Joseph William Mellor.

A comprehensive treatise on inorganic and theoretical chemistry (Volume 2) online

. (page 19 of 156)
Online LibraryJoseph William MellorA comprehensive treatise on inorganic and theoretical chemistry (Volume 2) → online text (page 19 of 156)
Font size
QR-code for this ebook

of chlorine condensed along with the hydrogen chloride by allowing the liquid to
evaporate ; passing the vapour through a quartz tube filled with mercury vapours ;


and condensing the purified hydrogen chloride, back to a liquid, or dissolving it in
water. Result, 35*461. R. W. Gray and F. P. Burt first determined the weight of
a normal litre of hydrogen chloride to be 1 '63885 grms. ; they then passed the gas
over heated aluminium, and measured the volume of the liberated hydrogen. Two
volumes of hydrogen chloride gave 1 '00790 vols. of hydrogen. With Morley's value
for the density of chlorine, it was found that 36 '4672 grms. of hydrogen chloride give
35'4594 grms. of chlorine, if the unit of hydrogen be 1'0078 grm. P. A. Guye and
C-. Ter-Gazarian's determination of the relative density of hydrogen chloride also
gave 35 '461 for the at. wt. of chlorine. J. Deutsch obtained a rather lower value.

m. The analysis of nitrosyl chloride. P. A. Guye and G. Fluss 9 first distilled a
known weight of purified nitrosyl chloride, NOC1, over heated silver this retained
the chlorine ; it was then passed over heated copper this retained the oxygen ;
and finally it was passed over metallic calcium -this retained the nitrogen. The
sum of the weights of chlorine, oxygen, and nitrogen so determined was usually a
little less up to 0'0012 grm. than the weight of the nitrosyl chloride employed.
The mean value for chlorine gave 35*468.

IV. Physical methods. Several physical methods have been employed for deter-
mining the mol. wt. of hydrogen chloride. The density of hydrogen chloride deter-
mined by R. W. Gray and F. P. Burt 10 was 1 '62698 grms. for a normal litre ; and
1*42762 for oxygen. Hence, with the mol. wt. of oxygen, 32, as the standard of
reference, the mol. wt. x of hydrogen chloride is given by the proportion 1 '42762 :
1-62698=32 : x, where z=36'469 ; with the at. wt. of hydrogen 10076, the at. wt. of
chlorine is 36 '469 1*0076=35 '461. The compressibility coeff. that is, the mean
deviation of hydrogen chloride from Boyle's law required for an application of the

D. Berthelot's method of limiting densities (1. 6, 8), has been determined by A. Leduc
to be 4,1=0-00758 ; E. Briner, 0'00750 ; R. W. Gray and F. P. Burt, 0'00743 ; and
A. Jaquerod and 0. Scheuer found for oxygen ^4 I =0'00097 ; R. W. Gray and
F. P. Burt, 0'000964. Hence, A. Leduc calculated the mol. wt. of hydrogen chloride
to be 36'460 ; and E. Briner, 36'462. A. Leduc's method of molecular volumes,
where it is assumed that all gases at corresponding temp, and press, have the same
mol. vol., gave the value 36'450 for the mol. wt. of hydrogen chloride ; and

E. Briner, 36 '453. The method by the reduction of the critical constants, based on
the equation: Mol. wt. referred to oxygen 32 is equal to 22 '412 W /(I +a)(l b),
where W denotes the weight of a litre of gas at and 760 mm., reduced to sea-level,
and latitude 45. P. A. Guye and G. Ter-Gazarian n found for a normal litre
of hydrogen chloride, W =1'6398 ; and (1+ a )(l &o)=l '00773, so that the mol. wt.
of hydrogen chloride is 36 '4393, and if the at. wt. of hydrogen is 1*0078, that of
chlorine will be 35'4615.

