Chapter XXII: , Note 35) respecting the combination of CuSO{4} (2)

[23] Silver chloride is almost perfectly insoluble in water, but is
somewhat soluble in water containing sodium chloride or
hydrochloric acid, or other chlorides, and many salts, in
solution. Thus at 100°, 100 parts of water saturated with sodium
chloride dissolve 0·4 part of silver chloride. Bromide and iodide
of silver are less soluble in this respect, as also in regard to
other solvents. It should be remarked that _silver chloride
dissolves in solutions of ammonia, potassium cyanide, and of
sodium thiosulphate_, Na_{2}S_{2}O_{3}. Silver bromide is almost
perfectly analogous to the chloride, but silver iodide is nearly
insoluble in a solution of ammonia. Silver chloride even absorbs
dry ammonia gas, forming very unstable ammoniacal compounds. When
heated, these compounds (Vol. I. p. 250, Note 8) evolve the
ammonia, as they also do under the action of all acids. Silver
chloride enters into double decomposition with potassium cyanide,
forming a soluble double cyanide, which we shall presently
describe; it also forms a soluble double salt, NaAgS_{2}O_{3},
with sodium thiosulphate.

Silver chloride offers different modifications in the structure of
its molecule, as is seen in the variations in the consistency of
the precipitate, and in the differences in the action of light
which partially decomposes AgCl (_see_ Note 25). Stas and Carey
Lea investigated this subject, which has a particular importance
in photography, because silver bromide also gives _photo-salts_.
There is still much to be discovered in this respect, since Abney
showed that perfectly dry AgCl placed in a vacuum in the dark is
not in the least acted upon when subsequently exposed to light.

[24] _Silver bromide_ and _iodide_ (which occur as the minerals bromite
and iodite) resemble the chloride in many respects, but the degree
of affinity of silver for iodine is greater than that for chlorine
and bromine, although less heat is evolved (_see_ Note 28 bis).
Deville deduced this fact from a number of experiments. Thus
silver chloride, when treated with hydriodic acid, evolves
hydrochloric acid, and forms silver iodide. Finely-divided silver
easily liberates hydrogen when treated with hydriodic acid; it
produces the same decomposition with hydrochloric acid, but in a
considerably less degree and only on the surface. The difference
between silver chloride and iodide is especially remarkable, since
the formation of the former is attended with a greater contraction
than that of the latter. The volume of AgCl = 26; of chlorine 27,
of silver 10, the sum = 37, hence a contraction has ensued; and in
the formation of silver iodide an expansion takes place, for the
volume of Ag is 10, of I 26, and of AgI 39 instead of 36 (density,
AgCl, 5·59; AgI, 5·67). The atoms of chlorine have united with the
atoms of silver without moving asunder, whilst the atoms of iodine
must have moved apart in combining with the silver. It is
otherwise with respect to the metal; the distance between its
atoms in the metal = 2·2, in silver chloride = 3·0, and in silver
iodide = 3·5; hence its atoms have moved asunder considerably in
both cases. It is also very remarkable, as Fizeau observed, that
the density of silver iodide increases with a rise of
temperature--that is, a contraction takes place when it is heated
and an expansion when it is cooled.

In order to explain the fact that in silver compounds the iodide
is more stable than the chloride and oxide, Professor N. N.
Beketoff, in his 'Researches on the Phenomena of Substitutions'
(Kharkoff, 1865), proposed the following original hypothesis,
which we will give in almost the words of the author:--In the case
of aluminium, the oxide, Al_{2}O_{3}, is more stable than the
chloride, Al_{2}Cl_{6}, and the iodide, Al_{2}I_{6}. In the oxide
the amount of the metal is to the amount of the element combined
with it as 54·8 (Al = 27·3) is to 48, or in the ratio 112 : 100;
for the chloride the ratio is = 25 : 100; for the iodide it = 7 :
100. In the case of silver the oxide (ratio = 1350 : 100) is less
stable than the chloride (ratio = 304 : 100), and the iodide
(ratio of the weight of metal to the weight of the halogen = 85 :
100) is the most stable. From these and similar examples it
follows that the most stable compounds are those in which the
weights of the combined substances are equal. This may be partly
explained by the attraction of similar molecules even after their
having passed into combination with others. This attraction is
proportional to the product of the acting masses. In silver oxide
the attraction of Ag_{2} for Ag_{2} = 216 × 216 = 46,656, and the
attraction of Ag_{2} for O = 216 × 16 = 3,456. The attraction of
like molecules thus counteracts the attraction of the unlike
molecules. The former naturally does not overcome the latter,
otherwise there would be a disruption, but it nevertheless
diminishes the stability. In the case of an equality or proximity
of the magnitude of the combining masses, the attraction of the
like parts will counteract the stability of the compound to the
least extent--in other words, with an inequality of the combined
masses, the molecules have an inclination to return to an
elementary state, to decompose, which does not exist to such an
extent where the combined masses are equal. There is, therefore, a
tendency for large masses to combine with large, and for small
masses to combine with small. Hence Ag_{2}O + 2KI gives K_{2}O +
2AgI. The influence of an equality of masses on the stability is
seen particularly clearly in the effect of a rise of temperature.
Argentic, mercuric, auric and other oxides composed of unequal
masses, are somewhat readily decomposed by heat, whilst the oxides
of the lighter metals (like water) are not so easily decomposed by
heat. Silver chloride and iodide approach the condition of
equality, and are not decomposed by heat. The most stable oxides
under the action of heat are those of magnesium, calcium, silicon,
and aluminium, since they also approach the condition of equality.
For the same reason hydriodic acid decomposes with greater
facility than hydrochloric acid. Chlorine does not act on magnesia
or alumina, but it acts on lime and silver oxide, &c. This is
partially explained by the fact that by considering heat as a mode
of motion, and knowing that the atomic heats of the free elements
are equal, it must be supposed that the amount of the motion of
atoms (their _vis viva_) is equal, and as it is equal to the
product of the mass (atomic weight) into the square of the
velocity, it follows that the greater the combining weight the
smaller will be the square of the velocity, and if the combining
weights be nearly equal, then the velocities also will be nearly
equal. Hence the greater the difference between the weights of the
combined atoms the greater will be the difference between their
velocities. The difference between the velocities will increase
with the temperature, and therefore the temperature of
decomposition will be the sooner attained the greater be the
original difference--that is, the greater the difference of the
weights of the combined substances. The nearer these weights are
to each other, the more analogous the motion of the unlike atoms,
and consequently, the more stable the resultant compound.

