abandon the chemical 
criteria as ultimate tests. The chemical properties of matter 
are so ill-defined that we can hardly regard them as the last 
word on the subject. The number of molecules contained in 
even the most microscopic quantity of a reagent is so enormous, 
that our results are the merest generalizations and tell us very 
little of the qualities of the individual atoms with which we 
1 Crookes, Trans., 1888, 53, 490. 


are dealing. Possibly at some future time our analytical 
methods will become more refined positive ray analysis fore- 
shadows something of the sort. 

2. The Periodic Table and the Atomic Volume Curve. 

The Periodic Table, as laid down by Mendeleeff in his 
writings, exhibits a symmetry which was one of its greatest 
assets. For some psychological reason, symmetry has an 
attraction for the human mind ; and we are always apt to 
prefer a regular arrangement to one in which irregularities pre- 
dominate. Psychological peculiarities are, however, undesirable 
guides in the search for truth ; and a careful examination of 
the Table in the light of our present knowledge will suffice to 
show that it can boast of no such symmetry as we are led to 
expect from the text-books of our student days. 

For example, owing to the omission of some of the rare 
earth elements and by the insertion of blanks, the Table in its 
original form attained a very high degree of regularity ; but 
since there are, as we know from the X-ray spectra results, 
only sixteen elements to fill the eighteen vacant spaces in the 
Table, it is evident that the symmetry of Mendeleeff s system 
is purely factitious. 

Again, the " typical elements " of the second series do not 
resemble the elements of the fourth series nearly so closely as 
they approach the odd series elements. For example, carbon, 
nitrogen, oxygen, and fluorine, all form alkyl derivatives and are 
thus parallel in their behaviour to germanium, arsenic, sulphur, 
and chlorine ; but no alkyl derivatives of titanium, vanadium, 
chromium or manganese have been prepared. 

Further, in order to produce the appearance of symmetry, 
Mendeleeff was forced to place copper, silver, and gold in the 
first group, although there is no known oxide Au 2 and the 
stable chloride of gold is AuCl 3 . 

Finally, while the elements of Group VIII. are placed 
together on the assumption that their highest oxides are of the 
type E0 4 , in actual practice only two, osmium and ruthenium, 
give such compounds, the other seven elements not yielding 
anything higher than E0 3 . 

These examples are well-known, and are mentioned here 


only for the purpose of enforcing the statement that the 
symmetry of MendeleefFs system cannot be sustained at the 
present day. Fascinating though its cut-and- dried regularity 
may be, we cannot afford to let symmetry dominate our minds 
when in actual fact there is no symmetry to be found. 

The most superficial examination shows that, instead of 
being a symmetrical whole, the Table is really pieced together 
from a series of discrete sections. In the first place we have 
the two short Periods, beginning with helium and ending 
with chlorine. These form a clearly-cut section by themselves ; 
for there is nothing to disturb the regular periodicity of the 
arrangement. Then come the first two long Periods, which 
also form a complete section in themselves, beginning with 
argon and ending with iodine. The long Periods do not repeat 
the exact sequence of the short Periods : for the difference 
between the A and B Groups makes its appearance, and the 
elements of Group VIII. break the parallelism between the 
long and short periods. The third long Period, beginning 
with xenon, marks the complete collapse of the symmetrical 
arrangement. Xenon, csesium, and barium follow the usual 
order; and then come the rare earths and confusion. There- 
after, the usual sequence appears once more at tantalum and 
continues, probably, down to the missing halogen with the 
atomic number 85. With the fourth long Period we reach the 
radio-elements and enter a new field in which radioactive 
properties make their appearance. 

From the above, it will be seen that the Table contains the 
following fragments : (1) two short Periods ; (2) two long 
Periods ; (3) a broken long Period ; (4) the rare earth group ; 
and (5) the radio-elements. The task of fitting together these 
five portions into a symmetrical whole has hitherto baffled those 
who have attempted it ; but there seems to be no reason for 
assuming that the problem is insoluble. 

Soddy l prefers to draw a distinction between two portions 
of the Table. The three series of elements, vanadium-ger- 
manium, colurnbium-tin, and tantalum -lead are regarded by 
him as " interpolated " elements ; and the table is thus 
divided into two short periods ; two long periods each with a 
hiatus between Groups IV. and V. ; an interrupted long period 
1 Soddy, Matter and Energy, p. 65. 


beginning with xenon; another long period commencing with 
niton, the rare earth elements ; and three series of " inter- 
polated " elements. 

Many other attempts 1 have been made to produce new 
tabular forms, but none of them combine symmetry with the 
accurate representation of the elemental relationships. 

Leaving the tabular arrangement, we may now proceed to 
the consideration of the curve of atomic volumes. In the past, 
it has been usual to plot the abscissse as atomic weights ; but 
in view of the fact that atomic weights are no longer to be con- 
sidered as physical constants, it seems desirable to replot the 
graph, using the atomic numbers on the horizontal axis. The 
diagram given at the end of this volume shows the result ; and 
it will be seen that the change leads to an even distribution of 
the points on the curve. 

The most casual examination of the figure shows that the 
curve is not a regular one. There is a steady increase in atomic 
volume as we pass from lithium, through sodium, potassium, 
and rubidium to caesium at the maxima of the graph ; but if 
the minima are investigated it will be found that no such 
regularity exists in that section.* The five known minima, 
instead of lying in a crescendo or diminuendo order, apparently 
fall at random, as the following figures show : 

Minimum Element. Carbon. Aluminium. Nickel. Ruthenium. Osmium. 
Atomic Volume 3'5 10'5 6'7 8'3 8'5 

Thus the minima do not occur, like the maxima, at elements 
analogous to each other in chemical nature ; nor do they fall 
into any clear order in respect to the actual volume of the 
minimum element in each cycle. 

1 See Garrett, The Periodic Law (1909), for a summary of work up to that 
date ; and also Martin, Researches on the Affinities of the Elements ; Adams, 
J. Amer. Chem. Soc., 1911, 33, 684 ; Hopkins, ibid., 1005; Harkins and Hall, 
ibid., 1916, 38, 169 (c/. Soddy, Chem. Soc. Ann. Report, 1916, 13, 254); 
Hackh, ibid., 1918, 40, 1023; Steinmetz, ibid., 733; Stackelberg, Zeitsch. 
physikal. Chem., 1911, 77, 79 ; Vogel, Zeitsch. anorgan. Chem., 1918, 102, 177 ; 
Schmidt., ibid., 1918, 103, 79; Loring, Studies in Valency., p. 28; Thornton, 
Phil. Mag., 1917, 34, 70 (cf. Richards, Trans., 1911, 99, 1201) ; Haughton, 
Chem. News, 1888, 58, 93, 102. 

* If the curve be plotted with atomic radii as ordinates instead of atomic 
volumes, it assumes a less irregular aspect, but this is so merely on account 
of the disparities being reduced by the method of plotting. 



In this connection, it must be borne in niind that our 
methods of determining the atomic volume of an element at 
the present time are not free from obvious errors. Atomic 
volume, in the sense in which the term is used in connection 
with the curve, is simply a figure obtained when the atomic 
weight of the element is divided by the density of the element in 
the solid state. It represents merely the ratio of mass to space 
occupied. But the " space occupied " includes in itself no less 
than four factors : (1) the space actually filled by the minute 
positive and negative charges of the atomic machine ; (2) the 
spaces between these charges within the atom ; (3) the spaces 
between the atoms in the molecule of the element ; and (4) the 
spaces between the molecules of the element in the mass with 
which we are dealing. It is self-evident that when we com- 
pare two different elements, these four factors are not neces- 
sarily identical. For example, in the case of a monatomic 
element like mercury, the third factor cannot be present ; 
whilst the difference in coefficient of thermal expansion between 
certain elements proves that the fourth factor varies with 

Mendeleeff l believed that when the atomic volume is large, 
the atoms in the molecule of the element stand comparatively 
far apart from each other ; whereas those elements possessing 
small atomic volumes have their atoms closely adjacent to each 
other in the molecular structure; and to this he traced the 
difference in reactivity between the two elemental types.* The 
fact that compressibility seems to be a periodic function 2 
suggests that the inter -molecular spaces depend upon the nature 
of the compressed element. Further, those elements with large 
atomic volumes are more readily compressible than the ones 
near the minima of the Curve. 

Again, when the elements in the same family are examined, 
it is found that the linear coefficient of thermal expansion 
changes from element -to element as we pass down the group. 

1 Mendeleefi, Annalen (Suppl), 1872, 8, 156. 

2 Richards, Zeitsch. physikal Chem., 1907, 61, 98; Trans., 1911, 99, 1201; 
Zeitsch. Elektrochemie, 1907, 13, 519. 

* Mendeleefi used the analogy of a sponge to represent a reactive element 
which can easily be penetrated by the atoms of other elements, whereas a 
non-reactive element was likened by him to a more solid and less permeable 
material (Principles of Chemistry, II., 33 (1897)). 


Thus magnesium has a coefficient of 27*5 ; zinc has 29 ; 
cadmium has 31 ; and mercury has 61 ; but on the other hand 
the order is reversed in the electronegative families ; for the 
coefficient of chlorine is 470 while bromine and iodine have 
390 and 83'7 respectively. 

From considerations such as these, it is evident that our 
present methods of estimating atomic volumes are extremely 
rough ; and hence the irregularity of the atomic volume curve 
is only what might be expected when we take iuto account the 
nature of the data employed in framing it. 

It seems not improbable that something of importance 
might be achieved by an attempt to recalculate the atomic 
volumes of the atoms by the insertion of corrections based on 
the known data with regard to compressibility and expansion 
coefficients. A compressibility correction would reduce the 
atomic volumes of the alkali series and the halogens ; while 
the correction for expansion would, as far as can be seen, tend 
to equalize the Curve considerably. 

In its present form (see the diagram at the end of this 
volume), the Atomic Volume Curve contains five waves. The 
first two of these correspond to the two earliest series in the 
Mendeleeff Table ; and each of them includes eight elements. 
The next two waves are longer, containing eighteen elements 
each. The extension is due to the appearance in the middle of 
the wave of a series of elements (titanium to zinc and zirconium 
to cadmium) which do not exactly reproduce the characters of 
Mendeleeff s typical elements. The fifth wave is yet more 
extended, since in it the rare earth elements are interpolated ; 
and it contains thirty-two elements. 

With regard to the intercalary sections of the third, fourth, 
and fifth waves, it is not without interest to observe that the 
iron, ruthenium, and osmium groups occupy minima on the 
Curve corresponding to the carbon minimum in the first wave ; 
and that while the element carbon is pre-eminent in its capacity 
for chain-formation, the " minimum " groups in the later 
periods are marked by their complex- salt production. It may 
be suggested that small atomic volume is connected in some way 
with the firm retention of atoms or groups in a compound. 

Turning to the relations between valency and the atomic 
volume curve, it may be worth our while to examine the 


conditions in the waves beginning with potassium and rubidium. 
The following figures give the highest valency exhibited in the 
oxygen compounds or acids : 

K Ca Sc Ti V Or Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr 

1234567 3 422 2 3 456 (5) 

Eb Sr Y Zr Cb Mo Ku Rh Pd Ag Cd In Sn Sb Te I X 

123456 ? 8 641 2 345670 

An examination of the above will show that if the abnormality 
of bromine be neglected, the first series falls into three distinct 
groups. From potassium to manganese there is a regular 
increase in valency ; from zinc to bromine, if the salts of per- 
bromic acid can be proved to exist, there is again a regular 
increase in valency ; whilst in the intermediate group, from 
iron to copper, there is no regularity traceable. In the second 
series, there is a regular increase in valency from rubidium to 
ruthenium and also from silver to iodine ; but again there is an 
intercalated group rhodium and palladium which diverges 
from the regular sequence. It is evident that the Atomic 
Volume Curve*, like Mendeleeff s Table, fails to bring out any 
perfect symmetry where valency relations are concerned. 

With the second type of intercalary elements, the rare earth 
group, it will be convenient to deal in a separate section of this 

The peculiar resemblances in certain properties between 
some elements which lie adjacent to one another in the atomic 
order, find a better expression in the Curve than in the Table. 
For example, in the case of boron and carbon we have the 
hydrides B 2 H 6 and C 2 H 6 , where the similarity in composition 
is reinforced by the fact that both substances are colourless 
gases. The close approximation of the two elements on the 
curve throws up their relationship very much better than does 
the ordinary tabular arrangement, especially in cases such as 
this where the valency of boron must actually be " abnormal " 
in order to produce the similarity. 

