LIB
QD
561
G441
B 401605
Gibbons, V. L
The potentials of silver in
of silver nitrate
nonaqueous solutions

GENERAL LIBRARY
NOV 12 1914
The Potentials of Silver in Non-
Aqueous Solutions of Silver
Nitrate
CHEMICAL
BRARY
QI
561
G44
A DISSERTATION
PRESENTED TO THE FACULTY OF BRYN MAWR COLLEGE IN PART
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY
BY
VERNETTE L. GIBBONS
1914
EASTON, PA.:
ESCHENBACH PRINTING COMPANY
1914

The Potentials of Silver in Non-
Aqueous Solutions of Silver
Nitrate
A DISSERTATION
PRESENTED TO THE FACULTY OF BRYN MAWR COLLEGE IN PART
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY
BY
VERNETTE L. GIBBONS
1914
EASTON, PA.:
ESCHENBACH PRINTING COMPANY
1914

5'3-510110

57
7
II
12
15
20
23
27
28
31
I. Introduction..
TABLE OF CONTENTS.
a. Aim of the Investigation.
b. Preparation of Materials.
II. Measurements of Electrical Conductance.
III. Determination of Transport Numbers..
IV. Determination of the Electromotive Force of Various Concentration Cells..
V. Discussion of Results.
VI. Summary.
VII. Chronological Bibliography.
VIII. Vita...

The Potentials of Silver in Nonaqueous Solu-
tions of Silver Nitrate.
I.
Introduction.

The study of the potentials of metals in nonaqueous solutions of their
salts was begun by Campetti¹ about twenty years ago. Since, however,
the drop electrode was employed in these experiments the results cannot
be considered trustworthy.
A few months later, Jones2 published results of measurements of the
combinations
Ag | AgNO3 in solvent I | AgNO3 in solvent II | Ag.
The solvents used were water, ethyl alcohol, methyl alcohol and acetone.
The concentrations of both solutions were identical for each individual
experiment. Electromotive forces of considerable numerical value were
obtained, but they varied with the combination employed. He concluded
that the solution tension of silver depends upon the nature of the solvent
in which the salt is dissolved.
Kahlenberg³ carried out a long series of investigations of the behavior
of ten different metals, in solutions of their salts in about thirty different
solvents. He measured the potential of silver in o.1 N solutions of silver
nitrate in twenty-four different solvents. The silver nitrate solution
formed one part of the cell, and in each case the other part was either
a normal or o.1 N calomel electrode. The metals dipped into solutions
contained in test tubes or open beakers and the connection between the
two vessels was made by means of a strip of filter paper folded to several
thicknesses. Thus the solutions were liable to change in concentration
because of evaporation or of absorption of moisture from the atmosphere.
The difference of potential at the junction of the two liquids was neglected
and no definite temperature was maintained. The change in electro-
motive force caused by the interaction of the ions of the two electrolytes
at the junction was entirely overlooked.
The following year Kahlenberg measured the electromotive force
of concentration cells of silver nitrate in pyridine and acetonitrile.
1 Atti accad. Torino, 28, 61 and 228 (1893).
2 Z. physik. Chem., 14, 346 (1894).
3 J. Phys. Chem., 3, 379 (1899).
4 Ibid., 4, 709 (1900).
6
He compared the results with the values calculated by means of the Nernst
equation
E =
==
RT
nF
loge
m₁A1
M2A2
(1)
assuming that the potential at the junction between the two solutions
was negligible. Although a fair agreement between experimental and
calculated values had been obtained when aqueous solutions were used,
here an utter lack of agreement was found, and he was forced to the con-
clusion that the formula was not applicable to cells involving nonaqueous
solutions.
In a later article, Kahlenberg¹ gave the results of measurements of the
electromotive force of the system
Ag | 0.1 N AgNO3 in water | 0.1 N AgNO3 in pyridine and water | Ag.
He found that, by increasing the amount of water in the mixed solvent,
the electromotive force of the cell diminished, and he concluded that the
potential difference between silver and silver nitrate depended upon the
composition of the solvent.
Carrara and D'Agostini,2 from their investigation of concentration cells
in methyl alcohol, concluded that in this solvent the Nernst formula was
applicable.
In 1905, Carrara³ turned his attention to the influence of the solvent
upon the solution pressure of the metals. He found that in methyl al-
cohol this pressure was less than in water and was, therefore, not inde-
pendent of the nature of the solvent.
Sackur criticized the work of Kahlenberg and also that of Carrara and
D'Agostini. He maintained that neither had disproved the applicability
of Nernst's law; that, in fact, in the case of the work of Carrara and D'Agos-
tini the results with three out of the four metals investigated could be
interpreted as in full accord with that law.
In the same year, Bjerrum attempted to eliminate the diffusion po-
tential at the junction of the two liquids by using a concentrated solu-
tion of potassium chloride as a "middle liquid" and obtained values more
nearly constant for each experiment and with less variation in a series
of experiments than when this "middle liquid" was not used.
The measurement of the electromotive force of concentration cells of
silver nitrate in methyl- and ethylamine was undertaken by Bodländer

1 Z. Elektrochem., 11, 385 (1905).
2 Atti R. Ist. Veneto, 62, 793 (1902).
3 Gazz. chim. ital., 35, I, 132 (1905).
4 Z. Elektrochem., 11, 385 (1905).
5 Z. physik. Chem., 53, 428 (1905). ·

7
and Eberlein' as a means of studying the complexes formed in these sol-
vents. They considered that these ions are represented by the formulas,
Ag(CH3NH2)2 and Ag(C2H5NH2)2, thus showing the similarity of these
solvents to ammonia.
The potential difference of concentration cells with liquid ammonia
as a solvent was measured by Cady in 1905.2 His conclusion was that
the Nernst formula could be used to calculate the electromotive force
when this solvent was used in the cells.
A large amount of experimental work was carried out by Neustadt and
Abegg for the purpose of determining the relation of the tension series
of the metals to the solvent. Ag | AgNO3 was always made one-half of
the cell. In only a few cases is mention made of the use of a "middle
liquid;" although, in several cases, the solution in the second half of the
cell was a chloride. They concluded that the tension series was quite
independent of the solvent. The liquid potentials in a given system were
considered at some length but without reaching any definite conclusion.
The behavior of cadmium in alcoholic solutions of its salts was investi-
gated by Getman.4 He measured the electromotive force of the metal
in different concentrations of its salts, the other half of the cell being in
each case a normal calomel electrode. The changes in potential with con-
centration in the alcoholic solutions were found to be opposite to the changes
in aqueous solutions, i. e., the potential of cadmium became more strongly
negative with increasing concentration of the solutions of its salts. The
electromotive force developed at the junction of the two solutions was
assumed to be negligible. The possibility of the application of Nernst's
equation to such nonaqueous solutions was considered doubtful.
In 1912 the potentials of zinc in alcoholic solutions of zinc chloride were
measured by Getman and Gibbons, one-half of the cell being in each case
the normal calomel electrode. It was found that the potentials of zinc
in solutions in ethyl alcohol became more strongly negative with increas-
ing concentration of the solutions. This was the same effect observed
by Getman in his study of solutions of cadmium salts. Solutions in methyl
alcohol showed fluctuations of potential of more than ten millivolts, an
effect which is to be studied further. Shortly after this, Dr. Laurie of
Edinburgh, in a personal letter to Dr. Getman, called attention to the
fact that the potential difference developed at the junction of the two liquids
might sometimes have considerable influence."
1 Ber., 36, 3945 (1903).
2 J. Phys. Chem., 9, 476 (1905).
3 Z. physik. Chem., 69, 486 (1909).
4 Am. Chem. J., 46, 117 (1911).
6 Proc. Roy. Soc. Edinb., 31, 375 (1911).
5 Ibid., 48, 124 (1912).
8
Roshdestwensky and Lewis' have studied concentration cells of silver
nitrate in acetone. They not only measured the electromotive forces
of various combinations of freshly prepared solutions, but they made
similar measurements both with solutions which had stood for six weeks
in the dark and with solutions which had stood one day exposed to the
sunlight. The agreement of the results in all cases is within the limits
of experimental error. The type of cell used was not described, but the
two solutions were in direct contact, thus possibly developing a potential
difference at the junction. In the following way, however, this potential
difference was shown to be very slight. A solution of 0.01 N silver ni-
trate was interposed as a "middle liquid" and the measurements repeated.
The results agree with those for a similar cell without this "middle liquid."
Therefore, assuming Nernst's law to be valid for these systems, Rosh-
destwensky and Lewis calculated the transport numbers for the ions and
found that in solutions ranging in concentration between the limits of
(0.02-0.007 N) the value for the anion lay between 0.60 and 0.58, and for
more dilute solutions (0.007-0.0005 N) the value was 0.56. As a result
of further attempts to eliminate the potential differences at the junction
of the two liquids, they concluded that the validity of "middle liquid"
methods for acetone solutions was doubtful since, in all cases, where an
effect was observed there was an increase instead of a decrease of electro-
motive force.
Measurements of the electromotive force of silver nitrate concentra-
tion cells with water and ethyl alcohol as solvents were carried out by
Bell and Feild.2 They used a closed apparatus of a U form, having an
outlet tube in the middle, closed with a three-way stopcock, so that the
two arms could be connected with each other or the solution drawn from
either one without disturbing that in the other. In this way, the solutions
could be brought to the same level in both sides of the tube without
intermingling and could be kept separated except during the time of
actual measurement. Apparently, the effect of gravity in causing inter-
diffusion was overlooked. Writing Nernst's law in the form