The instability of cupric chloride and nitric oxide, the absence
of compounds of fluorine with oxygen, whilst there are compounds
of oxygen with chlorine, the greater stability of the oxygen
compounds of iodine than those of chlorine, the stability of boron
nitride, and the instability of cyanogen, and a number of similar
instances, where, judging from the above argument, one would
expect (owing to the closeness of the atomic weights) a stability,
show that Beketoff's addition to the mechanical theory of chemical
phenomena is still far from sufficient for explaining the true
relations of affinities. Nevertheless, in his mode of explaining
the relative stabilities of compounds, we find an exceedingly
interesting treatment of questions of primary importance. Without
such efforts it would be impossible to generalise the complex data
of experimental knowledge.

_Fluoride of silver_, AgF, is obtained by dissolving Ag_{2}O or
Ag_{2}CO_{3} in hydrofluoric acid. It differs from the other
halogen salts of silver in being soluble in water (1 part of salt
in 0·55 of water). It crystallises from its solution in prisms,
AgFH_{2}O (Marignac), or AgF_{2}H_{2}O (Pfaundler), which lose
their water in vacuo. Güntz (1891), by electrolising a saturated
solution of Ag_{2}F, obtained _polyfluoride of silver_, Ag_{2}F,
which is decomposed by water into AgF + Ag. It is also formed by
the action of a strong solution of AgF upon finely-divided
(precipitated) silver.

Silver chloride may be decomposed, with the separation of silver oxide, by heating it with a solution of an alkali, and if an organic substance be added to the alkali the chloride can easily be reduced o metallic silver, the silver oxide being reduced in the oxidation of the organic substance. Iron, zinc, and many other metals reduce silver chloride in the presence of water. Cuprous and mercurous chlorides and many organic substances are also able to reduce the silver from chloride of silver. This shows the rather easy decomposability of the halogen compounds of silver. Silver iodide is much more stable in this respect than the chloride. The same is also observed with respect to the _action of light_ upon moist AgCl. White silver chloride soon acquires a violet colour when exposed to the action of light, and especially under the direct action of the sun's rays. After being acted upon by light it is no longer entirely soluble in ammonia, but leaves metallic silver undissolved, from which it might be assumed that the action of light consisted in the decomposition of the silver chloride into chlorine and metallic silver and in fact the silver chloride becomes in time darker and darker. Silver bromide and iodide are much more slowly acted on by light, and, according to certain observations, when pure they are even quite unacted on; at least they do not change in weight,[24 bis] so that if they are acted on by light, the change they undergo must be one of a change in the structure of their parts and not of decomposition, as it is in silver chloride. The silver chloride under the action of light changes in weight, which indicates the formation of a volatile product, and the deposition of metallic silver on dissolving in ammonia shows the loss of chlorine. The change does actually occur under the action of light, but the decomposition does not go as far as into chlorine and silver, but only to the formation of a subchloride of silver, Ag_{2}Cl, which is of a brown colour and is easily decomposed into metallic silver and silver chloride, Ag_{2}Cl = AgCl + Ag. This change of the chemical composition and structure of the halogen salts of silver under the action of light forms the basis of _photography_, because the halogen compounds of silver, after having been exposed to light, give a precipitate of finely-divided silver, of a black colour, when treated with reducing agents.[25]

[24 bis] The changes brought about by the action of light necessitate
distinguishing the photo-salts of silver.