When we turn to physical properties, however, the supreme 
advantage of the Curve over the Table becomes manifest. By 
the superposition of the odd and even series in the Table, the 
elements are juxtaposed in a manner which conceals almost 
completely the general periodicity of physical properties ; for, 


intermingled with the soft white metals of the alkali group we 
get the coloured metals, copper and gold ; and among the non- 
metallic halogens we find the definitely metallic manganese. 
In the Curve, each set of physical properties recurs at regular 
intervals, as can be seen from an inspection of the diagram at 
the end of this volume. 

Lothar Meyer 1 drew attention to four peculiarities in the 
distribution of the elements on the Curve. In the first place, 
elements with small atomic volumes do not appear to possess 
an electrochemical character which is either strongly positive or 
strongly negative : carbon, for instance, combines with both 
hydrogen and chlorine to produce stable derivatives, while iron 
forms ordinary salts and gives rise to the ferrates as well. In 
the second place, rapid changes in the slope of the Curve go 
hand in hand with great alterations in the chemical character 
from element to element, while a slight change in the inclina- 
tion of the Curve is paralleled by small differences in chemical 
nature between two adjacent elements. Thirdly, electro- 
positive elements in the first two waves occur at the maxima 
or on the downward slopes beyond the maxima, while the 
electronegative elements are found either at the minima or on 
the upgrade between a minimum point and the next maximum. 
In the case of the three longer waves, both maxima and minima 
are occupied by electropositive elements ; the elements 
adjacent to them are also positive, and electronegative (or 
else amphoteric) elements take up positions intermediate 
between the maxima and minima. For example, in the first 
long wave, potassium occupies the maximum and calcium is 
the next element on the down- slope; then come titanium, 
vanadium, and chromium, which, though metallic, give acidic 
oxides ; at the minimum lie iron, cobalt, and nickel, which are 
metallic in the main; then follow the elements copper, zinc, 
gallium and germanium, which, though amphoteric in some 
circumstances, are mainly electropositive ; and, finally, on the 
rising gradient lie the elements arsenic, selenium and bromine, 
which show increasingly electronegative character as we pass 
up the series. Lothar Meyer's fourth comment is concerned 
with the fact that in certain cases two elements may have 

1 Lothar Meyer, Modern Theories of Chemistry, 1888, p. 149; Annalen 
(Suppl), 1870, 7, 354. 


almost exactly the same atomic volume and yet may differ 
markedly in chemical properties, as is the case with sodium and 
chlorine or sulphur and indium. 

It seems a matter of interest to trace, if possible, some con- 
nection between these observations of Lothar Meyer, and the 
modern electrical theory of matter, in order to see whether our 
newer views throw any light upon the points in question. To 
do that, it is necessary to assume that the " atomic volume " 
expressed in the curve really gives us an approximate measure 
of the relative spaces occupied by the atoms themselves ; and, 
as has been noted, this is not necessarily anything but a very 
rough approximation. Let it be assumed, however, that the 
" atomic volume " is a measure of the space actually occupied 
by an atom. If so, the diameter of the outermost electronic 
orbits must be proportional to the atomic volumes. We may 
further assume that electropositive character implies a readi- 
ness of the system to lose electrons, whilst electronegative 
character corresponds to an attraction for extra negative 
electrons towards the system. 

Now it is possible to imagine a system in which the 
dimensions are such that the attraction of the central nucleus, 
being exerted across a large radius, is not sufficient to attract 
external electrons through the negative ring of the atom because 
of the repulsive influence of the ring-electrons which are acting 
at a comparatively short distance from the external negative 
electron ; and, further, that the attraction of the positive nucleus 
is sufficient to prevent any of the ring- electrons leaving the 
atom, because no other positive nucleus can approach close 
enough to exert an overpowering attraction on the ring-electrons. 
Such a case would represent the inert gas atoms. 

If the atomic radius be increased beyond this critical value, 
then the attraction of the positive nucleus will be correspond- 
ingly weakened ; and electrons will be more easily enabled to 
quit the atom under external attractions. This case would 
correspond to the alkali metals. 

If the critical radius be reduced, the pull of the central 
nucleus on the ring-electrons will be strengthened ; and hence 
they will show no tendency to leave the system : while at the 
same time, owing to tKe possibility of near approach of external 
electrons being increased, it becomes more possible for electrons 


to be drawn into the atom from other systems. This case, then, 
corresponds to the halogen atoms. 

The results of still further contraction of the orbital radius 
will be governed by two opposed factors. In the first place, 
the smaller the radius, the greater the extension of the influence 
of the positive nucleus as compared with- the diameter of the 
ring ; while in the second place the ring-electrons will be more 
closely adjacent in space in the contracted circuit and they will 
therefore tend to repel each other and so facilitate their with- 
drawal from the atomic system. It seems probable that the 
alternation of positive and negative characters in the elements 
in the lower parts of the wave is due to the ascendancy attained 
by each of the factors in turn. 

There remains the problem of accounting for the fact that 
elements such as sodium and chlorine differ entirely in 
properties whilst their atomic volumes are identical. Sodium 
has a positive nucleus of eleven charges while the positive 
nucleus of chlorine contains seventeen charges. Since the 
orbital radii are equal in the two cases, it is evident that the 
ring-electrons of sodium will be much more weakly attracted 
than is the case in chlorine ; and also that the tractatory power 
of the chlorine nucleus upon negative electrons just outside 
the ring will be greater than that of the sodium nucleus. The 
counter- acting factor repulsive influence of the ring-electrons 
upon external negative charges is, of course, also increased 
in the case of chlorine; but since the positive charges are 
grouped together in a small space whereas the negative charges 
are extended over a large circumference, it seems reasonable to 
assume that the repulsive effect (as far as purely local results 
are concerned) will only be slightly increased as compared with 
the increase in the pull of the positive nucleus. 

There is one factor which is not well expressed in the 
periodic curves as hitherto drawn. Some concrete examples 
will serve to make the point clear. In passing from lithium, 
with the atomic number 3, to beryllium, with atomic number 4, 
it is clear that the addition of a single positive charge to the 
nucleus represents an actual increase of the charge by one- 
third : whereas the addition of a positive charge to the zinc 
atom only results in a change of one-thirtieth of the original 
charge, since the atomic number of zinc is 30. When the 


curve is plotted either with atomic weights or atomic numbers 
as abscissae, this disparity in the relative values of the additional 
charge is not brought out ; and it seems of interest to discover 
some graphic method which will put the matter in a clearer light. 

If the abscissae be represented by the logarithms of the 
atomic numbers, the desired result is attained, as then equal 
changes in the abscissae represent equal fractional differences in 
the nuclear charge, no matter at what point on the curve they 
may be measured. Further, if the ordinates be made pro- 
portional to the logarithms of the atomic volumes, any multiple 
relationship between the two factors will be expressed in the 

The result is to be seen in Fig. 23. Examination of the 
curve will show that the wave beginning with lithium is much 
broadened as compared with that on the ordinary type of graph, 
while the succeeding waves are made more and more narrow 
as we pass along the line. The meaning of this will be 
apparent from a numerical example. Beryllium lies at the 
point 0*60 while boron is situated at 0-70, leaving a gap of 
0*10 between them ; magnesium is placed at 1'08 and aluminium 
at I'll, the gap in this case being only 0*03. The chemical 
differences between beryllium and boron are too obvious to 
need recalling to the reader ; whereas magnesium and aluminium 
are much more closely alike. Evidently, on this system of 
plotting, differences in the abscissae correspond to some extent 
to differences in chemical nature : * the greater the differences 
between the abscissae of two adjacent elements, the greater the 
chemical differences between them. In the light of this, the 
compression of the succeeding waves becomes explicable, for it 
is well known that with increase in atomic number even the 
electronegative halogens become endowed with semi-metallic 
character and resemble the rest of the elements more closely. 
Again, the differences in chemical and physical properties 
between iron, cobalt, and nickel on the one hand, the ruthenium 
and osmium groups on the other is illustrated by the broadened 
form of the curve at the iron triad as compared with the 
sharper points of inflexion at ruthenium and osmium. 

* This is merely given for the sake of easy reference, since actually the 
important factor would appear to be the length measured along the curve 
from element to element. 

o o o o 



3. Spiral Arrangements of the Elements. 

The first attempt to arrange all the elements in a periodic 
grouping took the form of a three-dimensional model the 
Telluric Helix of de Chancourtois ; l and it is not surprising 
that from time to time attempts have been made to utilize the 
third dimension as an aid to classification. It cannot be said 
that much light has been thrown on the matter by these 
essays ; but some account of them must be given here for the 
sake of completeness. 

Eeynolds 2 put forward the analogy of a vibrating knotted 
string to represent the periodic arrangement; and his proposal 
was modified by Crookes 8 a couple of years later. Crookes 
suggested that the Atomic Volume Curve represented the pro- 
jection of a spiral figure which might either be a simple spiral 
or one of the "figure-of-eight" type. The elements of low 
atomic weight were grouped in the upper turns of the spiral 
and the valency was represented by dividing the separate 
sections of the curve into sectors each of which contained an 
element, so that the members of the same family stood above 
one another on the spiral. In Crookes' view, his spiral repre- 
sented the effect of the interplay of two forces electrical 
charge and temperature. Movement along the spiral, if pro- 
jected on a horizontal plane, indicated the change in electrical 
character ; while the vertical fall occasioned by the descent of 
the point in travelling along the spiral represented a lowering 
of temperature. 

A somewhat similar model was proposed by Harkins and 
Hall. 4 It is differentiated from that of Crookes by the fact 
that the rare earth elements are all placed in a vertical line 
instead of being grouped round the spiral. Stewart 5 suggested 
a three-dimensional model on the principle of bringing into 
line all elements showing the same numerical valency and 

1 See Compt. rend., 1891, 112, 77; and also Nature, 1889, 41, 186, for an 
account of de Chancourtois' work which was first published in 1862. 
8 Eeynolds, Chem. News, 1886, 54, 1 ; Nature, 1895, 51, 486. 

3 Crookes, Chem. News, 1886, 54, 115 ; Trans., 1888, 53, 487. 

4 Harkins and Hall, J. Amer. Chem. Soc., 1916, 38, 169. 

5 Stewart, Recent Advances in Physical and Inorganic Chemistry, 3rd 
Edition, 1918. 


differentiating between the three different types of transition 
which occur at Group VIII., Group IV., and the Zero Group. 
Soddy 1 proposed a spiral which combines to some extent the 
normal and figure-of-eight curves in a single model. 

A plain spiral arrangement was put forward by Stoney, 2 
who showed that the atomic weight relationships of the elements 
could be depicted by means of a logarithmic spiral ; but owing 
to the modern disrepute into which atomic weights have fallen, 
it may be doubted if this plan meets the case. 

The main drawback to the spiral representation appears to 
be that in it no new facts are brought to light, and there is no 
fresh collocation of the allied elements which might give it an 
advantage over the ordinary forms of classification. Also, in 
most cases it is more difficult to grasp as a whole. Soddy's 
arrangement seems the most perspicuous, if the spiral grouping 
be adopted. 

4. The Periodic Surface. 

The Periodic System is a pure matter of classification ; for 
we have no inkling of the factors which produce the recurrent 
resemblances among the elements. Prolonged study of the 
Table always leaves on the mind the impression that behind 
this curious grouping there is some simple law which, when it 
is at last discovered, will seem almost self-evident : but up to 
the present nothing has been suggested which throws any real 
light upon the subject. Obviously our immediate effort should 
be directed to producing an arrangement which, regardless of 
pre-conceived ideas, will bring out as far as possible all the 
known similarities ; for when such a system is evolved there is 
a certain chance that it may suggest to our minds the esoteric 
influences which control the periodicity of properties among the 
various forms of matter with which we are acquainted. 