E
=
20
RT
loge 10 = K,
log10 C1/C2 u + v NF
(2)
they obtained from the data for concentration cells in water, values for
K that vary from 0.0560 to 0.0623. Taking the latter as the value of
K they calculated the value of v in the above formula and found it to be
0.523 instead of 0.528 as given by Lehfeldt. Since the values of v are
probably smaller in more concentrated solutions, K will, of course, for
1 J. Chem. Soc., 101, 2094 (1912).
2 J. Am. Chem. Soc., 35, 715 (1913).
3 Electrochemistry, p. 256 (1904).

9
such solutions, be proportionately smaller. In the same way, K was
calculated from the measurements of the electromotive forces of concen-
tration cells in ethyl alcohol and, since K varied, they assumed that the
migration ratios were not constant. The value of v calculated from the
highest value of K obtained was 0.62.
Bjerrum¹ sought to eliminate the potential at the junction by using,
in one experiment, a saturated, and in another, a half-saturated solution
of potassium chloride as a "middle liquid." By comparing the results
of the two measurements and by extrapolating, he concluded that better
results were obtained than when the potassium chloride solution was not
present.
Cumming and Abegg2 concluded that a saturated ammonium nitrate
solution formed an exceptionally good "middle liquid" for eliminating
the potential at the junction.
Thermodynamic principles were applied by Henderson³ in the deriva-
tion of the following equation for calculating the potential developed
between two aqueous solutions:
RT
E
=
(u1-v1)C1-(U2-V2) C2
—
F (u₁W₁ + v₁W1)C1 — (u₂W2 + v2W2)C2
(u1W1v1W1)C1
loge
(3)
(u₂W2 + V2W2) C2
C1 and C2 represent equivalent concentrations. U1, U1, U2, V2 represent
absolute ionic velocities. W1, W1, W2, W2 represent the valences of the
ions.
Subsequent investigation's by Bjerrum led him to conclude that, if
the electromotive force remained constant for some minutes after contact
between the liquids had been established, the effect of diffusion was neg-
ligible and Henderson's formula could be applied. He found, also, that
if the connection between the two parts of the cell was made through a
layer of sand much more constant results were obtained. In a second
paper, he showed that ammonium nitrate gave less satisfactory results
than potassium chloride when used as a "middle liquid," but that it
might be used in cases where the latter would give rise to chemical action.
He also showed that the values obtained by extrapolation, when a solution
of potassium chloride was used, agreed closely with those computed by
means of Henderson's formula, and that, if the degree of dissociation was
1 Z. physik. Chem., 53, 49 (1905).
2 Z. Elektrochem., 13, 17 (1907).
3 Z. physik. Chem., 59, 118 (1907); 63, 325 (1908).
4 Z. Elektrochem., 17, 58 (1911).
5 Ibid., 17, 388 (1911).
IO
calculated by means of Nernst's equation, the result did not at all agree
with the value obtained by calculation from conductance measurements.
The electrode and liquid potentials of nonaqueous solutions were studied
by Isgarischew¹ who applied Henderson's formula to methyl alcohol
solutions. He referred all his measurements to the normal hydrogen
electrode as a standard. He prepared a calomel electrode in methyl
alcohol, determined its electromotive force and, with the aid of this elec-
trode calculated the electrode potential of several metals. Certain
metals he found showed great tendency to become passive. The elec-
trode potential of silver was obtained by using as one-half of the cell a
Cd Cd(NO3)2 electrode which had been measured against the calomel
electrode. Since the potential of this cadmium electrode showed a ten-
dency to increase quite rapidly at first, the two half cells were not placed
in contact until a constant value had been obtained for the electromotive
force of this electrode.
Isgarischew2 made a further study of the passivity of metals and con-
cluded that the phenomenon was caused by oxidation. In the case of
copper, however, instead of an oxide, a complex compound of copper
chloride and alcohol was formed on the surface of the metal. He measured
the electromotive force of the combination
Ag | AgNO3 in methyl alcohol | AgNO3 in water | Ag,
and computed the liquid potential by means of Henderson's formula,
obtaining an electrode potential much larger than when the cadmium
electrode was used. He sought to measure similar combinations of ethyl
alcohol and aqueous solutions, and stated that it was impossible to apply
Henderson's formula to compute the liquid potential of such combinations,
but gave no reasons for this conclusion.
The following simplification of Henderson's formula has recently been
given by Cumming:3

RT K(2n-1) — K′(2n' — 1)
KW1-K'W2
E
=
F
loge
KW₁
K'W2
(4)
K and K' are the specific conductances, n and n' the migration ratios
and W₁ and W2 the corresponding valences. This equation requires only
data that can be determined experimentally; and, if it is assumed that the
migration ratios in the two solutions are the same, the equation becomes
identical with Nernst's equation. Cumming applied the equation to
the results of Bjerrum and others on the potentials of various chloride
solutions with calomel electrodes and obtained satisfactory values.
1 Z. Elektrochem., 18, 568 (1912).
2 Ibid., 19, 491 (1913).
3 Trans. Faraday Soc., 8, 86 (1913).

II
Since the experimental work to be described in the following pages
was completed, Cumming and Gilchrist¹ have published a method for
ascertaining the true values for the diffusion and electrode potentials
from the observed potential of the cell. They conclude that a new bound-
ary must be made shortly before the measurement is taken, and that
capillary tubes must be avoided in the construction of an electromotive
force cell.
Thus, as we have seen, considerable work has been done, especially
within the last three years, upon the subject of the potentials of metals
in aqueous and nonaqueous solutions of their salts. Many of these meas-
urements have been made with open cells connected by capillary tubes,
filter paper or sand, with "middle liquids" whose effect is more or less
uncertain, or with closed cells, no attempt being made to calculate the
potential at the junction between the solutions. It has, therefore, been
considered advisable to carry out an investigation of the behavior of silver
in nonaqueous solutions of silver nitrate. This combination was chosen
because silver nitrate can be obtained and kept in a state of great purity.
It is also soluble in a large variety of solvents. In some of these solvents
it might be expected, from conductance measurements, that abnormal
effects would be observed..
The development of the Cumming equation has made it possible to
compute the potential at the junction of the two liquids from purely ex-
perimental data, and has thus removed one source of error.
The apparatus employed for the electromotive force measurements was
free from the defects mentioned by Cumming and Gilchrist and embodied
several improvements over the ordinary type of cell. The investigation
divides itself into three parts as follows:
I. The measurement of the conductance of the solutions employed.
II. The determination of the transport numbers in as many of these
solutions as possible.
III. The measurement of the electromotive force of various concen-
tration cells in each solvent, and the determination of the electromotive
force of the following combinations:
Ag | 0.1 or 0.01 N AgNO3 in water | 0.1 or 0.01 N AgNO3 in solvent II | Ag
Preparation of Materials. The silver nitrate used in making up the
solutions was finely powdered and kept in a desiccator over phosphorus
pentoxide.
The ethyl alcohol (high-grade commercial "Absolute" 99.6%) was
dehydrated over lime ("aus Marmor") for several weeks. It was then
distilled, using a fractionating column, and the distillate was collected in
1 Trans. Faraday Soc., 9, 174 (1914).
12
a receiver protected from the moisture and gases of the atmosphere by
means of a tube filled with soda-lime. The boiling point was 78.8° at
757.4 mm. of mercury.
The methyl alcohol ("acetonfrei") was subjected to the same treatment
as the ethyl alcohol. The fraction boiling between 64.3° and 64.8° at
746 mm. of mercury was collected for use.
The acetone was dried over fused calcium chloride, distilled, allowed
to stand over anhydrous copper sulfate several weeks, then redistilled
and the fraction boiling between 56°-56.1° was collected. Its conductance
was so slight it could not be measured.
Commercial aniline, was distilled, allowed to stand over fused potassium
carbonate several days, then redistilled using a fractionating column.
The distillation was repeated until a nearly colorless distillate was ob-
tained which boiled between 181.5°-182° at 751.8 mm. of mercury. The
conductance was too slight to be measured.
The pyridine, of the best grade obtainable, was treated with fused
potassium hydroxide for several days, filtered and after repeated frac-
tionations the portion boiling between 114°-116° at 742 mm. of mer-
cury was collected.
The water used was distilled according to the method of Jones and
Mackay. It was collected in a bottle well protected from the gases of
the atmosphere.