[25] In photography these are called 'developers.' The most common
developers are: solutions of ferrous sulphate, pyrogallol, ferrous
oxalate, hydroxylamine, potassium sulphite, hydroquinone (the last
acts particularly well and is very convenient to use), &c. The
chemical processes of photography are of great practical and
theoretical interest; but it would be impossible in this work to
enter into this special branch of chemistry, which has as yet been
very little worked out from a theoretical point of view.
Nevertheless, we will pause to consider certain aspects of this
subject which are of a purely chemical interest, and especially
the facts concerning _subchloride of silver_, Ag_{2}Cl (_see_ Note
19), and the photo-salts (Note 23). There is no doubt that under
the action of light, AgCl becomes darker in colour, decreases in
weight, and probably forms a mixture of AgCl, Ag_{2}Cl, and Ag.
But the isolation of the subchloride has only been recently
accomplished by Güntz by means of the Ag_{2}F, discovered by him
(_see_ Note 24). Many chemists (and among them Hodgkinson) assumed
that an oxychloride of silver was formed by the decomposition of
AgCl under the action of light. Carey Lea's (1889) and A.
Richardson's (1891) experiments showed that the product formed
does not, however, contain any oxygen at all, and the change in
colour produced by the action of light upon AgCl is most probably
due to the formation of Ag_{2}Cl. This substance was isolated by
Güntz (1891) by passing HCl over crystals of Ag_{2}F. He also
obtained Ag_{2}I in a similar manner by passing HI, and Ag_{2}S by
passing H_{2}S over Ag_{2}F. Ag_{2}Cl is best prepared by the
action of phosphorus trichloride upon Ag_{2}F. At the temperature
of its formation Ag_{2}Cl has an easily changeable tint, with
shades of violet red to violet black. Under the action of light a
similar (isomeric) substance is obtained, which splits up into
AgCl + Ag when heated. With potassium cyanide Ag_{2}Cl gives Ag +
AgCN + KCl, whence it is possible to calculate the heat of
formation of Ag_{2}Cl; it = 29·7, whilst the heat of formation of
AgCl = 29·2--_i.e._ the reaction 2AgCl = Ag_{2}Cl + Cl corresponds
to an absorption of 28·7 major calories. If we admit the formation
of such a compound by the action of light, it is evident that the
energy of the light is consumed in the above reaction. Carey Lea
(1892) subjected AgCl, AgBr, and AgI to a pressure (of course in
the dark) of 3,000 atmospheres, and to trituration with water in a
mortar, and observed a change of colour indicating incipient
decomposition, which is facilitated under the action of light by
the molecular currents set up (Lermontoff, Egoroff). The change of
colour of the halogen salts of silver under the action of light,
and their faculty of subsequently giving a visible photographic
image under the action of 'developers,' must now be regarded as
connected with the decomposition of AgX, leading to the formation
of Ag_{2}X, and the different tinted photo-salts must be
considered as systems containing such Ag_{2}X's. Carey Lea
obtained photo-salts of this kind not only by the action of light
but also in many other ways, which we will enumerate to prove that
they contain the products of an incomplete combination of Ag with
the halogens, (for the salts Ag_{2}X must be regarded as such).
The photo-salts have been obtained (1) by the imperfect
chlorination of silver; (2) by the incomplete decomposition of
Ag_{2}O or Ag_{2}CO_{3} by alternately heating and treating with a
halogen acid; (3) by the action of nitric acid or Na_{2}S_{2}O_{3}
upon Ag_{2}Cl; (4) by mixing a solution of AgNO_{3} with the
hydrates of FeO, MnO and CrO, and precipitating by HCl; (5) by the
action of HCl upon the product obtained by the reduction of
citrate of silver in hydrogen (Note 19), and (6) by the action of
milk sugar upon AgNO_{3} together with soda and afterwards
acidulating with HCl. All these reactions should lead to the
formation of products of imperfect combination with the halogens
and give photo-salts of a similar diversity of colour to those
produced by the action of developers upon the halogen salts of
silver after exposure to light.

The insolubility of the halogen compounds of silver forms the basis of many methods used in practical chemistry. Thus by means of this reaction it is possible to obtain salts of other acids from a halogen salt of a given metal, for instance, RCl_{2} + 2AgNO_{3} = R(NO_{3})_{2} + 2AgCl. The formation of the halogen compounds of silver is very frequently used in the investigation of organic substances; for example, if any product of metalepsis containing iodine or chlorine be heated with a silver salt or silver oxide, the silver combines with the halogen and gives a halogen salt, whilst the elements previously combined with the silver replace the halogen. For instance, ethylene dibromide, C_{2}H_{4}Br_{2}, is transformed into ethylene diacetate, C_{2}H_{4}(C_{2}H_{3}O_{2})_{2}, and silver bromide by heating it with silver acetate, 2C_{2}H_{3}O_{2}Ag. The insolubility of the halogen compounds of silver is still more frequently taken advantage of in determining the amount of silver and halogen in a given solution. If it is required, for instance, to determine the quantity of chlorine present in the form of a metallic chloride in a given solution, a solution of silver nitrate is added to it so long as it gives a precipitate. On _shaking or stirring_ the liquid, the silver chloride easily settles in the form of heavy flakes. It is possible in this way to precipitate the whole of the chlorine from a solution, without adding an excess of silver nitrate, since it can be easily seen whether the addition of a fresh quantity of silver nitrate produces a precipitate in the clear liquid. In this manner it is possible to add to a solution containing chlorine, as much silver as is required for its entire precipitation, and to calculate the amount of chlorine previously in solution from the amount of the solution of silver nitrate consumed, if the quantity of silver nitrate in this solution has been previously determined.[25 bis] The atomic proportions and preliminary experiments with a pure salt--for example, with sodium chloride--will give the amount of chlorine from the quantity of silver nitrate. Details of these methods will be found in works on analytical chemistry.[25 tri]

[25 bis] In order to determine when the reaction is at an end, a few
drops of a solution of K_{2}CrO_{4} are added to the solution of
the chloride. Before all the chlorine is precipitated as AgCl, the
precipitate (after shaking) is white (since Ag_{2}CrO_{4} with
2RCl gives 2AgCl); but when all the chlorine is thrown down
Ag_{2}CrO_{4} is formed, which colours the precipitate
reddish-brown. In order to obtain accurate results the liquid
should be neutral to litmus.

[25 tri] _Silver cyanide_, AgCN, is closely analogous to the haloid
salts of silver. It is obtained, in similar manner to silver
chloride, by the addition of potassium cyanide to silver nitrate.
A white precipitate is then formed, which is almost insoluble in
boiling water. It is also, like silver chloride, insoluble in
dilute acids. However, it is dissolved when heated with nitric
acid, and both hydriodic and hydrochloric acids act on it,
converting it into silver chloride and iodide. Alkalis, however,
do not act on silver cyanide, although they act on the other
haloid salts of silver. Ammonia and solutions of the cyanides of
the alkali metals dissolve silver cyanide, as they do the
chloride. In the latter case double cyanides are formed--for
example, KAgC_{2}N_{2}. This salt is obtained in a crystalline
state on evaporating a solution of silver cyanide in potassium
cyanide. It is much more stable than silver cyanide itself. It has
a neutral reaction, does not change in the air, and does not smell
of hydrocyanic acid. Many acids, in acting on a solution of this
double salt, precipitate the insoluble silver cyanide. Metallic
silver dissolves in a solution of potassium cyanide in the
presence of air, with formation of the same double salt and
potassium hydroxide, and when silver chloride dissolves in
potassium cyanide it forms potassium chloride, besides the salt
KAgC_{2}N_{2}. This double salt of silver is used in silver
plating. For this purpose potassium cyanide is added to its
solution, as otherwise silver cyanide, and not metallic silver, is
deposited by the electric current. If two electrodes--one positive
(silver) and the other negative (copper)--be immersed in such a
solution, silver will be deposited upon the latter, and the silver
of the positive electrode will be dissolved by the liquid, which
will thus preserve the same amount of metal in solution as it
originally contained. If instead of the negative electrode a
copper object be taken, well cleaned from all dirt, the silver
will be deposited in an even coating; this, indeed, forms the mode
of _silver plating by the wet method_, which is most often used in
practice. A solution of one part of silver nitrate in 30 to 50
parts of water, and mixed with a sufficient quantity of a solution
of potassium cyanide to redissolve the precipitate of silver
cyanide formed, gives a dull coating of silver, but if twice as
much water be used the same mixture gives a bright coating.