In the foregoing pages it has been shown that Mendeleeff s 
Table and Lothar Meyer's Curve each possess valuable in- 
dividual characteristics. The Table shows at a glance the 

1 Soddy, Chemistry of the Radio-elements, II., 9, 1914. 

2 Stoney, Phil Mag., 1902, 4, 411, 504 ; see also Eayleigh, Pro. Roy. Soc., 
1911, A, 85, 471 ; and compare Harkins and Hall, J. Amer. Ghent. Soc., 1916, 
38, 189 ; Loew, Zeitsch. physikal. Chem., 1897, 23, 3 ; Stintzing, ibid., 1916, 
91, 500 ; Caruelley, Chem. News., 1886, 53, 183. 


general chemical relationships of the elements ; but without 
undue complication of lettering, it is impossible in it to bring 
out simultaneously the corresponding periodicity in physical 
properties. The Curve, on the other hand, exhibits the re- 
currence of physical properties ; but by its very nature it fails 
to demonstrate clearly the relations between the elements of 
a given family, since the members are scattered over the various 
waves and not collected together. Obviously it would be 
advantageous if the good points of both representations could 
be combined in a single system, exhibiting simultaneously 
chemical and physical properties. 

Before this can be done, however, it is necessary to re- 
arrange the Mendeleeff Table ; for, as it stands, the alternation 
of the A and B sub-groups would throw the physical properties 
into a disorderly series. The following arrangement was 
adopted to produce the model illustrated in the frontispiece 
to this volume. 

He Li Be B C N O F 

NeNaMgAl Si P S 01 

A K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br 
Kr Eb Sr Y Zr Cb Mo Eu Eh Pd Ag Cd In Sn Sb Te I 

Ta W Os Ir Pt Au Hg Tl Pb Bi Po - 
Nt EaAcThUX 2 U 

As can be seen, the two short periods have been broken up 
and the elements in them have been distributed as far as 
possible along with their nearest analogues in the long periods. 
Thus fluorine is classed together with the halogens, oxygen and 
nitrogen along with sulphur and phosphorus. Carbon has been 
placed with silicon rather than with titanium and its analogues. 
The positions of beryllium, magnesium, and aluminium have 
been selected on account of the metallic character of the 
elements ; but it must be admitted that they might with equal 
justice have been transferred^ to the corresponding positions 
over zinc and gallium. 

If now, from each point representing an element in the 
above scheme a perpendicular be raised with a length pro- 
portional to the atomic volume of the element and if the 
summits of these perpendiculars be united by a surface, the 


three-dimensional model shown in the frontispiece is pro- 

The advantages which this form of representation possesses 
over the other two are obvious on inspection (see frontispiece). 
In the first place, by getting rid of the alternation of A and B 
sub-groups, it brings into direct contiguity the " atomic 
analogues " of Mendeleeff and separates from one another 
the oxygenic and basigenic elements which are associated in 
the Mendeleeff Table. Secondly, it throws into prominence the 
abnormality of the rare earth group, as can be seen from 
the abrupt elevation in the centre of the model, which is quite 
out of harmony with the rest of the contours. Further, it will 
be found that all elements giving coloured chlorides in their 
highest stage of valency fall compactly together in the centre 
of the model. Finally, while the Atomic Volume Curve only 
shows the change of slope in passing from one element to the 
next in the atomic order, the Periodic Surface exhibits simul- 
taneously the volume relations of neighbouring elements and 
also those existing between members of the same family. 

It is frankly to be admitted that the model possesses very 
little of the symmetry to which we have been accustomed ; but 
as has already been pointed out, if we have to choose between 
factitious symmetry and actual correlation of facts, we must 
decide in favour of the latter, discomforting though the choice 
may be. 

5. The Problem of the Rare Earth Elements. 

Of all the problems centering in the Periodic Table, the 
position of the rare earth elements is the most difficult. From 
the X-ray spectra, it appears to be established that between 
lanthanum and tantalum there are fifteen elements in the 
atomic order ; and if celtium be placed next to tantalum, this 
leaves only one element of the group still to be detected, which 
will have the atomic number 61 and will lie between neody- 
mium and samarium. 

Numerous attempts have been made l to find a satisfactory 

* Full directions for the construction of a slightly improved model are 
given in the Appendix. 

i Thomson, Zeitsch. anorgan. Chem., 1895, 9, 190 ; Brauner, ibid., 1902, 
32. 1 ; Budorf. ibid., 1903, 37, 177 ; Benedicks, ibid., 1904, 39, 41 ; Vogel, 


arrangement of the rare earth elements ; but up to the present 
no generally accepted solution of the problem has been reached. 
The proposals hitherto put forward may be grouped under 
three heads : (1) the elements may be placed all together in 
one group of the Table ; (2) they may be made to form a 
bridge connecting lanthanum with tantalum, outside the 
ordinary tabular arrangement ; and (3) they may be distributed 
among the groups in a more or less symmetrical manner. 
These three views may be dealt with in turn. 

Owing to the general similarities in the rare earth family, 
it has been proposed that they should be arranged in a single 
group of the Table. Since they all form oxides of the type 
E 2 3 , the third group seems the most appropriate ; but there 
is an alternative suggestion by Brauner that lanthanum should 
be left in Group III. while the elements from cerium to the 
predecessor of tantalum are placed in Group IV. Brauner's 
actual arrangement, however, necessitates the existence of no 
less than twenty elements between lanthanum and tantalum ; 
so it does not agree with the present state of our knowledge. 
K. Meyer removes cerium to Group IV. and leaves all the rest 
of the family in Group III. ; but this arrangement deviates too 
much from the known atomic order. From these examples, it 
will be seen that any attempt to retain even the majority of 
the elements in a single group is open to grave objections. 

The second suggestion that the rare earth family has no 
place in the Table but forms a " bridge " between lanthanum 
and tantalum is one which can hardly recommend itself in 
view of the regular sequence into which the X-ray spectra fall. 
It has been proposed that the elements should be placed in 
a horizontal plane while the rest of the Table is inscribed on 
a vertical surface ; but this idea conveys nothing very definite 
to the ordinary mind. It has also been suggested by Vogel 
that the normal elements in the Table should be placed on 
a regular spiral of eight places ; while the rare earth group 

ibid., 1918, 102, 177 ; Betgers, Zeitsch. physical. Chem., 1895, 16, 651 ; Baur, 
ibid] 1911, 76, 569; Biltz, Ber., 1902,35, 562; Werner, ibid., 1905, 38, 914; 
Meyer, Physical Zeitsch., 1918, 19, 178; Brauner, Zeitsch. Elektrochem.,1908, 
14, 525; Hicks, Phil. Mag., 1914, 28, 139 ; Steele, Chem. News, 1901, 84,285; 
Harkins and-Hall, J. Amer. Chem. Soc., 1916, 38, 169 ; WyroubofE and Verneuil, 
Ann. chim. phys., 1906, 6, 466 ; Reychler, Les theories physico-chimiques, 1903, 
p. 50 ; Soddy, Chemistry of the Radio-elements, II., 9, 1904. 


should be accommodated on a subsidiary spiral inserted after 
barium. Here also there seems to be no very concrete reality 
behind the representation. 

With regard to the distribution of the rare earth group over 
the ordinary Table, it is impossible to fit the rare earth elements 
into 'MendeleefFs arrangement owing to the disparity in 
numbers between the blanks and the elements predicted by the 
X-ray spectra. Also, owing to the peculiar character of the 
rare earth group, it is straining probability too far if we assume 
that one of these elements can take its place among the 
halogens. If the rare earth family is to be distributed among 
the "normal" elements, it must be done by means of some 
new arrangement. 

The following new grouping seems worth considering. 
Although it has many good points, it is not to be regarded as 
a final solution, but is put forward mainly in the hope that an 
examination of it may suggest some more perfect system. 

According to Moseley's results, the rare earth elements 
follow in natural sequence after csesium and barium ; and it is 
therefore reasonable to look for their congeners in corresponding 
places in the rest of the Table. Taking valency as our first 
guide, we find that the rare earth elements, with the exception 
of cerium, form stable oxides of the type B^C^ ; and as a 
beginning we may search out those elements in the other series 
which also yield this oxide form. Very little consideration 
will show that the required elements form definite and limited 
series, one of which runs from scandium to cobalt,* another 
from yttrium to rhodium. Now if we follow out this train of 
analogy, we shall be led to group the rare earth elements in the 
following manner : 

Kb Sr Y Zr Cb Mo Eu Eh Pd Ag 

Cs Ba La Ce Pr Nd Sm Eu 

Gd Tb Dy Ho Er Tm Yb 

Lu Ct Ta W Os Ir Pt Au 

At this point it may be well to apply the atomic weight 

test, as a rough clue to the accuracy of the arrangement. Owing 

* The oxide Ni 2 3 is apparently a mixture of NiO and Ni0 2 and not a true 
compound ; palladium sesquioxide is unstable ; and the anhydrous form of 
Pt 2 3 has not been obtained. 


to the similarity in general character, it may be assumed that 
the intermediate row of elements beginning with gadolinium 
should have atomic weights approximately equal to the average 
of the atomic weights of the elements above and below them. In 
other words, the atomic weight of gadolinium should lie near 
the mean of those of lanthanum and lutecium. The following 
figures show the results, the figures marked " Calculated " being 
those of the averages : 

Gadolinium. Dysprosium. Holmium. Erbium. Thulium. Ytterbium. 

Experiment 157-3 162-5 163-5 167-7 168-5 173-5 
Calculated 157'0 161-0 164-1 167'4* 170-7 172-5 

It will be seen that the agreement is remarkably good in some 
cases and is not bad on the whole. On the other hand, the 
agreement is by no means good when we calculate a mean value 
for cerium by taking the average of the atomic weights of 
zirconium and terbium. 

Since the rare earth elements form salts of various tints, it 
seems advisable to see whether the suggested arrangement brings 
to light any parallelisms in this field. The table below shows the 
state of the matter, the chlorides being chosen as the simplest 
salts in which complications are unlikely to present themselves 
and the highest chloride being selected in each case. 

* Calculated from the values for Nd, Sm, W and Os. 







H 5 


e 3 




* s 













3 w 


An examination of the above scheme will reveal certain points 
of interest. In the first place, as shown by the heavy line, all 
the elements yielding coloured chlorides lie in a single compact 
area. Secondly, there is a certain symmetry in the arrange- 
ment, since on the right-hand side there is a regular shift to 
the left of one place as we reach the end of the coloured salts, 
while on the left-hand edge of the scheme there is also a 
regularity ; for all the elements giving colourless chlorides are 
grouped there.* 

Further examination of the arrangement will show that it 
contains certain harmonies in the salt-colours which are not 
without interest. All chlorides of elements in the first column 
are white. In the second column, cerium is the only exception 
to the regular whiteness ; and there is no doubt whatever as to 
the place of cerium in the arrangement, since it follows lantha- 
num in Moseley's sequence. The third column exhibits a 
regular gradation of tint in praseodymium, dysprosium, and 
tantalum ; and the same may be said of the adjacent grouping : 
neodymium, holmium, and tungsten. In the next column the 
pink of the manganese salts is brought into correspondence with 
the pink of the erbium derivatives. The next column brings 
into prominence the close colour-alliance of iron, ruthenium, and 
samarium on the one hand, and thulium and osmium on the 
other. Finally, europium is placed in a proper colour-relation 
to cobalt and rhodium. 

It must be admitted that this evidence presents features of 
interest, even if the weight assigned to it is not excessive. 

When we come to the oxides, it will be found that the 
arrangement suggested above brings out another point. Among 
the rare earth elements, cerium, praseodymium and terbium are 
marked out by the fact that they furnish oxides of the type K0 2 . 
It will be noticed that the system suggested here places both 
cerium and terbium in the titanium group. Titanium and 
zirconium yield oxides of the types R 2 3 and E0 2 , which brings 
these elements into line with cerium and terbium \ ; and, like 

* It may be noted that if cerous chloride had been chosen instead of eerie 
chloride, then the colour in that case would have been white and the only 
exception would disappear. 

t The impure "peroxide " of terbium appears to correspond to the formula 
Tb 4 7 and may be a mixture of Tb 2 3 and 



cerium also, titanium and zirconium yield the stable dioxide, 
while their sesquioxides are less stable. 

In the next column we find praseodymium classed along 
with vanadium, columbium, and tantalum, These last three 
elements yield both types of oxides R 2 3 and E0 2 ; and it is 
of interest to note that praseodymium gives rise to corresponding 
compounds Pr 2 3 and Pr0 2 . 

Thus the system suggested appears to fit in well with the 
valency peculiarities of the rare earth elements. 