2. Electrical Conductance.
All measurements were made at 25°. The thermostat consisted of
two concentric galvanized iron tanks, the annular space between them
being filled with sawdust. The water was well stirred and was heated
by means of an incandescent lamp that had been coated with a ruby varnish
to prevent the reducing action of light on the silver nitrate solutions.
The temperature was maintained constant at 25 0.05° by means of
an electrically controlled Ostwald thermoregulator.
#
The mother solution in each solvent was made up by direct weighing,
and the more dilute solutions were made from this by means of carefully
calibrated burets and flasks.
The measurements of electrical conductance were made by means of
the well-known Kohlrausch apparatus, a Leeds and Northrup cylindrical
bridge being used. The cells were of two types. The ordinary Arrhenius.
form closed by a tight-fitting ebonite cover was used for most of the work,
but the form described by Kreider and Jones¹ having a ground glass
stopper and concentric cylindrical platinum electrodes, was used for the
solutions in ethyl alcohol. The constants of these cells were determined
by means of a potassium chloride solution containing one mol of the salt
1 Am. Chem. J., 45, 295 (1911).

13
in 128 liters of water. Its equivalent conductance was taken as 142.4
at 25°. The cell constants were redetermined frequently during the
investigation and were found to remain practically constant. The average
value for the Arrhenius type of cell was 16.83 and for the Kreider and Jones
type 2.63. The electrodes of the cell were platinized in the usual manner.
After each measurement, the cell and electrodes were carefully cleansed,
rinsed first with distilled water, then several times with absolute ethyl
alcohol and dried in a current of pure dry air.
The cell was allowed to stand in the thermostat for twenty or thirty
minutes and the contents were thoroughly stirred to insure a uniform
temperature before a measurement was taken. From three to six readings
were taken for each dilution. The cell was allowed to stand twenty minutes
longer in the thermostat and the readings repeated. If the two sets of read-
ings did not agree, the cell was freshly filled and another set of readings taken
for comparison. Tables I to V give the results of these measurements.

TABLE I.-SILVER NITRATE IN ETHYL
ALCOHOL.
TABLE II.-SILVER NITRATE IN METHYL
ALCOHOL.
m.
V.
O.I
10.0
A25°.
10.81
m.
V.
A 25°.
O.I
IO.O
36.64
0.01
100.0
22.06
0.01
100.0
68.75
0.0078
128.4
23.72
0.006
166.7
74.88
0.003
333.3
28.14
0.003
333.3
83.14
0.001
1000.0
34.95
0.0012
833.3
87.30
0.00056
1786.0
37.79
0.001
1000.0
90.10
0.0003
3333.3
38.52
0.0001
10000.0
40.71
0.0003
0.0001
3333.3
87.82
10000.0
73.23
TABLE III.-SILVER NITRATE IN ACETONE. TABLE IV.-SILVER NITRATE IN ANILINE.
m.
V.
A25°.
m.
V.
0.01
100.0
10.51
O.I
10.0
A25°.
0.666
0.003
333.3
II.39
0.01
100.0
0.327
0.001
1000.0
15.43
0.003
333.3
0.436
0.0003
3333.3
20.78
0.001
1000.0
0.678
0.00016
6250.0
25.60
0.0001
10000.0
28.06
0.0003
0.0001
3333.3
1.082
10000.0
1.651
TABLE V-SILVER NITRATE IN PYRIDINE.
m.
V.
A 25°.
O.I
10.0
24.80
0.01
100.0
33.85
0.003
333.3
43.68
0.001
1000.0
54.13
0.0003
3333.3
66.14
0.0001
10000.0
73.50
0.00006
16666.7
75.82
3
In Fig. I are plotted the values given in Tables I-IV. As a comparison,
the results of other investigators are also given. The scale used in Fig. 1
14
being too small to show well the changes of conductance with dilution
in solutions of silver nitrate in aniline, the results of Table V are plotted

150
ΙΟΟ
Equivalent Conductance, A
3
4
8
50
ΙΟ
O
5
1, Conductance in Water.
5
6
Volume
ΙΟ
15
Fig. 1.
20
2, Conductance in Ethyl Alcohol.
tance in Methyl Alcohol. 5, Conductance in Acetone at 18°.
tance in Acetone at 25°. 8 and 9, Conductance in Pyridine.
2
7 Legend
Gibbons
Jones and Co-workers
Laszczynski
xLincoln
25
3 and 4, Conduc-
6 and 7, Conduc-
in Fig. 2 and compared with the results obtained by Sachanov¹ in more
concentrated solutions. Curves similar to this conductance curve have
been described by Franklin and Gibbs2 for silver nitrate in solutions in
2.4
2.0
1.6
1.2
0.8
0.4
Equivalent Conductance, A
O
5
IO
15
3√Volume
Fig. 2.-Conductance in Aniline.
1 Z. physik. Chem., 83, 129 (1913).
2 J. Am. Chem. Soc., 29, 1389 (1907).
20
x
Legend
o Gibbons
Sachanov

15
methylamine, and by Shinn' for solutions of silver nitrate in ethylamine.
The conductance curve for silver nitrate in acetone at 18°, described
by St. V. Laszizynski,2 is also plotted in Fig. 1. As it lies above the curve
obtained at 25° it indicates that the silver nitrate solutions in acetone
have a negative temperature-coefficient. This phenomenon was observed
by Cattaneo for solutions of hydrochloric acid in ether, and by Getman
and Gibbons for solutions of zinc chloride in methyl alcohol.
3. Transport Numbers.
Since the data at hand, concerning the transport numbers of the ions
of silver nitrate in nonaqueous solutions, were somewhat meager, it seemed
best to carry out as full a series of such measurements as possible for the
solutions under investigation.
Two forms of apparatus were employed for these determinations--the
form designed by Mather and modified by Jones and Basset,5 and a modified
form of the Nernst-Loeb apparatus which has been de-
signed in this laboratory and is illustrated in Fig. 3.
The electrodes were in both cases made of pure silver
wire fused to a stout piece of platinum wire which was
in turn sealed into the glass tube and protected with
fusion glass.
A copper coulometer was used to measure the total
current. Two copper plates, 2.2 by 5 cm. in dimen-
sions, served as electrodes, one of which, the cathode,
could be removed for cleaning and weighing. The
solution in the coulometer was made up as follows:
CuSO4.5H2O, 150 g.; H2SO4, 50 g.; C2H5OH, 50 g.;
H2O, 1000 g. During the passage of the current, a
slow stream of carbon dioxide bubbled through the
solution between the electrodes in the coulometer. The
laboratory electrical circuit served as a source of current
and it remained fairly constant during the time of each experiment.
The Jones and Bassett cell was carefully calibrated. To simplify the
method given by them, the apparatus was leveled and filled with distilled
water to a suitable height. After marking each arm of the apparatus to
indicate this level, it was emptied. and thoroughly dried. The stopcock
was closed and each arm filled with distilled water from a carefully cali-
brated buret. The volume required to fill each arm to the mark was
1 J. Phys. Chem., II, 537 (1907).
2 Z. Elektrochem., 2, 55 (1895).
3 Atti accad. Lincei, [5] 2, 295.
4 Am. Chem. J., 48, 124 (1912).
Fig. 3.
5 Ibid., 32, 429 (1904).
16
noted. The stopcock was then opened and water added until it again
stood at the same level. This operation was repeated three times with
the following results:
I.
Cc.
II.
Cc.
III.
Cc.
55.56
55.83
112.21
55.54
55.80
112.23
55.45
55.76
112.24
55.45
Average, 55.50
55.80
112.23
For each experiment 112.23 cc. of the solution were run into the ap-
paratus from the buret and after inserting the electrodes it was placed
in the thermostat. When sufficient time had elapsed for the temperature
to become uniform, it was connected in series with the coulometer, a vari-
able resistance and a milliammeter. The circuit was closed and the cur-
rent allowed to pass through the solution several hours. When a sufficient
amount of silver had separated at the cathode, the stopcock was closed,
the cell was removed from the thermostat and the two solutions were
filtered through glass wool into two 100 cc. flasks. Each arm was thor-
oughly rinsed several times and the wash water was added to the corre-
sponding solution. After the solutions had acquired the temperature
of the room they were diluted to 100 cc. and analyzed by Volhard's method.
The ammonium thiocyanate solutions used were o.1 and 0.025 N and were
standardized against a corresponding aqueous solution of silver nitrate.
The amount of silver in the solution was compared with the amount orig-
inally present.
This method was used for the measurement of the transport numbers
of most of the solutions in methyl and ethyl alcohols. The analysis of
the solution from the cathode chamber seldom gave the same value for
the transport number as the analysis of the solution from the anode cham-
ber. If any silver peroxide formed on the anode, the amount was very
slight and the silver separated out in beautiful, shiny needles on the cath-
ode. By taking the mean of the two values so obtained, the agreement
of results was fairly good.
It seemed probable that the cause of the irregularity observed was that
the unaltered portion of the solution did not lie midway between the elec-
trodes. This would seem logical when the ions do not share equally in
carrying the current. Schlundt¹ found that in the case of solutions of
silver nitrate in pyridine and acetonitrile the unaltered portion or "middle
layer" was much nearer the anode than the cathode.
This experience led to the designing of the form of apparatus illustrated
1 J. Phys. Chem., 6, 159 (1902).