Silver plating in the wet way has now replaced to a considerable
extent the old process of _dry silvering_, because this process,
which consists in dissolving silver in mercury and applying the
amalgam to the surface of the objects, and then vaporising the
mercury, offers the great disadvantage of the poisonous mercury
fumes. Besides these, there is another method of silver plating,
based on the direct displacement of silver from its salts by other
metals--for example, by copper. The copper reduces the silver from
its compounds, and the silver separated is deposited upon the
copper. Thus a solution of silver chloride in sodium thiosulphate
deposits a coating of silver upon a strip of copper immersed in
it. It is best for this purpose to take pure _silver sulphite_.
This is prepared by mixing a solution of silver nitrate with an
excess of ammonia, and adding a saturated solution of sodium
sulphite and then alcohol, which precipitates silver sulphite from
the solution. The latter and its solutions are very easily
decomposed by copper. Metallic iron produces the same
decomposition, and iron and steel articles may be very readily
silver-plated by means of the thiosulphate solution of silver
chloride. Indeed, copper and similar metals may even be
silver-plated by means of silver chloride; if the chloride of
silver, with a small amount of acid, be rubbed upon the surface of
the copper, the latter becomes covered with a coating of silver,
which it has reduced.

Silver plating is not only applicable to metallic objects, but
also to glass, china, &c. Glass is silvered for various
purposes--for example, glass globes silvered internally are used
for ornamentation, and have a mirrored surface. Common
looking-glass silvered upon one side forms a mirror which is
better than the ordinary mercury mirrors, owing to the truer
colours of the image due to the whiteness of the silver. For
optical instruments--for example, telescopes--concave mirrors are
now made of silvered glass, which has first been ground and
polished into the required form. The _silvering of glass_ is based
on the fact that silver which is reduced from certain solutions
deposits itself uniformly in a perfectly homogeneous and
continuous but very thin layer, forming a bright reflecting
surface. Certain organic substances have the property of reducing
silver in this form. The best known among these are certain
aldehydes--for instance, ordinary acetaldehyde, C_{2}H_{4}O, which
easily oxidises in the air and forms acetic acid, C_{2}H_{4}O_{2}.
This oxidation also easily takes place at the expense of silver
oxide, when a certain amount of ammonia is added to the mixture.
The oxide of silver gives up its oxygen to the aldehyde, and the
silver reduced from it is deposited in a metallic state in a
uniform bright coating. The same action is produced by certain
saccharine substances and certain organic acids, such as tartaric
acid, &c.