Turning to another line of evidence, the molecular volumes 
of the rare earth chlorides may be examined. It will be seen 
from the figures below that there is a steady contraction in 
volume from lanthanum to samarium ; and thereafter a rise, so 
that gadolinium may be regarded as the beginning of a new 
series. This agrees with the arrangement suggested. 

La Ce Pr Nd Sm (Eu) Gd Tb Dy 

63-8 62-9 60-8 60'6 ? 57'5 ? 58'3 61*1 73'3 

The evidence accumulated above is sufficient to show that 
the foregoing system deserves a certain amount of consideration. 
Against it may be urged the fact that it takes no note of the 
relations which are known to exist between the solubilities of 
the double salts of the rare earth group, since it splits the 
terbium group in half. But solubility is hardly a final test in 
a case such as this ; though it must be allowed its due weight 
as evidence. 

The complete Periodic Table drawn up on this basis is 
printed on the following page. 







! oo 







B -sl 

^ 3 S 
I S| 

^^ a 
H g ^ 

e Hi 


CQ BH 3"1 



r- i & -ta *& 


s^ o 


2^ 3 


6. The Theory of Meta-elements. 

The acceptance of Crookes' views on the composition of the 
materials which we term elements has been forced upon us by 
the experimental proof that isotopes exist ; but up to the 
present no attempt seems to have been made to follow out to 
its logical conclusion the train of reasoning which is suggested 
by this result. We have accepted the idea that a substance 
may respond to all our chemical tests as if it were homogeneous 
in nature and yet may consist of a mixture of different kinds 
of atoms in arbitrary proportions. Indeed, if the newer 
results of positive ray analysis be confirmed, we shall have to 
admit that the chemical element chlorine, for instance, actually 
can be separated into two forms of matter each of which is 
different from the other although neither of them is simpler 
than " chemical " chlorine. Thus as far as the factor of mass 
is concerned, we have made a definite breach in the old 
definition of elements ; and it now behoves us to examine other 
factors equally carefully in order to see whether they are also 
not subject to similar variations from the normal. 

Next to mass, the electrical character of an atom appears to 
be of the greatest interest. Just as in earlier days we were 
satisfied that atomic weights were physical constants, so now 
we assume that the electromotive force of an element is an 
easily definable quantity. But if what we term elements are 
really congeries of non-homogeneous atoms, it seems clear that 
each of these atoms will have its own E.M.F. ; so that the 
quantity actually measured in practice is an average value 
which may not correspond in practice to the E.M.F. of any 
particular atom in the group. It seems not impossible that by 
a prolonged series of "fractional electrolyses" a separation of 
the atoms having different E.M.F.'s might be accomplished.* 
The main difficulties to be anticipated are not concerned with 
actual separation, but are associated, as far as can be foreseen, 
with the proof of the separation after it has been attained. 

At first sight, spectra appear to offer a very accurate criterion 
of the homogeneity of matter; but the results found with 
ordinary and X-ray spectra in the case of the isotopes have 

* An attempt of this kind has been begun in the author's laboratory. 


proved that spectra are not to be relied on as tests for the 
differentiation of a mixture from a homogeneous material. It 
is possible that a revision of our spectroscopic methods combined 
with a photometric examination of the relative brightness in 
the spectral lines might render the test more rigid; but at 
present we cannot properly assume that the resolving power at 
our disposal is sufficient to solve all the possible problems 
which the meta-elements open before us. 

When we turn to the chemical properties of the elements, 
the poverty of our methods of separation cannot fail to strike 
us. In the main, we depend almost entirely upon three 
different processes for dividing one form of matter from its 
neighbours. Differences of solubility between elements or 
between their compounds are the commonest phenomena which 
we utilize in order to separate mixtures into their components ; 
but it is only necessary to consider the case of the rare earth 
group to see the difficulties which attend this method and the un- 
certainties into which it may lead us. Less usual is the method 
of separation by distillation, which also is a very crude process. 
Finally, we have the ordinary electrolytic modes of dividing 
one material from another. It must be admitted that on the 
chemical side our methods are hardly such as to offer any real 
satisfaction. They are quite sufficient for our everyday needs ; 
but when applied to the case of meta-elements they fail com- 
pletely, as is proved in the case of the isotopic groups. To use 
Crookes' analogy, they make us believe that we have discovered 
a fixed star when actually we are dealing with a nebula. 

It seems probable that the next great stride forward in our 
knowledge of matter may come about through the invention of 
new methods of separation. So far, we have been in the 
position of a statistician confronted with a digest of the Census 
figures. We are able to predict very accurately how things 
will behave en masse; just as the Census official can predict 
the probable birth-rate or death-rate for the coming year ; but 
finer values escape our grasp. We can prophesy how many 
radioactive atoms in a given mass will disintegrate within a 
fixed period; but we cannot isolate these particular atoms from 
the remainder of the material before the breakdown takes 


7. Conclusion. 

A survey has now been made of the main characteristics 
presented by the Periodic Law ; and it may not be amiss to 
indicate here some of the properties of elements and compounds 
which are also periodic in their nature, since it is only by the 
consideration of the largest possible number of allied phenomena 
that we can hope to evolve some hypothesis which may shed 
light upon the causes underlying this mysterious grouping. 

As can be seen from the diagram at the end of this volume, 
there is a regular periodicity in electrochemical character and 
magnetic properties. Ductility, malleability, and brittleness, 
all show similar fluctuations with change in atomic number. 
The spectra of the elements bear a certain relation to the 
Periodic Law. As has been shown above, the colours of 
metallic chlorides exhibit some regularities which can hardly 
be accidental. The refractivities of certain elements have been 
adduced l to show that in their case also there is periodicity. 
Hardness also appears to vary periodically throughout the 
series of elements. 2 The periodicity of compressibility among 
the elements has already been mentioned in this chapter. 
Thermal coefficients of expansion are periodic in their nature 
also, as are thermal and electrical conducting powers. Boiling- 
point and melting-point were shown by Carnelley 3 to be 
periodic both for elements and for compounds. 

From this formidable list, it will be seen that the Periodic 
Law contains material for endless speculation even apart from 
the fundamental problem of its origin. For instance, to choose 
two examples given by Lothar Meyer, 4 it may be asked why 
conductivity for heat and electricity should go hand in hand 
with ductility and malleability; or again, why the empirical 
rule should hold good that all elements possessing atomic 
weights less than forty obey Dulong and Petit's Law when 

1 Cuthbertson and Metcalfe, Phil. Trans., 1907, 207, 135. 

2 For Rydberg's table, see Landolt-Bornstein's Physikaltsch-chemische 

3 Carnelley's papers are to be found in the Phil. Mag., 1879-1885; a 
summary, with references, is given in Chem. News, 1886, 53, 157, 1 69, 183, 

4 Lothar Meyer, Modern Theories of Chemistry, p. 144-5, 1888, 


their density is approximately less than one and a half times 
that of water ? 

Turning to another point, the differences between elements 
on " ascending " and " descending " branches of the atomic 
volume Curve attract attention. In the case of sodium and 
chlorine, the atomic volumes are equal, both elements are 
monovalent; and yet they differ entirely in properties. The 
only difference exhibited by the graph is the difference in 
atomic number (or number of positive charges in the nucleus) ; 
and it seems "probable that some thought might profitably be 
expended in considering how this factor enters into the problem. 

Again, the valencies of the elements in the eighth group of 
Mendeleeff's Table differ among themselves in a peculiar 
manner. Thus iron in the ferrates has a valency of six ; 
cobalt forms an oxide in which the metal is quadrivalent ; and 
nickel also seems to exert a maximum valency of four in its 
oxide. In the series ruthenium, rhodium and palladium the 
metals in the highest oxides obtained appear to be respectively 
octovalent, hexavalent and quadrivalent. Finally, osmium is 
octovalent; iridium is certainly quadrivalent and possibly 
reaches hexavalence ; while platinum exhibits hexavalence. 
The differences seem worthy of further consideration, the more 
so since they form one of the greatest breaks in the symmetry 
of Mendeleeff's grouping. 

Other points of interest will suggest themselves to any one 
who examines Periodic Law in detail ; and it is possible that 
through such studies new light may be thrown upon the subject 
which may enable us to arrange the elements in a better order 
and thus finally be in a position to formulate some definite 
theory of the origin of the Periodic System. 1 

Whatever this theory may be, it will assuredly have to take 
into account the nature of the atomic nucleus, since this part of 
the elemental structure appears to be intimately connected with 
the atomic number and hence with the atomic order of the 
elements. It has been shown, in the foregoing pages, that the 
supposititious symmetry of the Mendeleeff system must be given 
up ; and the new theory, when it is brought forward, will have 
to account for a much looser fabric. It seems not improbable 

1 For suggestions of this type, see Carnelley, Cheni. News, 1886, 53, 197 ; 
Orookes, ibid., 54, 114 ; Trans., 1888, 53, 487. 


that the key to the enigma may finally be discovered by a 
careful examination of the problems suggested by the rare earth 
elements, since they present in an exaggerated form the 
difficulties involved in the relations of the eighth group 

The work of recent years has brought us back to the point 
which Crookes reached in 1886. Until a year or two ago, we 
divided element from element by means of differences in 
chemical properties between the two; but with the newer 
knowledge which we now possess it seems as though we should 
have to accept Crookes' suggestion l and take " not an external 
boundary, but an internal type " as our criterion. In the rare 
earth series, the chemical properties of the elements have so 
many points in common that it is hard to lay down a hard-and- 
fast test of elemental character by purely chemical means ; 
but the X-ray spectra appear to cleave a sharp line between 
one elementary form and its neighbours. Even here, however, 
the division is not a final one^ since the isotopes have identical 
X-ray spectra and yet we know from radioactivity measure- 
ments that there are fundamental differences between the 
members of this class. Even positive ray measurements will 
fail to separate from each other radium-D and the end-product 
of radium-C2, since these have the same atomic weight, although 
their stabilities are totally different. 

Enough has been said to show the complexities before us 
when we attempt to define elemental matter; and it seems 
clear that unless we. get some definite ideas on the subject we 
shall not be able to approach the problem of the Periodic 
Arrangement with any chance of solving it. 

1 Crookes, Trans., 1888, 53, 490. 



FOR close on thirty years the school of Ostwald dominated the 
field of physical chemistry ; and it is now possible to take 
stock of the results attained by their efforts, and to draw a 
comparison between the hopes which appear to have inspired 
them and the reality achieved by their labours. 

In the first place, it may be well to point out that most of 
the epoch-making discoveries had been made before this school 
came upon the scene at all. The Gas Laws, the Atomic Theory, 
the Law of Mass Action, Faraday's Law of Electrolysis, critical 
phenomena, van der Waals' equation, and the Kinetic Theory 
of Gases, to choose but a few examples, were all perfectly 
familiar to students of theoretical chemistry long before the 
name Physical Chemistry became current. 

None the less, about 1887, when the Zeitschrift fur phy- 
sikaUsche Chemie was founded, the physical chemists of the 
time took up an attitude which, in the light of our present 
knowledge, appears rather pathetic. More by their manner of 
stating things, perhaps, than by actual affirmation, they gave 
the impression that physical chemistry, this "new" subject, 
was to regenerate the outworn fabric of the older science. 

What has been the outcome of it all ? 

In the field of electrochemistry, a certain number of ad- 
vances stand to their credit ; but none of these can be placed 
upon a level with the greater generalizations which had already 
been made in the past. Let us take one or two examples of 
their achievements. The genius of van't Hoff evolved a theory 
of solution which agrees wonderfully with the facts within the 
limitations laid down by its author; but that theory has 
little or no application to the solutions which we commonly 
deal with in our test-tubes. Arrhenius cleared up many of our 
difficulties with regard to reactions in solution ; but his theory 
leaves untouched practically the whole of organic chemistry 



After thirty years' incessant work, physical chemistry of the 
old-fashioned school has left us still devoid of any general 
theory of solution. Ostwald has given us a Dilution Law 
which breaks down completely when pressed beyond one single 
group of electrolytes ; and his hypothesis as to indicators 
is chiefly notable on account of the fact that it missed the 
crux of the problem owing, probably, to his ignorance of 
the chemistry of the indicators themselves.* His attempts to 
dispense with the use of the atomic theory are now merely of 
antiquarian interest.! 