Starting with oxygen 16 as the standard of reference, F. W. Clarke (1910) regards
35 '4584 0'0002 as the best representative value for the at. wt. of chlorine if the at.
wt. of the silver be 107 '8800 '00029 ; and B. Brauner (1913) considers the best
representative value to be 35'457, if silver be 107 '880 ; 35 '456, if silver be 107 '876 ;
and 35*454, if silver be 107 '871. B. Brauner also inquires : What reliance can be
placed on the third decimal ? and answers that the uncertainty is smaller if silver
be regarded as a fixed basis with respect to oxygen 16, and greater when the at. wt.
is referred to oxygen 16 alone regarded as a fixed constant. If referred to silver
fixed at 107-876, the at. wt. of chlorine lies between 35'454 and 35'458, or 35'456
0'0019 ; but if the value for chlorine is made to depend upon the cycle of relations
between it and oxygen=16, " the at. wt. of chlorine lies between 35'453 and 35'459,
or possibly 35'452 and 35'460 ; " or the at. wt. of chlorine is 35'456 0'003, or possibly
35*456 0"004. B. Brauner illustrates the idea geometrically by means of Fig. 22,
which he says makes it much clearer than is possible by mathematical symbols,
how the greater the distance right or left from the middle line corresponding with
Ag=107'876, Cl=35"456, the smaller the probability that the corresponding value
for the at. wt. will be correct. The international value for the at. wt. of chlorine
is 35-46.


It is astonishing what a vast amount of labour has been expended in the struggle
for increased accuracy in the determination of the at. wt. ratios. Some of the later
determinations are masterpieces of precision. It is probable that the majority of
these researches has been directed towards the chlorine : oxygen or the chlorine :
hydrogen ratio on account of its fundamental importance. Even now J. F. W. Her-
schel's words 12 are not inapplicable :

It is doubtful whether such accuracy in chemical analysis has yet been attained as to
enable us to answer positively for a fraction not exceeding the three or four-hundredth part
of the whole quantity to be determined ; at least, the results of experiments obtained
with the greatest care often differ by a greater amount.

Many have had misgivings as to the utilitarian value of the enormous labour which
has been expended in this direction, and particularly when the best representative
values of all the best results are usually rounded off to the nearest tenth when the
at. wt. are employed in chemical calculations. Lord Kelvin's words are often quoted
as a stimulus to greater and still greater precision :

Accurate and minute measurement seems to the non-scientific imagination a less lofty
and dignified work than looking for something new ; discoveries of science have been the
reward of accurate measurement, and patient, long-continued labour in the minute sifting
of numerical results.

The eighty or more individual numbers we call at. wt., adds T. W. Kichards, " are
perhaps the most striking of the physical records which Nature has given us concern-
ing the earliest stages of the evolution of the universe. They are mute witnesses

= 107-871 107-872 /07-874 107-876 107-878 107-879 W7-887

CI=35<45Z 35-453 35-453 35-464 35-454 35*55 35456 35-456 55-457 35-457 35*58 35-459 35-453

FIG. 22. B. Brauner's Table of the Relation of the Atomic
Weight of Chlorine to that of Silver (Oxygen 16).

of the first beginnings of the cosmos out of chaos, and their significance is one of the
first concerns of the chemical philosopher."

We are now promised atoms of one element of different weight ; so that the
observed at. wt. is a kind of average. If the atoms of different weight could be
separated, the fractions would occupy the same position on the periodic table, and
be chemically identical they are called isotopes, and the subject is discussed in
Vol. III. According to the positive ray spectrograph, F. W. Aston 13 obtained
results which show that chlorine contains two isotopes of at. wt. 35 and 37 ; and
W. D. Harkins has claimed that in the atomysis of hydrogen chloride the density
of the fraction which remains in the diffusion tube increases at a rate corresponding
with the assumption that the chlorine isotopes have at. wts. 35 and 37 with possibly
a third of at. wt. 39. F. W. Aston added :

At first sight it may seem incredible that chlorine, whose chemical combining weight
has been determined more often and with greater accuracy than almost any other element,
should not have given evidence of its isotopic nature in the past ; but it must be remembered
that, in all probability, every one of these determinations has been performed with chlorine
originally derived from the sea in which the isotopes, if ever separate, must have been
perfectly mixed from the most remote ages. Chlorine from some other source,- if such can
be found, may well give a different result, as did radio-lead when examined.