17
in Fig. 3.
It was used in a few of the experiments with solutions in the
alcohols and with the solutions in pyridine. The apparatus was filled
to a level above the horizontal connecting tube and the electrodes were
inserted. The stoppers carrying the electrodes had a slight groove cut
on one side to allow for the expansion of the contents of the cell, but in
no case did the solution come in contact with these rubber stoppers. The
outlet tubes of the two arms were protected by means of rubber caps
sealed with glass rods. The apparatus was then placed in the thermostat
and, after uniform temperature was acquired, connection with the circuit
was made as in the previous experiments.
When sufficient silver had separated on the cathode, the following
method was adopted to ascertain the exact position of the "middle layer."
The solution was drawn from the anode arm into three or four tared glass
stoppered weighing bottles, and the entire solution from the cathode arm
was transferred to another weighing bottle. These portions were weighed
and their content of silver determined by Volhard's method. These
values were then compared with the amount of silver which would have
been present if the solutions had maintained their original concentration.
If there was a gain in silver, that portion was considered a part of the
solution around the anode; if there was a loss in silver, that portion was
considered a part of the solution around the cathode. If the change in
the amount of silver present was less than 1%, that portion was considered
the "middle layer." However, the amount of gain was added to the value
for the solution around the anode or the amount of loss was added to the
value for the solution around the cathode.
As an example, let us consider the measurements made with the 0.05 N
solution of silver nitrate in ethyl alcohol.

Wt. of
solution.
Orig. wt.
of silver.
I.....
II.
9.2174
0.06206
Final wt.
of silver.
0.10147
Change in
silver.
+0.03941
•
3.3788
0.02567
0.02533
+0.00034
III.
IV.
.
3.0431
0.02063
0.02045
+0.00018
9.1731
0.06229
0.05254
-0.00975
•
V.
19.8476
0.13486
0.10598
-0.02888
Total solution in anode chamber.
16.048 g.
Transport number of anion.....
0.613
Total solution in cathode chamber...... 29.021 g.
Transport number of anion....
0.593
Mean, 0.603
In this case, it is quite evident that if the division had been made ex-
actly in the middle of the solution, accurate values could not have been
obtained.
18
Tables VI to VIII give the results of the values obtained for the various
solutions:
m = concentration of the solutions (made up accurately by weight).
duration of the experiment in hours.
t
E
=
=
reading of the milliammeter in milliamperes.
A = the transport number of the anion (calculated from the analysis.
of the solution around the anode).
C
=
the transport number of the anion (calculated from the analysis
of the solution around the cathode).
TABLE VI.-SOLUTIONS IN ETHYL ALCOHOL.
Apparatus.
m.
1.
E.
A.
C.
Mean.
Mean of
series.
Jones and Bassett's.
O.I
4
O. I
5
O.I
5
3-4 0.626 0.707 0.667
3-4 0.668 0.683 0.675
3-4 0.632 0.718 0.676
0.673
Jones and Bassett's..
0.05
8
0.05 ΙΟ
0.05
6
Gibbons'..
0.05 42/3
Jones and Bassett's......... 0.04 ΙΟ
0.04 ΙΟ
2.5-3 0.751 0.752 0.751
2.5-3 0.697 0.729 0.713
2.5-3 0.724 0.744 0.734
4 0.613 0.593 0.603
2.5-3 0.728 0.693 0.710
2.5-3 0.705 0.717 0.711
0.733
0.603

}
0.7105
TABLE VII.-SOLUTIONS IN METHYL ALCOHOL.
Apparatus.
m.
t.
E.
A.
Gibbons'..
O.I
3
O.I
3
C.
4-5 0.587 0.570 0.578
4-5 0.584 0.572 0.578
Mean.
Mean of
series.
0.578
Jones and Bassett's...
Jones and Bassett's...
O.I
3.5
5-8 0.765 0.571
0.571 0.668
0.617
0.07
3
Gibbons'...
0.07 3
0.07 4
55
Jones and Bassett's....
0.05
0.05
33
3.5
Jones and Bassett's....
0.04
0.04
34
4
5.5
0.04 4
0.04 4
5.5
5.5
5 0.637 0.607 0.617
5 0.767 0.600 0.684
4-5 0.580 0.568 0.574
5-6 0.561 0.604 0.582
5-6 0.572 0.587 0.579
5-6 0.560 0.572 0.566
0.616 0.622 0.619
0.576 0.572 0.574
0.573 0.541 0.557
}
0.650
0.574
}。.
0.581
0.579
Apparatus.
Gibbons'...
m.
t.
E.
O.I
4
O.I
4
TABLE VIII.-SOLUTIONS IN PYRIDINE.
Mean.
Mean of
series.
C.
4-6 0.675 0.677 0.676
4-6 0.699 0.701 0.700 0.694
A.
O.I
4
4-6 0.698 0.717 0.707
Gibbons'...
0.05
0.05
54
3-5 0.667 0.691 0.679
3-5 0.656 0.697 0.676
0.677
19
Discussion of Above Tables.
Ethyl Alcohol. Since the differences in the values obtained when Jones
and Bassett's apparatus was used were so great, only the value obtained
when my apparatus was used will be compared with the values ob-
tained by other investigators.
Methyl Alcohol.-The values for o.1 and 0.07 N solutions obtained with
Jones and Bassett's apparatus are considered much less accurate than the
other values given in the above tables.
Acetone. Attempts were made to determine experimentally the trans-
port numbers in acetone solutions, but the maximum concentration that
could be obtained even by shaking silver nitrate with acetone for 24 hrs.
was only 0.01 N and the resistance of this solution was so great that the
measurements were not considered trustworthy. This was the conclusion
arrived at by Jones and Rouiller¹ in their work. Roshdestwensky and
Lewis, from their electromotive force measurements of concentration
cells, calculated the transport number for the anion of silver nitrate in
acetone as 0.60 to 0.58 for dilutions ranging between 0.02 and 0.007 N.
Jones and Rouiller determined the transport numbers for silver nitrate
in mixtures of acetone and water. By extrapolation from the curves
obtained by plotting their results, it seems probable that the value for
a 0.01 N solution in pure acetone might lie between 0.60 and 0.62.
Pyridine. The pyridine solutions could not be analyzed in the same
way as an aqueous or an alcoholic solution, because of the basic and solvent
properties of the pyridine. Two methods of eliminating the pyridine
were tried: (1) the pyridine was neutralized with concentrated nitric
acid; or (2) the greater part of the pyridine was distilled off from the
solution before neutralizing it. The mixture of silver and pyridine nitrates
was diluted with water and titrated with ammonium thiocyanate solu-
tion. The first method was less troublesome and gave fully as satisfac-
tory results.
Aniline. The analysis of the aniline solutions proved to be an in-
surmountable obstacle to the determination of the transport numbers
in that solvent. All known methods were found to be impracticable.
Comparison of the Values as Given by Various Investigators.
TABLE IX.-SOLUTIONS IN ETHYL ALCOHOL.