Accurate experiments, and more especially the _researches of Stas_ at Brussels, show the proportion in which silver reacts with metallic chlorides. These researches have led to the determination of the _combining weights_ of silver, sodium, potassium, chlorine, bromine, iodine, and other elements, and are distinguished for their model exactitude, and we will therefore describe them in some detail. As sodium chloride is the chloride most generally used for the precipitation of silver, since it can most easily be obtained in a pure state, we will here cite the quantitative observations made by Stas for showing the co-relation between the quantities of chloride of sodium and silver which react together. In order to obtain perfectly pure sodium chloride, he took pure rock salt, containing only a small quantity of magnesium and calcium compounds and a small amount of potassium salts. This salt was dissolved in water, and the saturated solution evaporated by boiling. The sodium chloride separated out during the boiling, and the mother liquor containing the impurities was poured off. Alcohol of 65 p.c. strength and platinic chloride were added to the resultant salt, in order to precipitate all the potassium and a certain part of the sodium salts. The resultant alcoholic solution, containing the sodium and platinum chlorides, was then mixed with a solution of pure ammonium chloride in order to remove the platinic chloride. After this precipitation, the solution was evaporated in a platinum retort, and then separate portions of this purified sodium chloride were collected as they crystallised. The same salt was prepared from sodium sulphate, tartrate, nitrate, and from the platinochloride, in order to have sodium chloride prepared by different methods and from different sources, and in this manner ten samples of sodium chloride thus prepared were purified and investigated in their relation to silver. After being dried, weighed quantities of all ten samples of sodium chloride were dissolved in water and mixed with a solution in nitric acid of a weighed quantity of perfectly pure silver. A slightly greater quantity of silver was taken than would be required for the decomposition of the sodium chloride, and when, after pouring in all the silver solution, the silver chloride had settled, the amount of silver remaining in excess was determined by means of a solution of sodium chloride of known strength. This solution of sodium chloride was added so long as it formed a precipitate. In this manner Stas determined how many parts of sodium chloride correspond to 100 parts by weight of silver. The result of ten determinations was that for the entire precipitation of 100 parts of silver, from 54·2060 to 54·2093 parts of sodium chloride were required. The difference is so inconsiderable that it has no perceptible influence on the subsequent calculations. The mean of ten experiments was that 100 parts of silver react with 54·2078 parts of sodium chloride. In order to learn from this the relation between the chlorine and silver, it was necessary to determine the quantity of chlorine contained in 54·2078 parts of sodium chloride, or, what is the same thing, the quantity of chlorine which combines with 100 parts of silver. For this purpose Stas made a series of observations on the quantity of silver chloride obtained from 100 parts of silver. Four syntheses were made by him for this purpose. The first synthesis consisted in the formation of silver chloride by the action of chlorine on silver at a red heat. This experiment showed that 100 parts of silver give 132·841, 132·843 and 132·843 of silver chloride. The second method consisted in dissolving a given quantity of silver in nitric acid and precipitating it by means of gaseous hydrochloric acid passed over the surface of the liquid; the resultant mass was evaporated in the dark to drive off the nitric acid and excess of hydrochloric acid, and the remaining silver chloride was fused first in an atmosphere of hydrochloric acid gas and then in air. In this process the silver chloride was not washed, and therefore there could be no loss from solution. Two experiments made by this method showed that 100 parts of silver give 132·849 and 132·846 parts of silver chloride. A third series of determinations was also made by precipitating a solution of silver nitrate with a certain excess of gaseous hydrochloric acid. The amount of silver chloride obtained was altogether 132·848. Lastly, a fourth determination was made by precipitating dissolved silver with a solution of ammonium chloride, when it was found that a considerable amount of silver (0·3175) had passed into solution in the washing; for 100 parts of silver there was obtained altogether 132·8417 of silver chloride. Thus from the mean of seven determinations it appears that 100 parts of silver give 132·8445 parts of silver chloride--that is, that 32·8445 parts of chlorine are able to combine with 100 parts of silver and with that quantity of sodium which is contained in 54·2078 parts of sodium chloride. These observations show that 32·8445 parts of chlorine combine with 100 parts of silver and with 21·3633 parts of sodium. From these figures expressing the relation between the combining weights of chlorine, silver, and sodium, it would be possible to determine their atomic weights--that is, the combining quantity of these elements with respect to one part by weight of hydrogen or 16 parts of oxygen, if there existed a series of similarly accurate determinations for the reactions between hydrogen or oxygen and one of these elements--chlorine, sodium, or silver. If we determine the quantity of silver chloride which is obtained from silver chlorate, AgClO_{3}, we shall know the relation between the combining weights of silver chloride and oxygen, so that, taking the quantity of oxygen as a constant magnitude, we can learn from this reaction the combining weight of silver chloride, and from the preceding numbers the combining weights of chlorine and silver. For this purpose it was first necessary to obtain pure silver chlorate. This Stas did by acting on silver oxide or carbonate, suspended in water, with gaseous chlorine.[26]

[26] The phenomenon which then takes place is described by Stas as
follows, in a manner which is perfect in its clearness and
accuracy: if silver oxide or carbonate be suspended in water, and
an excess of water saturated with chlorine be added, all the
silver is converted into chloride, just as is the case with oxide
or carbonate of mercury, and the water then contains, besides the
excess of chlorine, only pure hypochlorous acid without the least
trace of chloric or chlorous acid. If a stream of chlorine be
passed into water containing _an excess of silver oxide_ or silver
carbonate while the liquid is continually agitated, the reaction
is the same as the preceding; silver chloride and hypochlorous
acid are formed. But this acid does not long remain in a free
state: it gradually acts on the silver oxide and gives silver
hypochlorite, _i.e._ AgClO. If, after some time, the current of
chlorine be stopped but the shaking continued, the liquid loses
its characteristic odour of hypochlorous acid, while preserving
its energetic decolorising property, because the silver
hypochlorite which is formed is easily soluble in water. In the
presence of an excess of silver oxide this salt can be kept for
several days without decomposition, but it is exceedingly unstable
when no excess of silver oxide or carbonate is present. So long as
the solution of silver hypochlorite is shaken up with the silver
oxide, it preserves its transparency and bleaching property, but
directly it is allowed to stand, and the silver oxide settles, it
becomes rapidly cloudy and deposits large flakes of silver
chloride, so that the black silver oxide which had settled becomes
covered with the white precipitate. The liquid then loses its
bleaching properties and contains silver chlorate, _i.e._
AgClO_{3}, in solution, which has a slightly alkaline reaction,
owing to the presence of a small amount of dissolved oxide. In
this manner the reactions which are consecutively accomplished may
be expressed by the equations:

6Cl_{2} + 3Ag_{2}O + 3H_{2}O = 6AgCl + 6HClO;
6HClO + 3Ag_{2}O = 3H_{2}O + 6AgClO;
6AgClO = 4AgCl + 2AgClO_{3}.

Hence, Stas gives the following method for the preparation of
silver chlorate: A slow current of chlorine is caused to act on
oxide of silver, suspended in water which is kept in a state of
continual agitation. The shaking is continued after the supply of
chlorine has been stopped, in order that the free hypochlorous
acid should pass into silver hypochlorite, and the resultant
solution of the hypochlorite is drawn off from the sediment of the
excess of silver oxide. This solution decomposes spontaneously
into silver chloride and chlorate. The pure silver chlorate,
AgClO_{3}, does not change under the action of light. The salt is
prepared for further use by drying it in dry air at 150°. It is
necessary during drying to prevent the access of any organic
matter; this is done by filtering the air through cotton wool, and
passing it over a layer of red-hot copper oxide.