The condition into which physical chemistry had drifted 
towards the end of last century cannot be laid to the account 
of either van't Hoff or Arrhenius. Both of these men possessed 
strikingly original minds ; and neither of them confined him- 
self to a narrow groove in science. The leader upon whose 
shoulders the responsibility falls is Ostwald, the propagandist 
of the time. Under his guidance, physical chemistry de- 
generated to a great extent into a means of attacking the 
problems of pure chemistry instead of opening up new fields. 

In seeking the reason for this decline, we have no need to 
travel far. Mathematics is a good servant, but a very bad 
master; and in the closing years of the physical chemistry 
which is now passing away, there can be little doubt as to its 
influence. It is quite true that the more a subject becomes 
amenable to mathematical treatment, the more accurate our 
knowledge of it becomes ; but this is not quite the same thing 
as saying that the scientific value of a paper can be gauged 
from the number of signs of integration which it contains. 
Under the influence of Ostwald, numerical values assumed 
undue importance ; and groups of facts which could not be 
expressed in quantitative terms were regarded in a somewhat 
step-motherly fashion. 

Now, it is clear that, in the majority of cases, the first steps 
in a new field are qualitative rather than quantitative ones ; so 
that by throwing undue importance upon quantitative values 
the breaking of new ground is discouraged to a greater or less 

* This ignorance of organic chemistry appears to be a characteristic of the 
Ostwald school, few of whom have shown any grasp of the problems involved 
in the study of carbon compounds. 

t Even Ostwald himself has had to modify his attitude in this matter. 
See the VorbericM to his Grundriss der allgemeinen Chemie, 4th edition, 1909. 


extent ; and the incentive is given rather to the application of 
known methods to fresh cases. Perhaps it is not going too 
far if we see in this the origin of the sterility of the Ostwald 
school in the wider fields of physical chemistry. 

Another trait of the older school of physical chemists seems 
to have been their dislike of the use of mechanical models. 
As any organic chemist knows, these mechanical conceptions 
are of the utmost value in many cases ; but to the mathe- 
matically-minded physical chemist they appear to suggest 
nothing in particular. This is strange, since physicists of the 
rank of Clerk Maxwell have not despised their assistance. It 
seems probable that this contempt for mechanical illustrations 
accounts to a great extent for the failure of the Ostwald school 
in the field of organic chemistry. Ostwald himself suggested 
that organic compounds might be dealt with from the mathe- 
matical standpoint, and that their behaviour might be expressed, 
not in structural formulae but by mathematical symbols. 1 It 
is safe to say, however, that no results of any value whatever 
have been produced as yet ; and the mechanical conception 
has justified itself by countless successes while the chemical 
mathematicians have produced nothing in this field which is 
worth the paper on which it was written. Yet the reactions 
of organic chemistry have been open to attack at any time 
within the last generation. 

Lest there should be any misunderstanding it may be said 
quite plainly that the foregoing paragraphs are not intended to 
decry the value of mathematics when properly employed in 
chemical problems ; but are meant merely to indicate that its 
function is a restricted one. Mathematics cannot be dispensed 
with unless we are prepared to abandon the development of 
the atomic theory, which depends upon quantitative relations ; 
but when mathematics becomes anything more than a " brilliant 
second " in the chemical field the results are not likely to 
advance our knowledge. Not only so, but the individual who 
uses mathematics as an adjunct to chemistry must have a real 
understanding of mathematical technique if the results are to 
be of any value. FitzGerald, in his Helmholtz Memorial 
Lecture, 2 gave some examples of the pitfalls into which the 

1 Ostwald, Zeitsch. physikal. Chem., 1908, 61, 507. 

2 FitzGerald, Trans., 1896, 69, 885. 


mathematical smatterer may stumble; and it might be well 
were a wider circulation assured for the aphorism : " It is as 
risky for a chemist to apply mathematics as for a mathematician 
to lecture chemists." 

In yet another direction, the influence of Ostwald upon his 
school has been equally unhappy. One of the most original of 
modern physicists 1 once took the trouble to analyse Ostwald' s 
views on science and, after some trenchant criticism, described 
them as " a sort of well-arranged catalogue of facts without any 
hypotheses . . . worthy of a German who plods by habit and 
instinct." It is doubtful if any unbiassed person would wish 
to enter the Ostwald school after a perusal of this examination 
of Ostwald's fundamental ideas. Hypotheses and even 
theories are, of course, ephemeral things ; they serve the 
needs of their time and may pass away when we gain a fuller 
knowledge of facts. But a school which deliberately prefers to 
exclude hypotheses from its purview is certainly not likely to 
advance true knowledge to any great extent; though by 
drudgery and industry it may gather together a mass of 
experimental details. 

It so happens that recent developments have thrown a flood 
of light upon the relative values of the two methods, considered 
as instruments of progress. Up to the last few years, the 
determination of atomic weights was a branch of chemistry in 
which the Ostwald method reached its culmination, for no 
hypotheses were employed and the facts were arranged, year by 
year, in the annual tables of atomic weights. Something like 
thirty different determinations seem to have been made of the 
atomic weight of lead ; and among the experts there appears to 
have been a general agreement on a certain fixed value for the 
atomic weight. Suddenly, from the disintegration theory and 
especially from Soddy's far-reaching speculations, it became clear 
that the atomic weight of lead was not a physical constant. The 
application of an hypothesis and the subsequent testing of it over- 
threw at one stroke the whole value of previous investigations. 

It might be contended by supporters of the " facts without 
hypothesis " school that, after all, established facts are of 
value in themselves. But the lead problem has shown that 
even this small grain of comfort is more than they are entitled 

1 FitzGerald, Nature, 1896, 53, 441, 


to keep. It is now known that the elements are often mixtures 
of more than one type of atom; so that the whole of the 
laborious researches which have been devoted to " determining 
the atomic weights " prove in the end to be merely attempts to 
define the physical constants of a series of mixtures which may 
or may not be constant ones. What scientific value or even 
interest attaches to the fact that "lead" has an "atomic 
weight " of 207*095, when we now know that " lead " is in all 
probability a mixture of atoms having weights differing from 
each other by about one per cent. 

Fortunately, the Ostwald school of physical chemistry is 
dying out ; and a new and more hopeful change has come over 
the subject within the last few years. Some of the foregoing 
chapters have indicated the lines on which the fresh develop- 
ments are proceeding; and they are only examples of the 
trend in this region of chemistry. We seem at last, after 
thirty years of relative stagnation, to be breaking fresh ground ; 
and it may not be without interest to indicate briefly some of 
the points upon which our knowledge is still very incomplete. 

One of the commonest facts in chemistry is that certain 
liquids will mix with one another whereas others do not 
mingle ; yet our knowledge of the factors which govern these 
cases is very small. It is possible that the phenomena may 
be akin to " wetting " or " non-wetting " ; but this brings us 
very little further forward. 

The problem of non -ionic reactions such as esterification 
provides much material for speculation. The rapidity with 
which ionic reactions take place, even if no precipitation oc- 
curs, is extraordinary when we compare it with the compara- 
tive sluggishness characteristic of many organic reactions. 
The Law of Mass Action applies in both cases ; so that the 
only factor left to us to account for the difference is chemical 
affinity. If we regard organic reactions as totally different 
from non-ionic ones, we are bound to inquire wherein the 
two sets differ; and the problem of the neutralization and 
esterification of the same acid provides a point not without 
interest.* If we assume that in the one case the matter is 
concerned with electrical charges, what are we to postulate in 
the case of organic reactions ? Chemical affinity, according to 
* Possibly we have to do in this case with differences of electric potential. 


modern views, is also electrical in character. Are we, then, to 
assume two different modes of electrical operation ? 

As regards the nature of concentrated solutions, we appear 
to be returning more and more towards the solvate hypothesis ; 
but the whole subject is still in a very obscure condition. 1 

The action of semi-permeable membranes is one which 
raises acute controversy. 2 There are obviously three types of 
semi- permeable membrane : solid membranes, liquid membranes, 
and the boundary surfaces of solutions containing non-volatile 
solutes. No general theory has yet been proposed which 
includes all these cases. 

One obvious factor in the problem appears to have been 
disregarded in modern views upon semi-permeable membranes : 
the velocity of the particles which leave the surface of the 
membrane. In order that a particle may be permanently 
withdrawn from any surface, the particle must attain a 
velocity greater than the critical velocity of the surface in 
question. If its velocity be lower than this critical velocity, 
it will fall back upon the surface. Applying this to the case 
of the semi-permeable membrane, it is evident that molecules 
travelling with a high velocity will escape, whilst those en- 
dowed with lower velocities will be retained.* If we assume 
that solute molecules are always associated with a large 
number of solvent molecules, this furnishes a simple explana- 
tion of the action of semi-permeable membranes in the case of 
solutions ; for the friction upon the surface of the large sol- 
vate mass would lower its velocity considerably. Of course, in 
the case of aqueous solutions, since the solvent itself is associated, 
it is necessary to assume a very considerable association between 
solvent and solute. This agrees with the fact that in con- 
centrated solutions the osmotic pressure is higher than the 
theoretical value; since on the solvate hypothesis much sol- 
vent is employed in forming solvate complexes and is thus 
prevented from acting as a solvent medium. 

The solution problem may be approached from another 

1 For an account of the evidence in favour of the solvate view, see Jones, 
The Nature of Solution (1917). 

2 See Findlay, Osmotic Pressure (1913) ; and the discussion reported in 
Trans. Farad. Soc., 1917, 13, 119. 

* Cf. the speeds of hydrogen and nitrogen molecules in the case of a 
palladium membrane. 


point of view. At ordinary temperatures and pressures, solid 
sodium chloride is below its boiling-point and its vapour 
pressure is small. Yet if we apply water to it, it passes, 
according to the osmotic theory of solution, into something 
akin to the gaseous condition. It is evident from this that by 
applying water to its surface we have produced an effect 
parallel to that which we would bring about if we reduced the 
pressure on the surface of the salt. The external pressure, 
however, is apparently the same as before, if not actually 
greater. How can we explain the effect ? The simplest way 
seems to be to assume that the solvent exerts an attractive 
power upon the solute. Thus in the case of a dissolving sub- 
stance we should have the following forces in action: the 
vapour pressure of the solute, the attraction of the solvent, 
and the osmotic pressure in the solution. The first two factors 
would co-operate and would oppose the third. Saturation 
would occur when the three forces came into equilibrium. 

With regard to the effect of a solvent upon the intra- 
molecular structure of a solute dissolved in it we know very 
little indeed ; and an increase in our knowledge of this field 
might lead to interesting results. The problem is one which, 
from its nature, can hardly be attacked except by physical 
methods, such as an examination of absorption spectra, mag- 
netic rotation, refractivity, magnetic susceptibility or dis- 
persive power. 

These physical methods themselves, however, stand in con- 
siderable need of dispassionate examination. From the purely 
chemical side we have constructed a series of constitutional 
formulae which represent to a great extent the chemical be- 
haviour of compounds. From the physical side we have 
amassed a considerable amount of data. Instead of construct- 
ing models which will account for our physical results, we 
have simply adopted the chemical model in toto ; and have 
thus possibly over-loaded the scale on the chemical side instead 
of holding the balance evenly between the two. 

Among the numerous physical properties which have been 
examined, optical rotatory power is one of those which have 
gained least from our investigations during the last genera- 
tion. We are able to state from an examination of the 
structural formula of a compound whether or not the substance 


can show optical activity ; but beyond that our knowledge 
stops short. We cannot make even a rough guess at the value 
of the rotation of the substance. 

Crystallography is a subject which, until quite recently, 
lay almost outside the purview of the average chemist, whose 
views on crystal forms were generally confined to " small, flat 
plates " or " pointed leaflets " or some such description. The 
intricacy of crystal classification and the demands made by the 
subject upon the " three-dimensional imagination " sufficed to 
keep most dilettanti at a distance. 

In recent years, however, Federov has succeeded in reduc- 
ing to a form available for reference the data which have been 
accumulated with regard to many substances ; with the result 
that a new method of identifying compounds has been placed 
in our hands. Federov's index contains particulars of some ten 
thousand substances and in order rapidly to identify a crystal 
of any of these it is only necessary to take certain measure- 
ments of the crystal in question, an operation occupying not 
more than two hours. The reliability of Federov's method was 
shown by his identification of forty-eight out of fifty com- 
pounds which were presented to him unlabelled and which he 
was able to recognize by crystallographic measurements alone. 