t°.
m.
A.
20
0.108
0.594
Investigator.
Mather
25
0.076-0.101
0.5988
Jones and Bassett
25
0.02
0.616
Rouiller
25
0.05
0.603
Gibbons
1 Am. Chem. J., 36, 475 (1906).
20
TABLE X.-SOLUTIONS IN METHYL ALCOHOL.
1º.
m.
A.
?
?
0.533
Investigator.
Campetti
25
0.0093
0.523
Carrara
25
0.10
0.5797
Jones and Bassett
25
0.02
0.572
Jones and Rouiller
25
0.10
0.578
Gibbons
25
0.07
0.574
Gibbons
25
0.05
0.581
Gibbons
25
0.04
0.579
Gibbons
TABLE XI.-SOLUTIONS IN PYRIDINE.
21
O. I
0.613
Schlundt
22
0.025
0.564
Schlundt
22
0.025
0.554
Schlundt
25
O. I
0.694
Gibbons
25
0.05
0.677
Gibbons
4. Electromotive Force.
The electromotive force of cells of the following types was studied:
(1)
Ag|AgNO3, Conc. I | AgNO3, Conc. II | Ag,
the solvent in the two solutions being identical, and
(2) Ag|AgNO3, Aqueous | AgNO3, Nonaqueous | Ag,
the solvents being different but the concentrations being identical. The
more concentrated solutions were all made up at room temperature by
direct weighing. The dilutions were made from these by means of care-
fully calibrated burets and flasks. The solutions were kept in the dark,
and in no case was there more than a very slight tendency to turn brown,
even when left standing several weeks.
The compensation method of Poggendorff was employed, an enclosed
type of Lippmann electrometer being used as a zero instrument. Measure-
ments with this instrument were correct to one millivolt, except when the
poor conductance of the solutions or other disturbing factors introduced
a larger error. The experimental cell was balanced against a chloride
accumulator as a source of potential. The electromotive force of this
cell was determined by means of a certified Weston Standard Element,
before and after each measurement of the experimental cell. The elec-
tromotive force of this
Standard Element was
1.01948 volts at 25°.

The experimental cell, il-
lustrated in Fig. 4, con-
sisted of two parts con-
nected by means of a tight-
fitting piece of rubber tub-
ing. During an experiment,
the stopcock S, of fairly
large bore, was kept closed,
C
R
Fig. 4.

21
except while the measurements were actually being made. The electrodes
were made of pure silver wire sealed to a stout piece of platinum wire,
which, in turn, was sealed into a glass supporting tube and protected by
The electrical connections were made through mercury
placed in these glass tubes.
fusion glass.
At first, some difficulty was experienced in obtaining constant values
for the electromotive force of a given concentration cell, but by adopting
the following method of procedure the electromotive force remained
constant for at least 30 min. after contact between the two solutions had
been made. The two half-cells were connected together at C and with
the two stopcocks S and S' closed, the two solutions were poured into the
cell until they were at the same level on each side, the denser solution being
always put into the half-cell R. When the electrodes had been inserted
and the clamps T and T' closed, the cell was supported in the thermostat
and allowed to stand 25-30 min. in order to acquire a uniform tempera-
ture. Connection with the circuit was then established and the contact
between the two solutions made as follows. The stopcock S' and clamp
T' were opened and, by gently blowing into a pipet inserted at T', the
denser solution in R was forced up to the base of the tube carrying the
stopcock S'. The clamp T' was then closed and the stopcock S and clamp
T were opened. By blowing very gently into a pipet inserted at T,
the solution in L was brought into contact with the first solution. The
stopcocks S and S' and the clamp T were then closed. A measurement
TABLE XII.
mi.
m2.
E.
K.
mi.
m2.
Solutions in Ethyl Alcohol.
O.I
O.OI
0.0408
0.0591
O.I
0.003
0.0650
0.0587
0.01
0.01
E.
Solutions in Acetone.
0.003 0.0293 о.обої
0.001
0.0537 0.0630
K.
O.I
0.001
0.0922 0.0619
0.01
0.0003
0.0816
0.0665
O.I
0.0003
0.1297 0.0658
0.01
0.0001
0.1212
0.0771
O.I
0.0001
0.01
0.01
0.003
0.001
0.1648
0.0276
0.0538 0.0657
0.0680
Solutions in Aniline.
0.0640
O.I
O.I
0.01
0.003
0.0373
0.0285
0.0548
0.0321
0.01
0.0003
0.0858
0.0670
Solutions in Pyridine.
0.01
0.0001
0.1169 0.0674
O.I
0.01
0.0368
0.0425
Solutions in Methyl Alcohol.
O.I
0.003
0.0592
0.0460
O.I
0.01
0.0462 0.0636
O.I
O.I
0.003
0.0746
0.0639
O.I
0.001
0.0003
0.0831 0.0500
O.I
0.001
0.1008 0.0626
O.I
0.0001
0.1105
0.1375 0.0544
0.0577
O.I
0.0003 9.1358 0.0634
O.I
0.00006 0.1532 0.0560
O.I
0.01
0.01
0.01
0.01
0.0001
0.003
0.001
0.0003
0.0001
0.1719
0.0264 0.0600
0.0591 0.0670
0.0934
0.1337
0.0637
0.01
0.01
0.01
0.003
0.001
0.0003
0.0207
0.0437
0.0502
0.0549
0.0723 0.0587
0.0675
0.0677
0.01
0.0001
0.0978
0.0588
0.01
0.00006 0.1179 0.0630
22
of the electromotive force was made at once.
The stopcock S was opened
just long enough to determine the point of balance on the bridge wire.
Two other readings were made at intervals of ten minutes. The elec-
tromotive forces calculated from these three bridge readings seldom varied
from the mean value by more than one- or two-tenths of a millivolt.
The results of the measurements are given in Table XII. The constant
K was calculated from the Nernst formula, written in the following form.
The value of the transport number was assumed not to vary appreciably
with the concentration:
K =
E
log10 m1A1/m2A2
(5)
When it was assumed that the observed values of the transport numbers
were correct, and that they did not vary appreciably with the dilution,
the values for E calculated for the different cells showed, in no case, very
close agreement with those observed.
In Table XIII the calculated and observed values for the transport
number (n) of the anion of silver nitrate in the various solvents are given.
(n) is calculated from the formula

n =
ข
=
E
u + v 2 X 0.0595 log10 (M1A1/M2A2)
(6)
and the observed value is in each case the mean of the value given in Tables
IX-XI, excluding those found by Campetti and Carrara.
TABLE XIII.
mi.
m2.
n Calc.
n Obs.
mi.
m2.
Solutions in Ethyl Alcohol.
n Calc.
Solutions in Acetone.
n Obs.
O.I
0.01
0.496
0.603
O.I
0.003
0.493
0.01
0.01
0.003
0.001
0.505
0.621
0.541
....
O.I
0.001
0.520
0.01
0.0003
0.559
O.I
0.0003
0.553
0.01
0.0001 0.647
O.I
0.0001
0.571
Solutions in Aniline.
0.01
0.003
0.557
O.I
Ο.ΟΙ
0.239
0.01
0.001
0.565
O.I
0.003
0.490
0.01
0.0003
0.567
.....
Solutions in Pyridine.
0.01
0.0001
0.566
O.I
0.01
0.357
0.620
Solutions in Methyl Alcohol.
O.I
0.003
0.390
.....
O.I
0.01
0.534
0.577
O.I
0.001
0.420
O.I
0.003
0.537
O.I
.....
0.0003
0.443
O.I
0.001
0.526
O.I
0.0001
0.457
O.I
0.0003
0.533
O.I
0.00006
0.471
O.I
0.0001
0.535
0.01
0.003
0.422
0.01
0.003
0.504
0.01
0.001
0.461
0.01
0.001
0.563
0.01
0.0003
0.493
0.01
0.0003
0.554
0.01
0.0001
0.494
0.01
0.0001
0.569
0.01
0.00006
0.529
1 This value is extrapolated.