The atomic weight of bromine. The at. wt. of bromine has been determined
by methods which follow in principle those employed for chlorine. A. J. Balard
(1826), 14 the discoverer of bromine, transformed a known weight of potassium bromide
into the sulphate, and also reduced silver bromide to metallic silver by means of
zinc ; the numbers 74*7 and 75'3 were respectively obtained. J. von Liebig (1826)


also transformed potassium bromide into silver bromide, and obtained 75*2 for the
at. wt. of the element ; C. Lowig (1829) obtained 75*76. These numbers are very
low ; this is, no doubt, due to the impurities present in the salts. J. J. Berzelius (1828)
converted silver bromide into the chloride by the action of chlorine gas, and obtained a
value tres approchee du but, namely, 79*36. J. B. A. Dumas (1859), by the same method,
obtained 79*95. W. Wallace precipitated silver bromide from arsenic tribromide,
AsBr 3 , by the addition of silver nitrate, and obtained an at. wt. of 79738, when the
at. wt. of silver is 107*97, and arsenic 75. The preceding determinations are usually
disregarded in modern estimates of the at. wt. of bromine because of the then
imperfect state of the art of chemical analysis as involved in the work.

J. C. G. de Marignstc 15 decomposed potassium bromate by careful calcination,
and precipitated the bromine from the resulting potassium bromide by treatment
with silver nitrate. J. S. Stas reduced silver bromate by treatment with sulphurous
acid ; the ratio of silver to bromine was also determined by J. C. G. de Marignac
and by J. S. Stas either by synthesizing silver bromide from its elements ; or by
converting potassium bromide into silver bromide. J. S. Stas' value, 79*9628
0*0032, has been recalculated by F. W. Clarke and by J. D. van der Plaats, who
obtained respectively 79*951 and 79*955. A. Scott obtained values varying from
79*899 to 79*911 from his analysis of ammonium bromide. J. S. Goldbaum
(1911) by the electrolytic method used for the ratio Na : Cl, obtained for the at. wt.
of bromine 79*927, when the at. wt. of sodium is taken as 23'00. G. P. Baxter (1906)
dissolved the purest silver in nitric acid, precipitated silver bromide by the addition
of ammonium bromide, and finally fused the washed and dried product in bromine
vapour. He obtained values ranging from 79*914 to 79*918 average 79*915 as
the best representative value of his determinations. By collecting together all the
direct and incidental determinations of the silver-bromine ratios since about 1843,
F. W. Clarke obtained 79*9197 as the best representative value for the at. wt. of
bromine (silver=107*880, oxygen=16) ; B. Brauner obtained 79*916, if silver be
107-880 ; 79-913, if silver be 107*876 ; and 79*909, if silver be 107*871 ; the error in
the third decimal place is estimated to be about 0*004. P. A. Guye, E. Moles, C. K.
Reiman, and W. J. Murray have computed the at. wt. of bromine from determina-
tions of the density and compressibility of hydrogen bromide. The results lie
between 79*924 and 79*926, when H=1'0076. The International table gives 79'92
as the best representative value when silver is 107 '88.

The atomic weight of iodine. In his historic Memoire sur Viode (1814), J. L. Gay
Lussac 16 attempted to determine the at. wt. of iodine by the synthesis of zinc iodide,
and in this manner he obtained the value 125 for the constant. H. Davy obtained
132 by converting sodium hydroxide into iodide. By an analogous process to that
employed by J. L. Gay Lussac, W. Prout obtained 126. In 1825, T. Thomson
obtained 124 as a result of decomposing potassium iodide. In 1828, J. J. Ber-
zelius converted silver iodide into the chloride by the action of chlorine on the heated
salt, and obtained values ranging between 126*26 and 126*39; and by the same method,
J. B. A. Dumas obtained the value 126*59. In 1843, N. A. E. Millon determined the
percentage of oxygen in potassium iodate and silver iodate and obtained respec-
tively 126*697 and 125*33. The preceding determinations are usually disregarded
in modern estimates of the at. wt. of iodine because " the art of qualitative analysis
was then in its infancy."

J. C. G. de Marignac showed that the ignition of the iodate is not suited for the deter-
mination since some iodine is lost during the calcination, and he preferred the synthesis
of silver iodide by dissolving a known weight of silver in nitric acid and precipitating
the contained silver as iodide, by the addition of potassium iodide. The result
furnished 126*537 and 126*550. J. S. Stas 17 analyzed silver iodate and determined
both the water and the oxygen given off in each calcination, as well as the amount
of silver iodide. He also reduced the iodate to iodide by sulphurous acid, and
synthesized silver iodide by precipitation from silver nitrate by means of hydriodic
acid, and by treating silver sulphate with ammonium iodide. The results varied


from 126*85 to 126 '864 average 126*855. Both J. C. G. de Marignac's and J. S.
Stas' results were affected by constant errors, the chief one being due to the oc-
clusion of silver nitrate by the precipitated silver iodide. In 1902, A. Ladenburg
claimed that the value 126 '85 was about one-tenth too low ; and soon afterwards,
A. Scott obtained the value 126 '912 as a result of two syntheses of silver iodide.
P. Kothner and E. Aeuer synthesized silver iodide by heating the metal in iodine
vapour, and by precipitation from silver nitrate by the addition of hydriodic acid,
taking precautions to eliminate the occlusion of mother-liquor. As a result, they
obtained 126'936 (oxygen 16).