The usefulness of this new method of identification, espe- 
cially in cases where only small quantities of material are 
available, requires no advertisement here. In its rapidity it 
far surpasses ordinary analytical processes ; and it appears to 
yield results which, if they do not absolutely establish the 
identity of a compound, at least render its recognition a com- 
paratively easy matter. 1 

Turning to the properties of the elements, we find that 
many of them have recently been detected in a display of what 
may conveniently be termed abnormal valency. The old and 
apparently well-tried dogma of the quadrivalence of carbon has 
sustained a severe shock by the discovery of a large number of 
compounds in which carbon atoms behave as if they were tri- 
valent. Triphenylmethyl, (CeH^C, is the best-known case, 
but there are many others in which carbon atoms show an 
abnormal behaviour. Again, in the hydrides of boron the 

1 Particulars of Federov's method are given in Ann. Reports. 1912, 9, 261 ; 
1913, 10, 245 ; 1914, 11, 248 ; 1917, 14, 230. 



element acts as if it were quadrivalent. Divalent and quadri- 
valent nitrogen atoms have been detected in certain compounds. 
Sulphur and oxygen appear to act in certain cases as though 
they were monovalent; and the lead analogue of triphenyl- 
methyl has been isolated. 1 

The whole problem of valency, 2 however, bristles with 
difficulties ; and we still await some hypothesis which will 
bring complete illumination upon it. 

Why, for example, since we get PC1 3 and PH 3 , should we 
not expect to produce PH 5 just as easily as PC1 5 ? Why is 
PH 4 C1 stable when PHC14 has not been isolated ? How comes 
it that when sulphur combines with four chlorine atoms the 
resulting compound is not salt-like ; whereas the other form 
of quadrivalent sulphur, in the sulphonium salts, yields ions ? 
Why is H 2 S acidic in solution whilst NH 3 is basic and B 2 He is 
unstable under the same conditions ? Is there any connection 
between these facts and the capacity of ammonia to yield 
NH 4 C1 as well as N(C 2 H 5 ) 4 C1, whilst sulphur will give rise to 
(C 2 H 5 ) 3 S . Cl but not to H 3 S . 01 ? How can we explain the 
fact that while NH 3 , NC1 3 , and BC1 3 are known, the hydrogen 
compound of boron exists as B 2 H 6 ? What forces retain the 
iron, carbon, and nitrogen atoms in the ferrocyanide ion and 
what relation do these forces bear to ordinary valency ? Why, 
when a hydrogen atom is directly replaceable by a metallic 
atom, is the atom to which the hydrogen is attached always 
in a valency stage lower than that of its maximum valency ? 
Many such questions will suggest themselves ; and yet we have 
no clear answers to them. 

Enough has now been said to show that there is much fresh 
ground still awaiting investigation even among the commonest 
facts of chemistry ; and if the foregoing paragraphs have sug- 
gested anything of interest to the reader, they have amply 
fulfilled the object with which they were written. 

1 See Stewart, Recent Advances in Organic Chemistry (4th edition), for 
an account of these cases. 

2 See Friend, The Theory of Valency; also Loring, Studies in Valency; 
Martin, Researches on the Affinities of the Elements ; Nelson and Falk, School 
of Mines Quarterly, 1900, 30, 179; J. Amer. Chem. Soc., 1915, 37, 274 ; Nelson, 
Beans, and Falk, ibid., 1913, 35, 1810; Falk and Nelson, ibid., 1910, 32, 1637 ; 
1911, 33, 1140; Falk, ibid., 1912, 34, 1041; Noyes, ibid., 1912, 34, 663; Fry, 
ibid., 1912, 34, 664; 1914, 36, 248, 262, 1035; 1915, 37, 885; 1916, 38, 1323, 
1327, 1333 ; Falk and Nelson, Science, 1917, 46, 551. 




THE model of the Periodic Surface is most easily constructed in the 
following way. Plot the Atomic Volume Curve on squared paper, 
placing the elements at a distance of one inch from each other, and 
making ten units of atomic volume equal to an inch and a half. 
Divide the paper at the points occupied by the following elements : 
hydrogen, helium, beryllium, neon, aluminium, argon, krypton, 
xenon, europium, ytterbium,* niton. Paste these various sections 
upon three-ply wood (or fret-wood), each piece of wood being 
seventeen inches in length. In the case of hydrogen and helium, 
leave sixteen inches in the centre of the wood slip ; leave ten inches 
between beryllium and boron ; and ten inches between aluminium 
and silicon ; leave two inches blank before gadolinium and lutecium, 
nine inches blank after europium and ytterbium, and twelve inches 
blank after uranium. 

Now plot on squared paper the atomic volumes of hydrogen, 
lithium, sodium, potassium, rubidium, caesium, leaving an inch and 
a half between each element and reckoning, as before, that an inch 
and a half represents ten units of atomic volume. Four and a half 
inches beyond caesium, introduce a point to represent the undis- 
covered alkali metal, No. 87, which should be given the volume 
which seems best by extrapolation. Repeat the same process in the 
case of the elements helium, neon, argon, krypton, and xenon. 
Four and a half inches beyond xenon, insert the appropriate point 
for niton, and one and a half inches beyond niton, insert a point for 
No. 104. Paste these two graphs on wood also. 

The various sections are now cut out by means of a fretsaw. In 
the case of the gaps between hydrogen and helium, beryllium and 

* The volumes of most of the rare earths must be filled in as best one can 
as very few of them have been experimentally determined. 



boron, etc., leave a small tongue of wood, so as to make each series 
a complete piece. 

Set up the two end-pieces containing the alkali group and the 
inert gas group and connect them at the appropriate point with the 
longer pieces, each of which is a section of the curve. Fasten them 
all together with sprigs or fine tacks. 

The box-shaped arrangement thus produced is then fastened to a 
base-board seventeen inches by thirteen and a half, which it just 

Plaster of Paris is now poured into all the sections until it just 
overflows the upright divisions. The modelling of the higher parts 
of the curve is best done with material which has just begun to set. 
This method makes the model rather heavy to handle, and if light- 
ness is required, the lower parts of the spaces can be filled up with 
papier mache made from macerated newspaper with a finish of plaster 
of Paris on the top ; but the moist papier mache is apt to warp the 
wooden divisions. While the plaster of Paris is drying, it can be 
smoothed down with a putty-knife, and the curves of the surface 

The whole model is now coated with white paint ; and black 
paint is applied to those spaces in which no elements occur. A blue 
line can be painted in on the surface to divide the non-metallic 
elements from the others ; it should lie as shown by the dotted line 
on p. 260 of this volume. The region of elements yielding coloured 
salts can be indicated either by drawing a red line round it, as shown 
by the heavy line on p. 260, or by washing the whole region with 
a faint tint. The former method is better, as it permits the colour- 
ing of each square with the colour of the corresponding chloride, 
thus bringing out the relation between the salt-colours in the system. 

Finally, pins are inserted in the fretwood divisions at the proper 
points to represent the elements, and labels are glued to the pins. 
These labels are best prepared by punching discs from thin mill- 
board, by means of a large cork-borer and a hammer. The symbols 
of the elements are written on the discs with Indian ink. 

In order to illustrate the physical properties of the families, a 
glass case is made to contain the model, and small labels are pasted 
on the inside of the top at the appropriate places, to indicate whether 
the elements are malleable or brittle, fusible or refractory, etc. 



Acker, 13, 14 

Adams, 241 

Allen, 121, 122 

Ampere Electrical Company, 30 

Andrade, 114, 120 

Arbuthnot, 75 

Arhle, 37 

Armstrong, 60 

Arrhenius, 266 

Ashcroft, 14, 15 

Aston, 214, 217 


Baeyer, 40, 43 

Baker, 56, 164 

Baly, 64, 77, 78 

Barkla, 91, 94 

Bary, 163 

Baur, 254 

Baxter, 211 

Beans, 274 

Beatty, 93 

Becquerel, 154, 155, 163, 164, 165, 167 

Bender, 156 

Benedicks, 253 

Bennet, 50 

Benoist, 95 

Bernsthen, 26 

Berthelot, 29, 46 

Berzelius, 135 

Beyer, 19 

Bielecki, 85 

Biltz, 254 

Birkeland, 23 

Blumenfeld, 133 

Bohr, 117, 226 

Bose, 46 

Bosshard, 51 

Boyle, 179, 236 

Bragg, 102, 106, 113 

Branch, 227 

Brauner, 46, 253, 254 

Bray, 227 

Bredig, 38 

Broek, 117, 227 

Broglie, 114 
Brooks, 179 
Briihl, 38, 72 
Bruhat, 50; 
Biichner, 174 
Bumstead, 179 
Bunsen, 29 

CAMERON, 183, 191, 192 

Campbell, 174 

Carnelly, 251, 263, 264 

Caro, 29 

Castner, 13, 15 

Chancourtois, 250 

Chapman, 92, 93, 217 

Collie, 111, 149, 151, 182, 193, 195, 

196, 213 
Collier, 94 
Colson, 31 
Comte, 56 
Constam, 47, 50 
Cowles, 1 

Cranston, 173, 174 
Crookes, 18, 19, 90, 141, 154, 162, 167, 

207, 208, 214, 236, 238, 250, 261, 

262, 264, 265 
Crowther, 141, 147 
Crymble, 70, 73, 75, 78, 82 
Curie, M., 211 
Curie, Mme., 156, 158, 160, 163, 164, 

185, 192 
Curie, P., 96, 157, 164, 179, 182/185, 

Cuthbertson, 263 

DANNE, 179 

d'Ans, 46, 47 

Debierne, 158, 178, 182, 185, 188 

Desch, 64, 78 

Desfosses, 29 

Deville, 31 

Dewar, 188 

Dobbie, 18, 74, 83, 84 

Domcke, 56 

Donkin, 24 

Dony-Henault, 37 


2 7 8 


Dora, 172, 179 
Duane, 151 
Dubois, 50 

EBLEB, 156, 174 
Edminson, 73 
Egerton, 195 
Elod, 56 
Engelhardt, 70 
Ercolini, 45 
Exner, 200 
Eyde, 23 

FAJANS, 204, 205 

Falk, 274 

Faraday, 266 

Federov, 273 

Findlay, 271 

Finkelstein, 224 

FitzGerald, 268, 269 

Fleck, 201, 202, 203, 204, 205, 208, 

217, 231 
Foerster, 19 
Fowler, 55 
Fownes, 29 
French, 30 
Friederich, 46, 47 
Friedrich, 102, 105 
Friend, 124, 128, 132, 274 
Friman, 114 
Fry, 273 


Garrett, 241 

Geiger, 156, 175, 176, 226, 235 

Giesel, 183, 188 

Gimingham, 178 

Girod, 2 

Glasson, 93 

Gleditsch, 192 

Glendinning, 73 

Goldstein, 90 

Gran, 19 

Grant, 181 

Gray, J. A., 161 

Gray, B. W., see Whytlaw-Gray. 