The decomposition of the silver chlorate thus obtained was accomplished by the action of a solution of sulphurous anhydride on it. The salt was first fused by carefully heating it at 243°. The solution of sulphurous anhydride used was one saturated at 0°. Sulphurous anhydride in dilute solutions is oxidised at the expense of silver chlorate, even at low temperatures, with great ease if the liquid be continually shaken, sulphuric acid and silver chloride being formed: AgClO_{3} + 3SO_{2} + 3H_{2}O = AgCl + 3H_{2}SO_{4}. After decomposition, the resultant liquid was evaporated, and the residue of silver chloride weighed. Thus the process consisted in taking a known weight of silver chlorate, converting it into silver chloride, and determining the weight of the latter. The analysis conducted in this manner gave the following results, which, like the preceding, designate the weight in a vacuum calculated from the weights obtained in air: In the first experiment it appeared that 138·7890 grams of silver chlorate gave 103·9795 parts of silver chloride, and in the second experiment that 259·5287 grains of chlorate gave 194·44515 grams of silver chloride, and after fusion 194·4435 grams. The mean result of both experiments, converted into percentages, shows that 100 parts of silver chlorate contain 74·9205 of silver chloride and 25·0795 parts of oxygen. From this it is possible to calculate the combining weight of silver chloride, because in the decomposition of silver chlorate there are obtained three atoms of oxygen and one molecule of silver chloride: AgClO_{3} = AgCl + 3O. Taking the weight of an atom of oxygen to be 16, we find from the mean result that the equivalent weight of silver chloride is equal to 143·395. Thus if O = 16, AgCl = 143·395, and as the preceding experiments show that silver chloride contains 32·8445 parts of chlorine per 100 parts of silver, the weight of the atom of silver[26 bis] must be 107·94 and that of chlorine 35·45. The weight of the atom of sodium is determined from the fact that 21·3633 parts of sodium chloride combine with 32·8445 parts of chlorine; consequently Na = 23·05. This conclusion, arrived at by the analysis of silver chlorate, was verified by means of the analysis of potassium chlorate by decomposing it by heat and determining the weight of the potassium chloride formed, and also by effecting the same decomposition by igniting the chlorate in a stream of hydrochloric acid. The combining weight of potassium chloride was thus determined, and another series of determinations confirmed the relation between chlorine, potassium, and silver, in the same manner as the relation between sodium, chlorine, and silver was determined above. Consequently, the combining weights of sodium, chlorine, and potassium could be deduced by combining these data with the analysis of silver chlorate and the synthesis of silver chloride. The agreement between the results showed that the determinations made by the last method were perfectly correct, and did not depend in any considerable degree on the methods which were employed in the preceding determinations, as the combining weights of chlorine and silver obtained were the same as before. There was naturally a difference, but so small a one that it undoubtedly depended on the errors incidental to every process of weighing and experiment. The atomic weight of silver was also determined by Stas by means of the synthesis of silver sulphide and the analysis of silver sulphate. The combining weight obtained by this method was 107·920. The synthesis of silver iodide and the analysis of silver iodate gave the figure 107·928. The synthesis of silver bromide with the analysis of silver bromate gave the figure 107·921. The synthesis of silver chloride and the analysis of silver chlorate gave a mean result of 107·937. Hence there is no doubt that the combining weight of silver is at least as much as 107·9--greater than 107·90 and less than 107·95, and probably equal to the mean = 107·92. Stas determined the combining weights of many other elements in this manner, such as lithium, potassium, sodium, bromine, chlorine, iodine, and also nitrogen, for the determination of the amount of silver nitrate obtained from a given amount of silver gives directly the combining weight of nitrogen. Taking that of oxygen as 16, he obtained the following combining weights for these elements: nitrogen 14·04, silver 107·93, chlorine 35·46, bromine 79·95, iodine 126·85, lithium 7·02, sodium 23·04, potassium 39·15. These figures differ slightly from those which are usually employed in chemical investigations. They must be regarded as the result of the best observations, whilst the figures usually used in practical chemistry are only approximate--are, so to speak, round numbers for the atomic weights which differ so little from the exact figures (for instance, for Ag 108 instead of 107·92, for Na 23 instead of 23·04) that in ordinary determinations and calculations the difference falls within the limits of experimental error inseparable from such determinations.

[26 bis] The results given by Stas' determinations have recently
been recalculated and certain corrections have been introduced. We
give in the context the average results of van der Plaats and
Thomsen's calculations, as well as in Table III. neglecting the
doubtful thousandths.

The exhaustive investigations conducted by Stas on the atomic weights of the above-named elements have great significance in the solution of the problem as to whether the atomic weights of the elements can be expressed in whole numbers if the unit taken be the atomic weight of hydrogen. Prout, at the beginning of this century, stated that this was the case, and held that the atomic weights of the elements are multiples of the atomic weight of hydrogen. The subsequent determinations of Berzelius, Penny, Marchand, Marignac, Dumas, and more especially of Stas, proved this conclusion to be untenable; since a whole series of elements proved to have fractional atomic weights--for example, chlorine, about 35·5. On account of this, Marignac and Dumas stated that the atomic weights of the elements are expressed in relation to hydrogen, either by whole numbers or by numbers with simple fractions of the magnitudes 1/2 and 1/4. But Stas's researches refute this supposition also. Even between the combining weight of hydrogen and oxygen, there is not, so far as is yet known, that simple relation which is required by _Prout's hypothesis_,[27] _i.e._, taking O = 16, the atomic weight of hydrogen is equal not to 1 but to a greater number somewhere between 1·002 and 1·008 or mean 1·005. Such a conclusion arrived at by direct experiment cannot but be regarded as having greater weight than Prout's supposition (hypothesis) that the atomic weights of the elements are in multiple proportion to each other, which would give reason for surmising (but not asserting) a complexity of nature in the elements, and their common origin from a single primary material, and for expecting their mutual conversion into each other. All such ideas and hopes must now, thanks more especially to Stas, be placed in a region void of any experimental support whatever, and therefore not subject to the discipline of the positive data of science.