A very careful series of determinations were made by G. P. Baxter about the
same time. He precipitated silver iodide from silver nitrate by treatment of the
soln. with ammonium iodide in the presence of an excess of ammonia ; and also by
converting iodine into ammonium iodide and precipitating the silver iodide by the
addition of silver nitrate, taking precautions to avoid an excess of the latter reagent.
As a result of the whole work, 126 '929 was obtained for the at. wt. of iodine.
G. P. Baxter and G. S. Tilley also measured the ratio of silver to iodine pentoxide,
and found '646230, which makes the at. wt. of iodine 126 '891. M. Guichard de-
composed iodine pentoxide into oxygen and iodine by heat ; the former was absorbed
by red-hot copper, and the iodine condensed. He found that if oxygen has an at.
wt. of 16, iodine is 126 '915. G. Gallo electrolyzed a soln. of silver salt so that silver
was deposited on the cathode, and the iodine liberated at the anode by titration
with sodium thiosulphate. His determinations varied from 126 '82 to 126 '98
average 126 '89.

Summing up the various determinations, F. W. Clarke obtained 126*920 0*00033
as the best representative value for the at. wt. of iodine ; and B. Brauner 126*932,
if silver be 107 "880 ; 126'927, if silver be 107 '876 ; and 126'921, if silver be 107 '871.
It is very probable that the at. wt. is greater than 107 '870 (oxygen 16) and smaller
than 107*880, consequently, says B. Brauner, " the uncertainty in these values for
the at. wt. of iodine does not extend to many units in the third decimal place."
This makes the at. wt. of iodine 126*93. The International table gives 126'92 as
the best representative value.

The anomaly in the at. wt. of iodine and tellurium with respect to their position
in the periodic table has greatly stimulated researches on the at. wt. of these
elements. It has been asked : Does iodine contain an undiscovered halogen
element of higher atomic weight than iodine ? In answer, G. P. Baxter converted
iodine into hydriodic acid by hydrogen sulphide, and the hydriodic acid was con-
verted back to iodine by distillation with potassium permanganate in small quantities
at a time so as to obtain four fractions. No difference could be detected in the
different fractions. If a halogen element were present in iodine with properties to
be expected from the analogies with other members of the halogen family, it should
have accumulated in the first fraction. Hence it is unlikely that iodine contains a
halogen element of higher at. wt. than iodine. E. Kohlweiler, however, claims to
have evidence of the existence of iodine isotopes.

The molecular weights of iodine, bromine, and chlorine. When the vapour
density determinations of all known volatile chlorides are collected together, a table
illustrated by the excerpt Table XI is obtained. The smallest combining weight of
chlorine in any one of these compounds corresponds with the combining 35 '46-
oxygen=16 and accordingly this number is taken to represent the at. wt. of
chlorine. The at. and eq. wt. of chlorine have the same numerical value. Similar
results are obtained with the volatile fluorides, bromides, and iodides. The results
show that the best representative values for the at. wt. of the halogens are Fl, 19 '0 ;
Cl, 35'46 ; Br, 79 '92 ; and I, 126'92. The vapour densities of the halogens corre-
spond with diatomic molecules; at elevated temp., as already shown, there are
signs of dissociation into monatomic molecules, and this more particularly with
iodine and bromine than with chlorine and fluorine. E. Paterno and R. Nasini 1S
have determined the mol. wt. of bromine in aq. and acetic acid soln. by the f.p.



method, and the results correspond with the formula Br 2 . Iodine in dil. benzene
soln. corresponded with the molecule I 2 , but in more cone, soln., the molecule seemed
to be more complex. In acetic acid soln. the results were intermediate between those
corresponding with mono- and di-atomic molecules. There has, however, been
much discussion on the molecular condition of iodine in different solns.


Volatile chloride.

Vapour density.

Formula of com-
pound : Mol. wt. =
vapour density.

Amount of chlorine
in the molecule.



Hydrogen chloride .