Grover, 211 

Gruszkiewic?, 29 

Guldberg, 19 

HABEE, 19, 26, 28 

Hackh, 241 

Halm, 173 

Hall, 241, 250, 251, 252 

Hansen, 47 

Hardy, 164, 165 

Harkins, 116, 211, 227, 241, 250, 251, 

Hartley, Sir N., 64, 66, 67, 70, 74, 84, 

Hartley, 217 

Haschek, 200 

Hauenstein, 43 

Haughton, 241 

Hemptinne, 24, 25 

Henri, 85-87 

Henriot, 174 

Heroult, 2 

Herringa, 46 

Herweg, 114 

Hevesy, 199, 200, 204 

Hicks, 254 

Hilditch, 73 

Himstedt, 188 

Hinsberg, 196 

Hittorf, 224 

Hofer, 19 

Hoff, van't, 266, 267 

Holmes, 190 

Honigschmid, 158, 210, 211 

Hopkins, 241 

Horovitz, 211 

Horry, 7 

Howies, 19, 22 

Hoyermann, 29 

Hulin-Schumann, 37 

Hyman, 210, 212 

James, 131 
Jantsch, 133 
Jaubert, 50 
Joly, 217 
Jones, 271 

KAILAN, 165 
Kaiser, 28 
Kaye, 91, 93, 112 
Kelvin, 225 
Kennedy, 174 
Kingzett, 38 
Klemensiewiez, 200 
Knipping, 102, 105 
Knox, 18, 19, 27 
Koenig, 19, 56 
KcBnigsberger, 224 
Korosy, 165 
Kuzma, 46 

Lambilly, 25, 26 
Landauer, 152 
Lantsberry, 196 
Lauder, 83, 84 
Laue, 102, 105 
Le Alanc, 19 
Lee, 19 

Lembert, 211, 212 
Le Rossignol, 26, 178 
Levi, 45 
Levin, 174 
Lewis, G. N., 227 



Ley, 70 

Lind, 165 

Lindemann, 114, 176, 217 

Lindt, 151 

Little, 124, 128, 132 

Loew, 251 

Loring, 241, 274 

Lowry, 58 

Luther, 46 

Lyman, 121 

MACBETH, 70, 76 

MacDougall, 19, 22 

Mackenzie, 195 

McLennan, 174 

Magini, 74 

Makower, 156, 179 

Marsden, 196, 226 

Marshall, 42, 45 

Martin, 15, 17, 18, 133, 241, 274 

Massini, 47 

Masson, 195, 196 

Maxwell, 268 

Mayer, 225 

Mehner, 31 

Meitner, 173 

Melikoff, 39, 41, 50 

Mendeleef, 236, 237, 239, 240, 242, 

243, 244, 251, 252, 253, 255, 264 
Merck, 37 
Merton, 195 
Metcalfe, 263 

Meyer, L., 236, 237, 245, 246, 251, 263 
Meyer, 188, 254 
Migliorini, 45 
Millikan, 225 
Moissau, 1, 2, 4 
Moore, 97 
Moseley, 114-116, 118-120, 123, 255, 


Mulder, 46 
Miiller, 224 
Muthmann, 19, 126 

NELSON, 274 
Nernst, 19, 22 
Nicholson, 227 
Norton, 17 
Noyes, 227, 274 
Niiranen, 19 
Nuttall, 175, 176, 235 

O'NEILL, 29 

Oordt, 26, 28 
Ostwald, 266-70 
Owen, 92 
Owens, 177 

PANETH, 199, 200, 216 
Pascal, 73 

Patterson, H. S., 149, 151, 193, 195, 

196, 213 
Pauling, 23 
Peltner, 48 
Perkin, Sir W. H.,73 
Perman, 164, 192 
Pichou, 1 
Pickering, 182 
Pissarjewsky, 39, 41. 50 
Playfair, 29 
Pokorny, 46 
Poole, 217 
Pouzenc, 50 
Precht, 158 
Price, 42, 45 
Pring, 15 
Prout, 214, 236 

RAMSAY, 24, 157, 158, 179, 180, 181, 

182, 183, 184, 187, 188, 191, 192, 

197, 208, 232 
Baschig, 46 
Eayleigh, 19, 251 
Rea, 73 
Regnault, 24 
Retgers, 254 
Reyehler, 254 
Reynolds, 250 
Richards, 211, 212, 242 
Richardson, 225 
Rideal, 3, 15 
Riedel, 34 
Rillet, 69 
Rontgen, 154 
Rossi, 200 
Royds, 182, 192 
Rudorf, 252 
Ruer, 174 
Runge, 158 
Russ, 19 

Russell, 200, 204, 205 
Rutherford, 114, 117, 120, 160, 166, 

168, 177, 179, 182, 185, 186, 187, 

192, 196, 197, 226, 229 
Rydberg, 116, 263 

SADLER, 91, 94 
Sagnac, 96 
Schlundt, 156 
Schlutius, 25 
Schmidlin, 47 
Schmidt, 241 
Schone, 40 
Schonherr, 24 
Schumann, 152 
Schutzenberger, 31 
Scott, 23 
Serpek, 27 
1 Siegbahn, 114, 211 



Siemens, 1 

Smiles, 59, 72, 223 

Soddy, 117, 118, 165, 166, 168, 170, 
171, 173, 174, 176, 177, 182, 183, 
184, 187,^188, 195, 198, 199, 201, 
204, 205, 206, 208, 210, 211, 212, 
215, 217, 219, 220, 231, 240, 241, 
251, 254, 269 

Soret, 69 

Spencer, 124 

Stackelberg, 241 

Stansfield, 3, 15 

Steele, 181, 254 

Steinmetz, 241 

Stenstrom, 211 

Stewart, A. W., 70, 73, 74, 75, 76, 77, 
78, 81, 220, 229-33, 250, 274 

Stewart, J. Q., 233 

Stintzing, 251 

Stoney, 251 

Strong, 174 

Strutt, 53-8, 159, 161, 189, 195 

Sudborough, 164 

Sulc, 46 

TAFEL, 39 

Tanatar, 46, 48, 50 

Tarugi, 45 

Thiele, 71, 72 


Thomson, Sir J. J., 141,145, 148,149, 

150, 151, 152, 195, 213, 225, 226 
Thornton, 241 
Thorpe, 158 
Tiede, 56 
Tinkler, 83, 84 
Turrentine, 45 

UEBAIN, 125, 133 

U.S.A. Bureau of Mines, 156 

Usher, 164, 192 

VAVON, 174 
Vcrneuil, 254 
Villiger, 40, 43 
Vogel, 241, 253 

WAAGE, 19 

Waals, van der, 266 

Wadsworth, 211, 212 

Warburg, 19 

Watson, 46, 182 

Watts, 158 

Wegscheider, 216 

Weiss, 126 

Welsbach, 125, 135 

Wendt, 151-2 

Werner, 254 

Wheeler, 179 

Whiddington, 93, 97 

Whytlaw-Gray, 158, 179, 180, 181 

Willcock, 164 

Willson, 7, 116 

Willstatter, 43 

Wilson, 227 

Witt, 76, 77 

Wohler, 31 

Wolff, 176 

Wolftenstein, 39, 48 

Wright, R, 70, 73, 75, 78, 81 

Wright, 3, 15 

Wyrouboff, 254 

YOUNG, 24 


a-PAETiCLE, 162, 169, 170, 201, 205 
o-ray change, 170, 174, 201-2 
a-rays, 160, 162, 166, 167, 169 

deflection of, 161 

nature of, 162 

scattering of, 225 
Absorption band, 64 

" head " of, 64 

"persistence " of, 65 

Absorption curves, calculation of, 86 

curves, plotting of, 62 ff. 

general, see General Absorption 
- X-rays of, 94 ff. 

selective, see Selective Absorption 

spectra, 59 ff. 

and chemical change, 83 

valency, 81 ff. 
definition of, 60 

Hartley method in, 66 

method of observing, 60 ff. 

spectrum of quinone, 62, 65 
Acetone, spectrum of, 77 
Acetyl-acetone, 78 
Acetylene, 8, 29 

Actinium, 172 
Actinium-A, 174 

-B, 174, 205 
0, 174, 206 

-D, 174, 207 

-X, 174, 198 
Actinium emanation, 174 

series, 174, 178 
Active deposit, 186 

nitrogen, action of, on elements 

and compounds, 54 

effect of external influences on, 


influence of impurities on, 55 

preparation of, 53 

properties of, 54 

spectrum of, 55 

Activity, excited, 185 
Additive properties, 59 

Alkali halides, crystal structure of, 


Alkali manufacture, 13 
Alkyl halides, 70, 78 

Aluminium nitride, 27 
Alundum, 11 

Ammonia, manufacture of, 24 
Ammonium, 122 

hydroperoxide, 39 
. persulphate, 42 

" Antihypo," 49 

" Aquadag," 6 

Arc furnaces, 1, 22-3 

Argon, 214 

Asterium, 119 

Astronomical elements, 120, 227 

Atom, model, see Model Atom 

Atomic core, 229 

nucleus, Moseley on, 118 

numbers and atomic weights, 117 

chemical sequence, 117, 119 

definition of, 117 

positive nucleus, 226 

structure, 225 ff. 

Atomic volume and valency, 243 
Atomic volume curve, 237, 241 ff., 251 
Atomic weight, 219, 220 

weight and X-rays, 92-7 

nature of, 234 

not a constant, 210 

of lead, 209, 270 

-BAY change, 170, 205, 222, 223 
0-rays, 160, 161, 162 

and ionic charges, 208 

deflection of, 161 

velocity of, 162 
Barium carbide, 30 
Barium cyanamide, 30 

cyanide, 30 

hydroperoxide, 39 

peroxide, 40-1 
Becquerel rays, 158 ff. 

Becquerel rays, complexity of, 158 ff . 

effects of, 164 

ionising power of, 164 

physiological effects of, 165 

Beer's Law, 64 
Benzaldoximes, 74 
Benzene system, 76, 111 
Brassiaic acid, 75 




CALCIUM carbide, 6 
Calcium cyanamide, 31 

fluoride crystal, 111 

hydride, 28 

hydroperoxide, 39 

nitrate, 23 

nitride, 28 

permutite, 35 

- silicide, 11 
Camphorquinone, 77 
Canal rays, 80 
Carbides, 6 ff, 30 
Carbon disulphide, 12 
Carborundum, 9 
Caro's acid, 43, 44, 46 

constitution of, 44 

Catalysis in nitrogen fixation, 26, 28 
Cathode rays, action of, on hydrogen, 


nature of, 90 

velocity of, 93 

Celestial elements, 120 

Chalcolite, 156 

Chlorine isotopes, 214 

Chlorosulphonic acid, 43 

Chromium, 224 

Chromogens, 77 

Chromophores, 76 

Citraconic acid, 73, 74 

Cobalt, 224 

Cobalt tri-electride, 232 

Colligative properties, 59 

Conjugated double bonds, 70 ff . 

Constitutive properties, 59 

Copper crystal, 112 

Coronium, 119 

Corpuscular rays, 96 

Cotarnine, 83 

Crookes' tube, 90 

Crystallography, 273 

Crystal structure, analysis of, 98 ff . 

Cyanamide, 30 

Cyanides, manufacture of, 28 

" DAG," 6 

Denitrifying bacteria, 17 
Deposit, active, 186 
Diacetyl, 77 

Diactinic substances, 69 
Diamagnetism, 73 
Diamond crystal, 111 
Diffraction, 98 ff. 

grating, crystal's action as, 104 
Dilution Law, 267 
Dimethyl-diacetylene, 76 
Distintegration, multiple, 169 

of copper, 192 

of nitrogen, 196 

of thorium, 192 

- theory, 167 ff . 
Disintegrations, chart of, 207 

Double bonds, conjugated, 72 
Dowson gas, 25, 29 

ELECTRIC furnace, 1 ff., 22 ff., 27, 30 

arc type, 1 

construction, 3 

for alkali manufacture, 13 

alundum, 11 

calcium carbide, 6 

carbon disulphide, 12 

carborundum, 9 

fixation of nitrogen, 22 ff ., 27, 


graphite, 5 

- phosphorus, 12 

silicides, 10 

silicon, 8 

carbide, 9 

essentials of, 3 

ingot type of, 7 

linings, 4 

resistance type, 2 

rotary type of, 5, 7 

tapping type of, 7 

types of 1, ff . 

Electrides, 232 

Electromotive force of atoms, 261 

Electron as an element, 232 

Electronic theories of atom, 225 

Elementary groups, 238 

Elements, analogy and identity of, 208 

celestial," 120, 227 

homogeneous and heterogeneous, 


interpolated, 240 

passivity of, 224 

" pure " and " mixed," 216 
Emanations, see Actinium, Thorium, 

and Niton 
End-products of radioactive change, 

171-3, 210, 216 
Equilibrium, radioactive, 169 
Erucic acid, 75 
Ether, luminiferous, 226 
Ethyl iodide, 78-9 
Ethylene iodide, 78-9 
Excited activity, 185 


Fixation of nitrogen, 16 ff . 

Fluorescence produced by Becquerel 
rays, 163 

Fractional crystallisation, 131 

Fumaric acid, 74 

Furnace, electric, see Electric Fur- 

7-RAYS, 160 ff. 

nature of, 161 

origin of, 161 
Geiger-Nuttall relation, 175, 235 



General absorption and conjugation, 
70 fi. 

unsaturation, 70 ff . 

factors affecting, 68 ff . 

of cyclic compounds, 70 

homologues, 69 

isomers, 70 ff . 