23
The electromotive forces of the combinations
Ago.1 N AgNO3 in water | 0.1 N AgNO, in solvent II | Ag
were measured and the results are tabulated below. The first column
designates the nature of the solvent II, E the observed potential, E1
the electrode potential in the aqueous solution, E2 the potential at the
junction between the two liquids which is calculated according to Cum-
ming's formula given on page 10, and E, is the electrode potential in
solution II. The electrode potentials are referred to that of the hydrogen,
electrode as zero. The electrode potential E, in aqueous solution is cal-
culated by means of the formula E Ep-0.0595 log m, in which the
value of E, is taken as 0.798 volt.
=
TABLE XIV.
Solvent II.
E.
E1.
E2.
Es.
Ethyl alcohol......
Methyl alcohol..
+o. 1004
+0.7331
+0.00217
+0.8314
+0.0967
Acetone...
+0.1712
Aniline.
Pyridine.
-0.2559
-0.3618
+0.7331
+0.7331 +0.00202
+0.7331
+0.7331 -0.00017
-0.00005
+0.8298
+0.9022
....
+0.3711
5. Discussion of Results.
An inspection of Figs. 1 and 2 brings to light certain differences between
the conductance curves for nonaqueous solutions and the curve for aqueous
solutions of silver nitrate. Considering the latter to represent a normal
curve of conductance, the curves for nonaqueous solutions present certain
abnormalities. This abnormality is most clearly shown in the curve for
solutions in aniline and less clearly in the curves for solutions in acetone
and pyridine. The curve for solutions in ethyl alcohol is quite similar
to that for solutions in water, except that in the more concentrated region
it appears somewhat flatter; while the curve for solutions in methyl al-
cohol shows a tendency to pass through a maximum in the more dilute
regions.
In seeking a cause for these abnormalities it is important to consider
the determination of the molecular weight of silver nitrate in these sol-
vents. In certain cases, at the temperature of the boiling solvent, there.
is considerable interaction between the solvent and the salt, which renders
the determination of the molecular weight extremely difficult. This
is especially true when aniline is the solvent and true to a lesser degree
in the other solvents. It is, however, a well known fact that these liquids
show a great tendency to combine with the ions and that there are also
present in such solutions polymerized molecules of the dissolved salt.
Jones¹ showed that sodium iodide, cadmium iodide, and ammonium thio-
cyanate were polymerized to an appreciable extent when in solution in
1 Am. Chem. J., 27, 16 (1902).
24
acetone. It has been shown by many investigators that in solutions in
ethyl and methyl alcohol the solute frequently exists in a polymeric state.
Getman and Gibbons¹ have shown this to be the case for zinc chloride.
The molecular weight of silver nitrate in pyridine, as determined by Werner-
Smujlow,2 is 165.42 (mean value of series ranging from 160.02-169.82)
but, as determined by Speransky and Goldberg, it is 207.5 (mean value
of series ranging from 199-210). This latter result was confirmed by
Schroeder¹ who found the molecular weight to be 203.6 (mean value of
series ranging from 186-214). By studying the equilibrium of the system,
silver nitrate in pyridine from -65° to +110°, Kahlenberg and Brewer5
discovered that these substances combine to form definite compounds
at different temperatures and that, between -24° and +48.5°, the solid
phase had the composition AgNO3.3C5H5N, while above 79° there was no
combination of the two substances, pure silver nitrate being in equilibrium
with the solvent. This shows clearly that the molecular complexity of
the solute at 25° would be even greater than at 115°, the boiling point
of pyridine.
That there is a possibility of three different reactions taking place in
solution where such complexes exist was discussed by Sachanov.6 The
complexes Cm may decompose into simpler molecules, or they may ionize
to form complex ions, or the simple molecules may ionize to form simple
ions. These reactions can be represented as follows:

Cm
Cm
mc,
A'+K',
CA+K,
where A' and K' denote the complex anion and cation, and A and K
denote the simple anion and cation. The first two equations would hold
true whether the complex was a polymerized molecule of the solute or
a compound of solute and solvent. Walden' considers that ionization
is caused and induced by the process of disaggregation of the polymerized
salt molecules. In case the solvent has a high dielectric constant, this
depolymerization into simple ions will be abrupt, and in a solvent with a
low dielectric constant this change will be more gradual or the effect of
the complexes will be more marked and prolonged. Arranging these
solvents in the order of their dielectric constants, we see that the degree
1 Am. Chem. J., 48, 124 (1912).
2 Z. anorg. Chem., 15, 18 (1897).
3 Z. physik. Chem., 39, 369 (1902).
4 Z. anorg. Chem., 44, 20 (1905).
5 J. Phys. Chem., 12, 283 (1908).
6 Z. physik. Chem., 83, 129 (1913).
7 J. Am. Chem. Soc., 35, 1649 (1913)

25
of the abnormality of the conductance curves follows practically the same
order. This is shown in the following tabulation:
Solvent.
Water...
Methyl alcohol.
Ethyl alcohol..
Acetone..
Pyridine.
Aniline...
D. C.
81.1
31.2
25.8
21.5
12.4
6.85
Sachanov and Prscheborowsky¹ investigated the conductance of various
electrolytes in six different solvents of very small dielectric constants.
Many electrolytes showed maxima in very concentrated solutions, but
in no case was a minimum observed. They concluded that this might
be due to the very small dielectric constant of the solvent.
The diminution in conductance with increasing molar concentration,.
as shown by extremely concentrated solutions of silver nitrate in aniline,
has been ascribed by Kraus2 to the influence of the increasing viscosity
of the solutions. Sachanov³ has shown that, if the equivalent conductance
A is multiplied by the ratio of the viscosity of the solution to that of the
solvent, n'/n, values are obtained which correspond to A for solutions where
the two viscosities are approximately equal.
As the temperature is increased the abnormalities in a conductance
curve diminish. Jones and his co-workers clearly stated the relations be-
tween the temperature and the conductance of an electrolyte in an aqueous
solution. They showed that (1) the temperature coefficient of con-
ductance of aqueous solutions of electrolytes is greater the greater the
hydrating power of the electrolyte, that (2) the temperature coefficients
of conductance for any given electrolyte increase with the dilution of the
solution and this increase is greatest for those substances with large hy-
drating power. These two statements led to the conclusion that the
decreasing complexity of the hydrates with rise of temperature is a very
important factor in conditioning the large temperature coefficients of
conductance shown by those substances which have a large hydrating
power. It would seem that the abnormalities observed in nonaqueous
solutions might also be explained in a similar way.
The values of the transport numbers calculated from the electromotive
force measurements of concentration cells are quite different from those
obtained experimentally. In all cases the experimental values are the
larger. The direct measurement could only be carried out with solutions
1 Z. Elektrochem., 20, 39 (1914).
2 J. Am. Chem. Soc., 36, 60 (1914).
4 "Hydrates in Aqueous Solution," Publ. Carnegie Inst., 60, pp. 156-7 (1907).
3 Z. physik. Chem., 83, 129 (1913).
26
of fairly large concentrations (not less than 0.02 N). It can be seen that
the values calculated for the extremely dilute solutions approach more
nearly the experimental values. The transport numbers, in all the sol-
vents for which experimental or calculated data could be obtained, show
an increase in the value for the anion with increasing dilution, with the
exception of the experimental values for solution in pyridine where the
opposite effect is apparent. The values for solutions in methyl alcohol
show a much greater degree of uniformity for calculated and observed
values as well as much better agreement between the two series. In the
determination of the transport numbers it was observed that the "middle
layer" did not lie midway between the electrodes. If one examines the
experimental data of Schlundt¹ for a solution of silver nitrate in pyridine,
one notices this same fact. He found a middle layer" less than one-
third the distance from the anode. The middle layer for the alcoholic
solutions was situated much nearer the median line of the cell. It has
been proved by Kahlenberg and Brewer2 that pyridine and silver nitrate
combine at ordinary temperatures to form complex molecules, and Mor-
gan and Kanolt³ found that in a mixture of water and pyridine a large
proportion of the pyridine traveled with the silver ions to the cathode.
LeBlanc, in his "Textbook of Electrochemistry" (page 76), calls atten-
tion to Kohlrausch's discovery that, with monatomic univalent ions,
the transport numbers of all solutes in aqueous solutions approach the
value 0.50 with increasing temperature. He states that at the same time
the difference in mobility of the two ions in each case actually increases.
The greatest shifting of the "middle layer" was observed when pyridine
was used as the solvent for the silver nitrate and definite compounds of
solvent and solute are known to exist in such solutions. If the cation
carries with it a part of the solvent, the concentration of the silver in the
cathode chamber will be abnormally low and a corresponding increase
will be observed in the anode chamber. It seems probable that the move-
ment of the solvent with the cation causes the shifting of the "middle
layer" toward the anode. If the solvent moves with the anion, the "middle
layer" should then shift toward the cathode. These ionic complexes
break down more or less completely with increasing temperature and cause
the transport numbers to approach a limiting value in aqueous solutions.
Probably, when sufficient experimental data have been accumulated, it
will be found that the transport numbers in nonaqueous solutions approach
similar limiting values.
If the transport numbers do not increase with the dilution of the solu-
1 J. Phys. Chem., 6, 159 (1902).