Chlorine .... 70*9

C1 2


Mercuric chloride . . . 2 73 '6

HgCl 2


Arsenic trichloride . . . 182*1

AsCl 3


Tin tetrachloride . . . 260-2

SnCl 4


Phosphorous pentachloride . 208 '3

PC1 5


The valency of chlorine, bromine, and iodine. Compounds are known in which
the three halogens act as uni-, ter-, quinque-, or septa- valent elements. Usually,
however, these elements are univalent. In chlorine dioxide, C10 2 , the chlorine is
bi- or quadri- valent. 19 In M. Berthelot's hydrogen perchloride, HC1 3 , the chlorine
is probably tervalent, and R. Meldola (1888) showed that the oxygen in the hydro -
chloride of methyl oxide is best regarded as quadrivalent, the chlorine tervalent ;
thus, (CH 3 ) 2 : : CLH. Iodine also appears to be tervalent in the so-called iodoninm

When chlorine is passed into a chloroform soln. of iodobenzene, C 6 H 5 I, an addition
compound phenyliododichloride, C 6 H 5 I.C1 2 , is produced in which the halogen atoms are
probably tervalent. When this compound is treated with alkali hydroxide, iodosobenzene,
C 6 H 5 IO, is formed ; and when this compound is boiled with water, it produces iodoxy-
benzene, C 6 H 5 IO 2 , and iodobenzene, C 6 H 5 I. By the action of silver hydroxide upon a
mixture of eq. amounts of iodosobenzene, C 6 H 5 IO, and iodoxybenzene, C 6 H 5 IO 2 , followed
by treatment of the clear nitrate with potassium iodide. V. Meyer prepared the so-called
diphenyl iodonium hydroxide, (C 6 H 5 ) 2 : I. OH. The precipitated hydroxide forms an iodide,
(C 6 H 5 ) 2 : I.I, which crystallizes from alcohol in yellow needle-like crystals melting between
175 and 176. The iodonium bases and salts resemble those of lead and silver but par-
ticularly those of thallium. The colour and solubility of the halides resemble the corre-
sponding salts of these metals. The chloride is white, the bromide pale yellow, and the
iodide yellow. The hydroxide and carbonate are soluble in water, and the soln. give an
alkaline reaction, as is the case with the corresponding thallium salts. Diphenyl iodonium
hydroxide gives a precipitate with ammonium sulphide, which looks like freshly precipitated
antimony sulphide, and it consists mainly of a trisulphide possibly I S S S 1 = (C 6 H 5 ) 2 .
The nitrate, (C 8 H 5 ) 2 I.NO 3 , acid sulphate, chromate, periodide, and other salts have been
prepared. The base in question seems to be a derivative of an hypothetical iodonium
hydroxide, HO I H 2 , analogous with hydroxylamine, HO N=H 2 . Monoiododiphenyl-
iodonium derivatives have been prepared by treating iodosobenzene, C 6 H 5 IO, with sulphuric
acid at a low temp., and afterwards treating the dil. soln. with potassium iodide. The
iodide, (C fl H 5 )I=I CgH^ I, and the other salts, as well as the base, (C 6 H 5 )OH=I C 6 H 4 I,
resemble the corresponding salts of diphenyliodionium.

Compounds of the type KBr 3 are usually supposed to contain a tervalent halogen ;
the chlorates, bromates, and iodates, and compounds of the type CsI 5 , to contain
quinquevalent halogens ; and the perchlorates and periodates and compounds of
the type CsI 7 , to contain septivalent halogens.


1 F. W. Clarke, A Recalculation of the Atomic Weights, Washington, 1910 ; B. Brauner, Abegg\t
Uandimch der anorganiachen Chemie, Leipzig, 4. ii, 50, 1913 ; J. D. van dcr Plaats, Ann. Chim.
Phy., (6), 7. 499, 1886.

2 J. J. Berzelius, Pogg. Ann., 8. 1, 1826 ; T. Thomson, Ann. Phil, 15. 89, 1820.


3 F Penny, Phil. Trans., 129. 30, 1839 ; T. J. Pelouze, Compt. Rend., 15. 959, 1842 ; C. Ger-
hardt, ib., 21. 1280, 1845 ; P. A. Guye and G. Ter-Gazarian, ib., 143. 411, 1907 ; J. C. G. de
Marignac, Liebig's Ann., 44. 11, 1842 ; L. Maumene, Ann. Chim. Phys., (3), 18. 71, 1846 ; V. Faget,
ib., (3), 18. 20, 1846 ; J. S. Stas, Mem. Acad. Belgique, 35. 3, 1865 ; A. Stabler and F. Meyer, anorg. Chem., 71. 378, 1911 ; T. W. Richards and H. H. Willard, Journ. Amer. Chem. Soc.,
32. 4, 1910.