Geology and radioactivity, 189 
Goldstein rays, 80 
Graphite electric furnace, 5 
Grating, diffraction, see Diffraction 

" Gredag," 6 


Hartley method, 66, 85 
Helium, 162, 169, 175, 187-9, 192, 
193-5, 213, 214 
vacuum tubes, 193 ff . 

produced from niton, 187 

Collis-Patterson, work on, 193 

triatomic, 149 ff . 
Hetero baric elements, 220 
Heterotopic elements, 220 
Hexatriene, 76 
Hydrocyanic acid, 29 
Hydrogen electride, 232 

peroxide, constitution of, 38 
preparation of, 37 

properties of, 38 

spectrum, 120 
Hydroperoxides, 39 
Hydroxycarbanil, 66 ff. 

INCANDESCENT mantle, 135 ff. 

- branding, 138 
burning, 140 

collodionising, 140 

fabrics, 138 
- fixing, 138 

history of, 135 
impregnation, 138 

- testing, 140 
Indicators, theory of, 267 
Indium salts, 222 

Indium's place in Periodic Table, 96 

Interpolated elements, 240 

Iodine, 78-9, 81 

lodoform, 78-9 

Ionic and non-ionic reactions, 270 

charges and -rays, 208 
Ionium, 172, 198, 199, 206 

spectrum of, 200 

Ions, absorptive power of, 82 
Iron, ferrous and ferric, 123, 221 ff. 

E.M.F. of, 224 
Isobarbituric acid, 87 
Isobares, 218 ff., 233 

definition of, 219 

Isobaric elements, see Isobares 
Isotopes, 120, 198 ff., 218-9, 235, 238 

among common elements, 212 

atomic numbers of, 120 
volume of, 212 

definition of, 199 

melting-points, 212 

separability of, 212, 217 

spectra of, 120, 200, 211 

theory of, 198 

of hydrogen, 120 
Isotope theory, tests of, 200 
Isotopy, see Isotopes 
Itaconic acid, 73 

Kanalstrahlen, 80 
Krypton's isotopes, 214 


Lactam and lactim forms, 66 

Lattice, see Space lattice 

Laue's experiment, 105 

Law, Periodic, see Periodic Law and 

Periodic Table 
Lead, 120, 189, 209, 270 

atomic weight of, 209, 270 

and age of rocks, 189 
Lightning, spectrum of, 182 
Lyman spectrum of hydrogen, 121 

Magnesium di-electride, 232 

nitrides, 32 
Magnetic rotation, 73 
Magnets, Mayer's, 225 
Maleic acid, 74 
Manganese permutite, 35 
Mantles, incandescent, see IncpJides- 

cent mantles 

Mass and electrical charge of atoms, 
208, 234 

of atoms and place in Periodic 

Table, 208, 215 
Mass-spectra, 141 
Membrane, palladium, 271 
Membranes, semipermeable, 271 
Mercury isotopes, 214 
Mesaconic acid, 74 
Mesothorium-1, 171, 221 ff. 

-2, 171, 221 ff. 

Meta-elements, 207, 214, 236, 261 ff. 
Metaneon, 213 

Methyl iodide, 78-9 

Methylene iodide, 78-9 

Microbalance, 181 

Minerals, age of, 189 

Model atoms, 225 ff. 

Model atom and radioactivity, 233 

rare earths, 231 

Molecular numbers, 121 

2 8 4 


Monazite sand, treatment of, 136 ff . 

Monox, 10 

Multiple disintegration, 169 


Neon, 192, 195, 213 

Net, 104 

Nickel, 224 

Niton, 172, 179 ff ., 191-2, 201 

action on water, 183 

atomic weight of, 180 

disintegration of, 184 

heat emitted by, 182 

kinship with inert gases, 182 

liquefaction of, 180 

place in Periodic Table, 181 

properties of, 179 

spectrum of, 182 
Nitre beds, 17 

Nitric acid, manufacture of, 18 ff . 
Nitrides, manufacture of, 31 
Nitrifying bacteria, 17 
Nitrogen, active form of, see Active 

a third form of, 58 

circulation in nature, 16 

disintegration of, 196 

as ammonia, 24 

cyanides, 28 

nitrides, .31 

oxides, etc., 18 

Nitrolim, 31 

Nitrous acid, manufacture of, 18 ft. 
Non-ionic reactions, 270 

" OIL-DAG," 6 
Opacity, specific, 96 
Optical rotatory power, 73. 
Osmotic pressure, 271-2 

PALLADIUM membrane, 271 
Partial valencies, 71 
Passivity, 224 
Per-acids, 37 ff. 

definition of, 41 

elements forming, 41 
general character of, 41 
Perborates, 50 ff. 
Percarbonates, 47 ff . 
Percolumbic acid, 47 
Perdisulphuric acid, 41, 43-5 
Periodic Law, 236 ff. 

properties of elements, 264 

Spirals, 250 

Surface, 251 ff. 

Table and radio-elements, 198 ff., 


anomalies in, 119, 238 

factors governing elements, 

places in, 208, 215 
gaps in, 118 

Periodic Table, new form of, 260 
Permonosulphuric acid, 41, 43-5 
Permutites, 33 ff . 

in compound preparation, 36 

in water purification, 34 ff. 

removal of alkali by, 35 

iron by, 35 

Pernitric acid, 46 
Peroxide of hydrogen, 37 
Peroxides, metallic, 39 
Perphosphoric acid, 46 
Persulphates, 42 ff. 
Persulphuric acids, 42 ff . 
Pertantalic acid, 47 
Pervanadic acid, 47 
Phosphorus, manufacture of, 12 
Pitchblende, 156, 211 
Positive ray analyser, 141-2 
analysis, 141 ff. 

example of, 147 

. theory of, 143 

photographs, 145 ff . 

rays, 90, 141-153 

Potassium halides, crystal structure 

of, 108 ff. 
perborate, 50 

percarbonate, 47-49 

persulphate, 42 

radioactivity of, 174 

tri-iodide, 78-9 
Precipitation, fractional, 130 
Proto-elements, 120, 227 
Protofluorine, 227 
Protyle, 236 
Pseudo-isobarism, 224, 233 

QUAETZ balance, 181 

EADIO ACTINIUM, 174, 198, 206 
Radioactive change, end products of, 
171-173, 210, 216 

constants, 171 ff. 

emanations, see Actinium, 
Thorium, and Niton 

equilibrium, 169 

recoil, 174 

series, 170 ff. 

Radioactivity, disintegration theory 
of, 166 

history of, 154 

possible cause of, 233 

and geology, 189 
Radio-elements and Periodic Table, 

205 ff. 

chemistry of, 202 
Radio-lead, 186 

Radiothorium, 171, 198, 199, 223 
Radium, 155 ff., 172, 185, 198, 201, 

210, 284, 235 

atomic weight of, 158 

chemical effects of, 164 



Radium, constants of, 172 

discovery of, 156 

emanation, see Niton 

metallic, 158 

physiological effects of, 165 

purification of salts of, 156, 157 

salts, properties of, 157 
series, 172 
Radium-A, 172, 175, 186, 201 

-B, 120, 172, 175, 186, 202 

-C, 172, 186, 202, 215 

-C x , 172, 216 

-C 2 , 173, 210, 216 

-D, 172, 186, 216, 223 

-E, 172, 186, 223 

-F, 172, 186, 210, 223 
Range of a particle, 176 

Rare earth elements, 123, 124 ff., 
253 ff. 

and model atom, 231 

applications of, 133 

atomic numbers of, 125 

basicity of, 127 

carbides of, 127 

chemical character of, 126 

classes of, 124-5 

colours of chlorides, 257 

double sulphates of, 129 
fractional crystallisation of, 

precipitation of, 130 

hydrides of, 127 
nitrides of, 127 

number of, 118 

oxides of, 127 

physical properties of, 126, 


purification of, 128 

sources of, 125 

their place in Periodic Table, 

123, 252 ff. 

Rays, a-, -, 7- and x> see under o, 
j8, etc. 

cathode, see Cathode rays. 

corpuscular, 96 

positive, see Positive rays. 
Reactions, ionic and non-ionic, 270 
Recoil efficiency, 175 

radioactive, 174 

Refractive index, 73 

Refractory materials, 4 

Residual affinity, 89 

Resistance furnaces, 2 

Rontgen rays and Becquerel rays, 

161, 163 
Row, 104 
Rubidium, radioactivity of, 174 

SCATTERING of a-particles, 225 
Selective absorption, factors affecting, 
76 ff. 

Semipermeable membranes, 271 
Series, radioactive, see Radioactive 

Silicides, 10 
Silicon, 8 

carbide, 9 

nitrides, 31 
Siloxicon, 10 
Silundun, 9 

" Slice," 104 

Sodium, manufacture of, 14 
Sodium chloride, crystal structure of, 

hydroperoxide, 39 

percorate, 50 

permodocarbonate, 48 

permutite, 34-5 

peroxide, 40, 50 

persulphate, 43 
Soddy's Law, 203 
Solution, nature of, 271, 272 
Solvent action, 80 

Space lattice, definition of, 103 
Spectra, absorption, see Absorption 

mass, see Positive ray 
- of isotopes, 120, 200, 211 
Spectrometer, X-ray, 106 
Spectrophotometer, 84 ff. 
Spectroscope, 60 
Spectrum of hydrogen, 120 

lightning, 182 

Spinthariscope, 162 

Structure of atom, see Atomic struc- 
ture and Model atoms 
Surface, periodic, 250 ff. 

TANATAR'S salt, 48 
Tartaric acids, 74 
Thorite, 210, 212 

Thorium, 136, 137, 139, 168, 185, 198, 
199, 206, 209, 210 

and uranous salts, 209 

emanation, 177, 185 

series, 171 
Thorium-A, 171 

-B, 171, 202-3 

-C, 171, 202-3 

-O lf 170, 171 

-D, 171 

Thorium-X, 168, 171, 198 
Transmutation, 191 ff . 
Transparency, equivalent, 96 
Triketopentane, 77 
Triphenylmethyl, 273 

UNSATUBATION and absorptive power, 

70 ff. 

Uraninite, 211 
Uranium, 154, 155, 167, 172, 173, 189, 

205, see also Uranium- 1 



Uranium and age of rocks, 189 


-1, 172, 173, 209 

-2, 172, 173, 204, 205, 208', 209, 223 

-X, 167 

-X,, 172, 173, '205, 208, 209, 223 
X 2 , 172, 173, 204, 223 

-Y, 173, 174 

Z, 173, 174 

Uranium-radium series, 172 
Uranous salts and thorium salts, 208, 

Uric acid, 87 

VALENCY, abnormal, 273 

and absorption power, 81 ff. 

problems, 273 

WATER purification, 34 
" Wetting," 270 

X 3 , 149 ft. 

X-ray and cathode ray energies, 93 

X-ray spectra, 106, 114 ff., 215, 216, 
239, 255, 265 

spectra and atomic numbers, 114 ff . 

spectra of isotopes, 120 

spectrometer, 106 

spectrum of hydrogen, 120 

wave-length determination, 113 
X-rays, 90 ff., 98 ff., 114 ff., 160, 161, 


absorption of, by elements, 94 ff. 

and atomic properties, 90 ff. 

and crystal structure, 98 ff. 

characteristic, 91 ff. 

emission of, by elements, 91 ff. 

" hard " and " soft," 92 

nature of, 91, 92 

penetrating power of, 92 
Xenon, isotopes of, 214 


Zinc blende crystal, 111 









VI 40 


Malleable Brittle 




ATOMIC NUMBERS i 2 3 i s e 7. a 9 ip 

Electro positive man 

Hydroxides soluble in both acid and alkali 
Electro negative i 

lalleable Brittle Malleable Brittle Malleable Brittle Malleable 








Coloured Salts 

12 13 14- 15 16 17 18 19 20 21 22 23 24- 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 

Malleable Brittle Malleable Malleable 


loured Salts 

.Coloured Salts, .Col Salts. 

2 43 4-4 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64- 65 66 67 68 69 70 71 

l I e a b I e 

Brittle Malleable 

/ \ 

Col Salts, 


12 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 J 









I s 

<Dl O 

-Pi O 

CO! <D 

University of Toronto 








Acme Library Card Pocket 

Under Pat. "Ref. Index File"