[27] This hypothesis, for the establishment or refutation of which so
many researches have been made, is exceedingly important, and
fully deserves the attention which has been given to it. Indeed,
if it appeared that the atomic weights of all the elements could
be expressed in whole numbers with reference to hydrogen, or if
they at least proved to be commensurable with one another, then it
could be affirmed with confidence that the elements, with all
their diversity, were formed of one material condensed or grouped
in various manners into the stable, and, under known conditions,
undecomposable groups which we call the atoms of the elements. At
first it was supposed that all the elements were nothing else but
condensed hydrogen, but when it appeared that the atomic weights
of the elements could not be expressed in whole numbers in
relation to hydrogen, it was still possible to imagine the
existence of a certain material from which both hydrogen and all
the other elements were formed. If it should transpire that four
atoms of this material form an atom of hydrogen, then the atom of
chlorine would present itself as consisting of 142 atoms of this
substance, the weight of whose atom would be equal to 0·25. But in
this case the atoms of all the elements should be expressed in
whole numbers with respect to the weight of the atom of this
original material. Let us suppose that the atomic weight of this
material is equal to unity, then all the atomic weights should be
expressible in whole numbers relatively to this unit. Thus the
atom of one element, let us suppose, would weigh _m_, and of
another _n_, but, as both _m_ and _n_ must be whole numbers, it
follows that the atomic weights of all the elements would be
commensurable. But it is sufficient to glance over the results
obtained by Stas, and to be assured of their accuracy, especially
for silver, in order to entirely destroy, or at least strongly
undermine, this attractive hypothesis. We must therefore refuse
our assent to the doctrine of the building up from a single
substance of the elements known to us. This hypothesis is not
supported either by any known transformation (for one element has
never been converted into another element), or by the
commensurability of the atomic weights of the elements. Although
the hypothesis of the formation of all the elements from a single
substance (for which Crookes has suggested the name protyle) is
most attractive in its comprehensiveness, it can neither be denied
nor accepted for want of sufficient data. Marignac endeavoured,
however, to overcome Stas's conclusions as to the
incommensurability of the atomic weights by supposing that in his,
as in the determinations of all other observers, there were
unperceived errors which were quite independent of the mode of
observation--for example, silver nitrate might be supposed to be
an unstable substance which changes, under the heatings,
evaporations, and other processes to which it is subjected in the
reactions for the determination of the combining weight of silver.
It might be supposed, for instance, that silver nitrate contains
some impurity which cannot be removed by any means; it might also
be supposed that a portion of the elements of the nitric acid are
disengaged in the evaporation of the solution of silver nitrate
(owing to the decomposing action of water), and in its fusion, and
that we have not to deal with normal silver nitrate, but with a
slightly basic salt, or perhaps an excess of nitric acid which
cannot be removed from the salt. In this case the observed
combining weight will not refer to an actually definite chemical
compound, but to some mixture for which there does not exist any
perfectly exact combining relations. Marignac upholds this
proposition by the fact that the conclusions of Stas and other
observers respecting the combining weights determined with the
greatest exactitude very nearly agree with the proposition of the
commensurability of the atomic weights--for example, the combining
weight of silver was shown to be equal to 107·93, so that it only
differs by 0·08 from the whole number 108, which is generally
accepted for silver. The combining weight of iodine proved to be
equal to 126·85--that is, it differs from 127 by 0·15. The
combining weights of sodium, nitrogen, bromine, chlorine, and
lithium are still nearer to the whole or round numbers which are
generally accepted. But Marignac's proposition will hardly bear
criticism. Indeed if we express the combining weights of the
elements determined by Stas in relation to hydrogen, the
approximation of these weights to whole numbers disappears,
because one part of hydrogen in reality does not combine with 16
parts of oxygen, but with 15·92 parts, and therefore we shall
obtain, taking H = 1, not the above-cited figures, but for silver
107·38, for bromine 79·55, magnitudes which are still further
removed from whole numbers. Besides which, if Marignac's
proposition were true the combining weight of silver determined by
one method--_e.g._ by the analysis of silver chlorate combined
with the synthesis of silver chloride--would not agree well with
the combining weight determined by another method--_e.g._ by means
of the analysis of silver iodate and the synthesis of silver
iodide. If in one case a basic salt could be obtained, in the
other case an acid salt might be obtained. Then the analysis of
the acid salt would give different results from that of the basic
salt. Thus Marignac's arguments cannot serve as a support for the
vindication of Prout's hypothesis.

In conclusion, I think it will not be out of place to cite the
following passage from a paper I read before the Chemical Society
of London in 1889 (Appendix II.), referring to the hypothesis of
the complexity of the elements recognised in chemistry, owing to
the fact that many have endeavoured to apply the periodic law to
the justification of this idea 'dating from a remote antiquity,
when it was found convenient to admit the existence of many gods
but only one matter.'

'When we try to explain the origin of the idea of a unique primary
matter, we easily trace that, in the absence of deductions from
experiment, it derives its origin from the scientifically
philosophical attempt at discovering some kind of unity in the
immense diversity of individualities which we see around. In
classical times such a tendency could only be satisfied by
conceptions about the immaterial world. As to the material world,
our ancestors were compelled to resort to some hypothesis, and
they adopted the idea of unity in the formative material, because
they were not able to evolve the conception of any other possible
unity in order to connect the multifarious relations of matter.
Responding to the same legitimate scientific tendency, natural
science has discovered throughout the universe a unity of plan, a
unity of forces, and a unity of matter; and the convincing
conclusions of modern science compel every one to admit these
kinds of unity. But while we admit unity in many things, we none
the less must also explain the individuality and the apparent
diversity which we cannot fail to trace everywhere. It was said of
old "Give us a fulcrum and it will become easy to displace the
earth." So also we must say, "Give us something that is
individualised, and the apparent diversity will be easily
understood." Otherwise, how could unity result in a multitude.