2 Ibid., 12, 283 (1908).
3 Z. physik. Chem., 48, 365 (1904); J. Am. Chem. Soc., 28, 572 (1906).

27
tion, K as calculated in Table XII should have a constant value, but it
is only in the case of solutions in methyl alcohol that K has even an ap-
proximate constancy. In the four other solvents, K increases steadily
in value with the increasing difference in the concentration of the solutions
in the cell. Likewise, the values of the transport numbers calculated in
Table XIII, show a steady increase with the changing difference of
concentration in the cell. Both of these phenomena may be due to the
same cause. The degree of the dissociation of the conducting complexes
into simple ions will increase as the solution is diluted. In the more con-
centrated solutions the molar concentration is not correctly given by the
weight of silver nitrate in a given volume of solution, hence mA is the true
value of the ionic concentration only in dilute solutions. The change in
value of this factor would cause an increase in the computed value for
both n and K (as is shown by Equations (5) and (6)). The fact that this
abnormality is greatest in the case of aniline and pyridine, where the effect
of the complexes is most clearly shown in the conductance curve, seems
to substantiate this view. The Nernst equation, therefore, is not a
true expression of all the factors which determine the difference of poten-
tial in these concentration cells.
In comparing the electromotive forces found when an aqueous solution
was joined to a nonaqueous solution of the same concentration, it is at
once noticed that, whereas the solutions in the alcohols and acetone are
positive toward the aqueous solution, the solutions in aniline and pyridine
are negative toward the aqueous solution. The value obtained by Is-
garischew¹ for the electrode potential of silver in solutions of silver nitrate
in methyl alcohol is slightly greater than the one given in Table XIV;
but he used the Henderson formula for calculating the potential at the
junction of the two liquids and this involved the use of A ∞ which is by
no means surely established. The value here given. therefore, seems
more probable, since only experimental data were used in this calculation.
6. Summary.
The results obtained may be summarized as follows:
1. Conductance measurements of solutions of silver nitrate in ethyl
alcohol, methyl alcohol, acetone, aniline and pyridine have been made
for concentrations ranging between o.1 and 0.0001 N. It has been shown
that the conductance curve for solutions in aniline approaches a second.
maximum as the solutions become more dilute. It has also been shown
that there is a relationship between the dielectric constant and the ab-
normalities of all the conductance curves.
2. The transport numbers of silver nitrate in ethyl alcohol, methyl
alcohol and pyridine have been determined experimentally.
1 Z. Elektrochem., 19, 491 (1913).
28
3. Measurements of the electromotive force have been made for a large
number of concentration cells in the various solvents, an apparatus being
used which is free from defects due to capillarity and constant communica-
tion of the two liquids.
4. Values for the electrode potential of silver in the various solvents
have been calculated.
5. The experimental data seem to prove that the abnormalities ob-
served in nonaqueous solutions of silver nitrate are due to the combination
of the solvent and the solute to form complex compounds which dissociate
more or less gradually.
7. Chronological Bibliography.
Difference of Potential.
1890. "Über die Potentialdifferenz zwischen zwei verdünnten Lösungen binärer
Electrolyte," Planck, Wiedemann's Ann., 40, 561.
1893. "Über die Potentialdifferenz zwischen alkoholischen und wässerigen
Lösungen eines und desselben Salzes," Campetti, Atti accad. Torino, 28, 61 and 228.
1894. "Über die Lösungstension von Metallen," H. C. Jones, Z. physik. Chem.,
14, 346.
1899. "The Electromotive Force of Concentration Cells," J. E. Trevor, J. Phys.
Chem., 3, 95.
1899. "Difference of Potential between Metals and Nonaqueous Solutions of
Their Salts," Kahlenberg, J. Phys. Chem., 3, 379.
1900. Ibid., 4, 709.
1900.
"The Solution Tension of Zinc in Ethyl Alcohol," Jones and Smith, Am.
Chem. J., 23, 379.
1902. "Über die elektromotorischen Kräfte zwischen einigen Metallen und den
Lösungen ihrer Salze in Wasser und in Methyl alkohol," Carrara und D'Agostini,
Atti R. ist. Veneto, 62, 793.
1903. "Über die Zusammensetzung der in Lösungen existirenden Silber-ver-
bindungen des Methyl- und Aethyl-amins," Bodländer und Eberlein, Ber., 36, 3945.
1905. "Über die elektromotorischen Kräfte zwischen einigen Metallen und den
Lösungen ihrer Salze in Wasser und in Methylalkohol," Carrara und D'Agostini,
Gazz. chim. ital., 35, I, 132.
1905. "Über die Elimination des Diffusionspotentials zwischen zwei verdünnten
wässerigen Lösungen durch Einschalten einer konzentrierten Chlorkaliumlösung,
Bjerrum, Z. physik. Chem., 53, 428.
1905. "Elektrochemie der nichtwässerigen Lösungen," Carrara, Gazz. chim. ital.,
37, I, 525.
1905.
1905.
"Concentration Cells in Liquid Ammonia," Cady, J. Phys. Chem., 9, 476.
"Zur Theorie der elektrolytischen Lösungstension," Brunner, Z. Elektro-

chem., 11, 915.
1905. "Die Spannungsreihe der Metalle," Kahlenberg. "Elektromotorische
Kräfte zwischen Metallen und Lösungen ihrer Salze in Wasser und Methylalkohol,"
Carrara und D'Agostini. "Criticism of Above Articles," Sackur, Z. Electrochem., II,
385.
1907. "Zur Eliminierung der Flüssigkeitspotentiale," Cumming und Abegg,
Z. Elektrochem., 13, 17.