4 E. Turner, Phil. Trans., 119. 291, 1829 ; 123. 529, 1833 ; F. Penny, ib., 129. 28, 1839 ;
J. C. G. ds Marignac, Liebig's Ann., 44. 11, 1842 ; J. B. A. Dumas, Ann. Chim. Phys., (3), 55. 134,
1859 ; J. S. Stas, Mem. Acad. Belgique, 35. 3, 1865 ; T. W. Richards and R. C. Wells, A Revision
of the Atomic Weight of Sodium and Chlorine, Washington, 1905.

5 J. C. G. de Marignac, Liebig's Ann., 44. 11, 1842 ; T. J. Pelouze, Compt. Rend., 20. 1047,
1845 ; J. B. A. Dumas, Ann. Chim. Phys., (3), 55. 134, 1859 ; J. S. Stas, Mem. Acad. Belgique, 35.
3, 1865 ; L. Maumene, Ann. Chim. Phys., (3), 18. 41, 1846 ; A. Thiel, Zeit. anorg. Chem., 40.
313, 1904; A. Scott, Journ. Chem. Soc., 79. 147, 1901 ; T. W. Richards and R. C. Wells, A
Revision of the Atomic Weight of Sodium and Chlorine, Washington, 1905 ; T. W. Richards and

E. H. Archibald, Proc. Amer. Acad., 38. 456, 1903 ; E. H. Archibald, Tram. Roy. Soc. Canada, 3.
47, 1904 ; T. W. Richards and A. Staehler, Further Researches concerning Atomic Weights,
Washington, 1907 ; T. W. Richards, P. Kothner, and E. Tiede, Journ. Amer. Chem. Soc., 31. 6,
1909 ; J. S. Goldbaum, ib., 33. 35, 1911.

6 T. W. Richards, Journ, Chem. Soc., 99. 1201, 1911 ; Methods used in Precise Chemical Investi-
gation, Washington, 1910.

7 F. W. Clarke, Amer. Chem. Journ., 27. 321, 1902.

8 H. B. Dixon and E. C. Edgar, Phil. Trans., 205. A, 169, 1905 ; E. C. Edgar, ib., 209. A. 1, 1908 ;
W. A. Noyes and H. C. P. Weber, Journ. Amer. Chem. Soc., 30. 13, 1908 ; R. W. Gray and

F. P. Burt, Journ. Chem. Soc., 95. 1633, 1909 ; Proc. Chem. Soc., 24. 215, 1908 ; R. W. Gray,
ib., 23. 119, 1907 ; P. A. Guye and G. Ter-Gazarian, Compt. Rend., 143. 1233, 1906 ; J. Deutsch,
Essai sur une nouvelle methode pour la determination du poids atomique du More, Geneve, 1905.

9 P. A. Guye and G. Fluss, Journ. Chim. Phys., 6. 732, 1908.

10 R. W. Gray and F. P. Burt, Journ. Chem. Soc., 95. 1633, 1909 ; A. Leduc, Recherches sur les
gaz, Paris, 1898 ; Compt. Rend., 140. 642, 1905 ; P. A. Guye, Journ. Chim. Phys., 3. 321, 1905 ;
0. Scheuer, Zeit. phys. Chem., 68. 575, 1909 ; E. Briner, Journ. Chim. Phys., 4. 476, 1906 ;
A. Jaquerod and 0. Scheuer, Mem. Soc. Geneve, 35. 659, 1908 ; K. Olschewsky, Monatsh., 5.
127, 1884.

11 P. A. Guye, Compt. Rend., 138. 1213, 1904 ; Journ. Chim. Phys., 3. 321, 1905 ; G. Ter-
Gazarian, ib., 7. 337, 1910 ; 9. 101, 1911.

12 J. F. W. Herschel, A Preliminary Discourse on the Study of Natural Philosophy, London, 307,

Online LibraryJoseph William MellorA comprehensive treatise on inorganic and theoretical chemistry (Volume 2) → online text (page 19 of 156)