'After a long and painstaking research, natural science has
discovered the individualities of the chemical elements, and
therefore it is now capable, not only of analysing, but also of
synthesising; it can understand and grasp generality and unity, as
well as the individualised and multifarious. The general and
universal, like time and space, like force and motion, vary
uniformly. The uniform admit of interpolations, revealing every
intermediate phase; but the multitudinous, the
individualised--such as ourselves, or the chemical elements, or
the members of a peculiar periodic function of the elements, or
Dalton's multiple proportions--is characterised in another way. We
see in it--side by side with a general connecting
principle--leaps, breaks of continuity, points which escape from
the analysis of the infinitely small--an absence of complete
intermediate links. Chemistry has found an answer to the question
as to the causes of multitudes, and while retaining the conception
of many elements, all submitted to the discipline of a general
law, it offers an escape from the Indian Nirvana--the absorption
in the universal--replacing it by the individualised. However, the
place for individuality is so limited by the all-grasping,
all-powerful universal, that it is merely a point of support for
the understanding of multitude in unity.'

Among the platinum metals ruthenium, rhodium, and palladium, by their atomic weights and properties, approach silver, just as iron and its analogues (cobalt and nickel) approach copper in all respects. _Gold_ stands in exactly the same position in relation to the heavy platinum metals, osmium, iridium, and platinum, as copper and silver do to the two preceding series. The atomic weight of gold is nearly equal to their atomic weights;[28] it is dense like these metals. It also gives various grades of oxidation, which are feeble, both in a basic and an acid sense. Whilst near to osmium, iridium, and platinum, gold at the same time is able, like copper and silver, to form compounds which answer to the type RX--that is, oxides of the composition R_{2}O. Cuprous chloride, CuCl, silver chloride, AgCl, and aurous chloride, AuCl, are substances which are very much alike in their physical and chemical properties.[28 bis] They are insoluble in water, but dissolve in hydrochloric acid and ammonia, in potassium cyanide, sodium thiosulphate, &c. Just as copper forms a link between the iron metals and zinc, and as silver unites the light platinum metals with cadmium, so also gold presents a transition from the heavy platinum metals to mercury. Copper gives saline compounds of the types CuX and CuX_{2}, silver of the type AgX, whilst gold, besides compounds of the type AuX, very easily and most frequently forms those of the type AuCl_{3}. The compounds of this type frequently pass into those of the lower type, just as PtX_{4} passes into PtX_{2}, and the same is observable in the elements which, in their atomic weights, follow gold. Mercury gives HgX_{2} and HgX, thallium gives TlX_{3} and TlX, lead gives PbX_{4} and PbX_{2}. On the other hand, gold in a qualitative respect differs from silver and copper in the _extreme ease_ with which all its compounds are _reduced to metal_ by many means. This is not only accomplished by many reducing agents, but also by the action of heat. Thus its chlorides and oxides lose their chlorine and oxygen when heated, and, if the temperature be sufficiently high, these elements are entirely expelled and metallic gold alone remains. Its compounds, therefore, act as oxidising agents.[29]

[28] It might be expected from the periodic law and analogies with the
series iron, cobalt, nickel, copper, zinc, that the atomic weights
of the elements of the series osmium, iridium, platinum, gold,
mercury, would rise in this order, and at the time of the
establishment of the periodic law (1869), the determinations of
Berzelius, Rose, and others gave the following values for the
atomic weights: Os = 200, Ir = 197, Pt = 198, Au = 196, Hg = 200.
The fulfilment of the expectations of the periodic law was given
in the first place by the fresh determinations (Seubert, Dittmar,
and Arthur) of the atomic weight of platinum, which proved to be
nearly 196, if O = 16 (as Marignac, Brauner, and others propose);
in the second place, by the fact that Seubert proved that the
atomic weight of osmium is really less than that of platinum, and
approximately Os = 191; and, in the third place, by the fact that
after the researches of Krüss, Thorpe, and Laurie there was no
doubt that the atomic weight of gold is greater than that of
platinum--namely, nearly 197.

[28 bis] In Chapter XXII., Note 40, we gave the thermal data for
certain of the compounds of copper of the type CuX_{2}; we will
now cite certain data for the cuprous compounds of the type CuX,
which present an analogy to the corresponding compounds AgX and
AuX, some of which were investigated by Thomsen in his classical
work, 'Thermochemische Untersuchungen' (Vol. iii., 1883). The data
are given in the same manner as in the above-mentioned note:

R = Cu Ag Au
R + Cl +33 +29 +6
R + Br +25 +23 0
R + I +16 +14 -6
R + O +41 + 6 -?

Thus we see in the first place that gold, which possesses a much
smaller affinity than Ag, evolves far less heat than an equivalent
amount of copper, giving the same compound, and in the second
place that the combination of copper with one atom of oxygen
disengages more heat than its combination with one atom of a
halogen, whilst with silver the reverse is the case. This is
connected with the fact that Cu_{2}O is more stable under the
action of heat than Ag_{2}O.

[29] Heavy atoms and molecules, although they may present many points
of analogy, are more easily isolated; thus C_{16}H_{32}, although,
like C_{2}H_{4}, it combines with Br_{2}, and has a similar
composition, yet reacts with much greater difficulty than
C_{2}H_{4}, and in this it resembles gold; the heavy atoms and
molecules are, so to say, inert, and already saturated by
themselves. Gold in its higher grade of oxidation, Au_{2}O_{3},
presents feeble basic properties and weakly-developed acid
properties, so that this oxide of gold, Au_{2}O_{3}, may be
referred to the class of feeble acid oxides, like platinic oxide.
This is not the case in the highest known oxides of copper and
silver. But in the lower grade of oxidation, aurous oxide,
Au_{2}O, gold, like silver and copper, presents basic properties,
although they are not very pronounced. In this respect it stands
very close in its properties, although not in its types of
combination (AuX and AuX_{3}), to platinum (PtX_{2} and PtX_{4})
and its analogues.