29
1907. "Zur Thermodynamik der Flüssigkeitsketten," Henderson, Z. physik.
Chem., 59, 118.
1908. Ibid., 63, 325.
1908. "Elektrochemie der nichtwässerigen Lösungen,' Carrara.
1909.
"Uber elektrochemische Potentiale in nichtwässerigen Lösungemitteln,"
Neustadt und Abegg, Z. physik. Chem., 69, 486.
1911. "Differences of Potential between Cadmium and Alcoholic Solutions of
Some of its Salts," Getman, Am. Chem. J., 46, 117.
1911. "On the Temperature Coefficient of Concentration Cells, in which the Same
Salt is Dissolved in two Different Solvents," Laurie, Proc. Roy. Soc. Edinb., 31, 375.
1911. "Über die Gültigkeit der Planckschen Formel für das Diffusionspotential,"
Bjerrum, Z. Elektrochem., 17, 58.
"Ueber die Elimination des Flüssigkeitspotentials bei Messungen von
Elektrodenpotentialen," Bjerrum, Z. Elektrochem., 17, 389.
1912.
"The Potentials of Zinc in Alcoholic Solutions of Zinc Chloride," Getman
and Gibbons, Am. Chem. J., 48, 124.
1912. "The Electrochemistry of Solutions in Acetone. Part II. The Silver
Nitrate Concentration cell," Roshdestwensky and Lewis, J. Chem. Soc., 101, 2094.
1912. "Über die Normalen und Flüssigkeitspotentiale nichtwässeriger Lösungen,”
Isgarischew, Z. Elektrochem., 18, 568.
1913. "Berichtigung zu der Abhandlung- Über die Normalen und Flüssigs-
potentiale nichtwässeriger Lösungen," Isgarischew, Z. Elektrochem., 19, 491.
1913. "Über die Passivierung der Metalle und ihre Abhängigkeit vom Lösungs-
mittel," Isgarischew, Z. Elektrochem., 19, 491.
1913. "The Electromotive Force of Silver Nitrate Concentration Cells," Bell
and Feild, J. Am. Chem. Soc., 35, 715.
1913. "Elimination of Potential due to Liquid Contact," Cumming, Trans.
Faraday Soc., 8, 86.
1914. "The Potential Due to Liquid Contact," Cumming and Gilchrist, Trans.
Faraday Soc., 9, 174.
Conductance.
1894. "Die electrische Leitfähigkeit von einigen Salzen in Aethyl- und Methyl-
alkohol," Völlmer, Wiedemann's Ann., 52, 166.
1895. Über die Leitfähigkeit der Lösungen einiger Salze in Aceton," St. v. Laszc-
zynski, Z. Elektrochem., 2, 55.
1899. "The Electrical Conductivity of Nonaqueous Solutions," Lincoln, J.
Phys. Chem., 3, 457.
1903. "Über organische Lösungs- und Ionisierungsmittel," Walden, Z. physik.
Chem., 46, 126.
1904. "Conductance of Silver Nitrate in Water, Ethyl and Methyl Alcohols,"
Jones and Bassett, Am. Chem. J., 32, 420.
1906. "The Relative Migration Velocities of the Ions of Silver Nitrate in Water,
Methyl Alcohol, Ethyl Alcohol and Acetone, and in Binary Mixtures of these Solvents,"
Jones and Rouiller, Am. Chem. J., 36, 427.
1907. "The Electrical Conductivity of Methylamine Solutions," Franklin and
Gibbs, J. Am. Chem. Soc., 29, 1389.
1907. "The Electrical Conductivity of Solutions in Ethylamine," Shinn, J. Phys.
Chem., II, 537.
1907. "Hydrates in Aqueous Solution," H. C. Jones, Publ. Carnegie Inst., No. 60.
1908. "Elektrochemie der nichtwässerigen Lösungen," Carrara.
30
1909. "Limiting Conductivities of Some Electrolytes in Ethyl Alcohol," Dutoit
and Rappeport, C. A., 3, 269.
1911. "The Conductivity of Certain Salts in Methyl and Ethyl Alcohols at
High Dilution," Kreider and Jones, Am. Chem. J., 45, 282.
1912. "Lösungsmittel mit kleinen Dielektrizitätskonstanten," Sachanov, Z..
physik. Chem., 80, 13.
1913. "Die anomalen Veränderungen der Leitfähigkeit," Sachanov, Z. physik..
Chem., 83, 129.
1913. "Studies of the Electrical Conductance of Nonaqueous Solutions," Shaw,
J. Phys. Chem., 17, 162.
1913. "The Dielectric Constants of Dissolved Salts," Walden, J. Am. Chem. Soc.,.
35, 1649.
1914.
"Conductance and Viscosity of Electrolytic Solutions," Kraus, Ibid., 36, 35..
1914. "Einige Lösungsmittel mit kleinen Dielektrizitätskonstanten," Sachanov
and Prscheborowsky, Z. Elektrochem., 20, 39.
1914. "Verbindung des gelösten Körpers und des Lösungsmittels in der Lösung,"
Dhar, Z. Elektrochem., 20, 57.
Ionic Velocities.
1888. "Ueberführungszahlen und Leitvermögen einiger Silbersalze," Loeb and
Nernst, Z. physik. Chem., 2, 948.
1894. "Der Einfluss der Lösungsmittel auf die Beweglichkeit der Ionen," Campetti,.
Nuovo Cimento, [3] 35, 225.
1901. "A New Apparatus for Determining the Relative Velocity of Ions; with
Some Results for Silver Ions," Mather, Am. Chem. J., 26, 473.
1902. "On the Relative Velocities of the Ions in Solutions of Silver Nitrate in.
Pyridine and Acetonitrile," Schlundt, J. Phys. Chem., 6, 159.
1904. "Determination of the Relative Velocities of Ions of Silver Nitrate in Mix-
tures of Alcohol and Water," Jones and Bassett, Am. Chem. J., 32, 409.
1904. "Über die Verbindung der Lösungsmittel mit den Ionen," Morgan und
Kanolt, Z. physik. Chem., 48, 365.
1906. "The Compounds of a Solvent with the Ions," Morgan and Kanolt, J..
Chem. Soc., 28, 572.
1906. "The Relative Migration Velocities of the Ions of Silver Nitrate in Water,
Methyl Alcohol, Ethyl Alcohol, and Acetone, together with the Conductivity of Such.
Solutions," Jones and Rouiller, Am. Chem. J., 36, 427.
"Calculation of Ionic Velocities of Silver Nitrate in Acetone Solutions,"
Roshdestwensky and Lewis, J. Chem. Soc., 101, 2097.

Molecular Weight.
1897. "Molekulargewichtsbestimmungen in Pyridin," Werner und Schmujlow,.
Z. anorg. Chem., 15, 18.
1902. "The Molecular Weights of Certain Salts in Acetone," H. C. Jones, Am.
Chem. J., 27, 16.
1902. "Elektrolyse einiger anorganischen Salze in organischen Lösungsmitteln,"
Speransky und Goldberg, Z. physik. Chem., 39, 369.
1905.
"Verhalten von Silbersalzen in Pyridin," Schroeder, Z. anorg. Chem., 44, 20.
1908. "Equilibrium in the System Silver Nitrate and Pyridine," Kahlenberg
and Brewer, J. Phys. Chem., 12, 283.

VITA
I, Vernette L. Gibbons, was born in Franklin, N. Y., January 5, 1874.
My parents were Marshville Gibbons and Augusta Foote Gibbons. In
the autumn of 1892, after completing a course of study at the Delaware
Literary Institute, in Franklin, I entered Mount Holyoke College, from
which I received the degree of Bachelor of Science in 1896. The follow-
ing year was spent in teaching, after which I returned to Mount Holyoke
College and was, for two years, 1897-1899, Assistant and for two years,
1899-1901, Instructor in Chemistry. The degree of Bachelor of Arts
was conferred upon me by Mount Holyoke College in 1899. During the
following summer I studied six weeks at the Summer School of Cornell
University. I spent the year 1901-1902, at the University of Chicago.
studying Chemistry. From 1902-1906 I taught Chemistry at Wells
College, the first two years as Instructor and the last two years as Associate
Professor, having charge of the department during the absence of the
Professor in 1904-1905. I returned to the University of Chicago in Octo-
ber, 1906, and completed the work required for the degree of Master of
Science, which was conferred upon me in March, 1907. My thesis sub-
ject was "Acetoacetic Ester-Its Formation and its Reactions." I was
four years, 1907-1911, Professor of Chemistry at the Huguenot College,
Wellington, South Africa. In February, 1908, the University of the Cape
of Good Hope conferred upon me the degree of Master of Science (ad
eundem).
I was awarded the Resident Fellowship in Chemistry at Bryn Mawr
College for the year 1911-1912 and the President's European Fellowship
for the year 1912-1913. I spent the year studying at the University of
Munich. The Mount Holyoke '86 Fellowship was awarded me by that
college for the year 1913-1914.
My major subject has been Inorganic and Physical Chemistry and my
minor subjects have been Organic Chemistry and Physiology. My Pro-
fessors at the University of Chicago were Dr. Nef, Dr. Stieglitz, Dr. Alex-
ander Smith and Dr. McCoy. At the University of Munich, I attended
lectures in physiology, physics and chemistry and worked in the physiological
laboratory with Professor E. F. Weinland upon the research problem
"Eine Untersuchung ueber den Gaswechsel der Süsswassermuschel
Anodonta." In my work at Bryn Mawr, I have studied organic chemistry
with Dr. Kohler and Dr. Brunel, physiology with Dr. Moore, mathematics
with Dr. Helen S. Huff and physical chemistry with Dr. Getman. In
the American Chemical Journal for August, 1912, the results of my work
32
า
on "The Potentials of Zinc in Alcoholic Solutions of Zinc Chloride" were
published.
My dissertation has been carried out under the direction of Dr. Getman,
to whom I wish to express my gratitude for inspiration and guidance during
the two years spent at Bryn Mawr College.

UNIVERSITY OF MICHIGAN
3 9015 02437 3113