m

JSii-M

'■■'■' '■'.'■■•■■■:'' '
,•■' .<:■■■■■ •■

''.-■.■■■:•■'■■:■

".'■■■■■■■■■■■■'

:'"' ': ;' - '■'•'■''' ' •'.
■ .■;'>■ vv. •...-.^■■■..,;

■.'''■' ,:

'WiV

■■■''!'•

Kffi : ■

1 ,■'•::■•'■•:•'

-:•:

Si®

THE ABSORPTION SPECTRA OF SOLUTIONS AS AFFECTED BY

TEMPERATURE AND BY DILUTION: A QUANTITATIVE

STUDY OF ABSORPTION SPECTRA BY MEANS

OF THE RADIOMICROMETER

By

HARRY C. JONES and J. SAM GUY

WASHINGTON, D. C.
Published bt the Carnegie Institution of Washington

1913

o

CARNEGIE INSTITUTION OF WASHINGTON

Publication No. 190

3H-1L

PRESS OF GIBSON BROTHERS, INC
WASHINGTON, D. C.

PREFACE.

The effect of high temperatures on the absorption spectra of nonaqueous
solutions was worked out in the Johns Hopkins University and published
in monograph No. 160. By means of a form of apparatus devised by Dr.
Strong, this work has now been extended to aqueous solutions and the
results are herein recorded.

Our previous work on the absorption spectra of solutions, which has now
been in progress continuously for eight years, had shown that the effect of
dilution on absorption spectra is much less than had hitherto been supposed.
A form of apparatus and method of procedure were worked out by Professor
Anderson, one of my former coworkers in this field, and this method has been
applied, with unusual skill, by Dr. Guy, to the effect of dilution on absorp-
tion spectra. The results that he has obtained are also recorded in this
monograph.

The grating spectrograph as a means of studying absorption spectra
has now supplanted the prism spectroscope. The grating spectrograph,
however, has its limitations. The results are photographed. This means
that the method is limited to the range of the photographic plate. This is,
for the best plates, from about 0.2^t to 0.8ju. It is, however, very desirable
to study absorption spectra in the region of wave-lengths which are much
greater than 0.8^. For this purpose, some method had to be devised which
did not make use of the photographic plate. The i adiomicrometer was the
obvious instrument to use, if it could be built sufficiently sensitive and at
the same time with sufficiently short period. This has been accomplished
by Dr. Guy.

With this instrument the absorption spectra of a number of salts have
already been mapped, and some surprising results have been obtained in
reference to the relative absorption of free water as compared with water
of hydration.

It gives me pleasure to express our thanks to Dr. E. J. Shaeffer, who has
assisted in making the radiomicrometer readings during the second half
of the past year, and who has also aided in the chemical work. Dr. Shaeffer
will continue the work on the absorption spectra of solutions, using the
radiomicrometer. We are especially indebted to Professor A. H. Pfund for
a large number of valuable suggestions, and for frequent advice during the
progress of this week. Professor J. S. Ames has kindly placed ample space
at our disposal for carrying out this investigation.

I am deeply grateful to the Carnegie Institution of Washington for
financial aid in carrying out this entire work, and in publishing the results
obtained. Without this aid, the work recorded in monographs Nos. 60,
110, 130, 160, and herein could not have been done.

Harry C. Jones.

hi

CONTENTS.

Chapter I. Introduction 1

Chapter II. The Absorption Spectra of Aqueous Solutions as Affected

by Temperature 5

The Making of a Spectrogram 7

Neodymium Chloride in Water 8

Neodymium Bromide in Water 9

Neodymium Nitrate in Water 9

Neodymium Acetate in Water 11

Neodymium Sulphate in Water 12

Cobalt Chloride in Water 12

Praseodymium Chloride in Water 13

Praseodymium Nitrate in Water 13

Uranyl Nitrate in Water 14

Uranyl Sulphate in Water 14

Uranyl Acetate in Water 15

Chapter III. The Effect of Dilution on the Absorption of Light by

Solutions 17

Making a Dilution Spectrogram 18

Neodymium Chloride in Water 18

Neodymium Bromide in Mater 20

Neodymium Nitrate in Water 20

Neodymium Sulphate in Water 21

Neodymium Acetate in Water 22

Praseodymium Chloride in Water 23

Praseodymium Nitrate in Water 24

Uranyl Chloride in Water 24

Uranyl Bromide in Water 24

Uranyl Nitrate in Water 2.5

Chapter IV. The Absorption Spectra of Aqueous Solutions of Certain

Salts of Neodymium a.s Studied by Means of the Radiomicrometer . 29

Method of Procedure 31

Discussion of the Results 36

Possible Explanation 38

Chapter V. The Absorption of Light by Water Changed in the Presence
of Strongly Hydrated Salts, as Shown by the Radiomicrometer.

New Evidence for the Solvate Theory of Solution 43

Absorption of Free and Combined Water 43

Hydrated and Nonhydrated Substances 44

Method of Procedure 44

Results 45

Discussion of the Results 52

Explanation of the Results 54

Chapter VI. Absorption Spectra of a Number of Salts as Measured by

Means of the Radio.micrometer 61

Mode of Procedure 62

Description of Cells Used 63

Discussion of Tables and Curves 65

Neodymium Chloride in Water 65

Neodymium Nitrate 72

Neodymium Acetate 74

Praseodymium Chloride 76

Praseodymium Nitrate 78

Nickel Chloride 79

Nickel Nitrate 80

Nickel Sulphate 80

Salts of Cobalt 81

Chapter VII. General Summary of Results 85

v

THE ABSORPTION SPECTRA OF SOLUTIONS AS AFFECTED BY

TEMPERATURE AND BY DILUTION: A QUANTITATIVE

STUDY OF ABSORPTION SPECTRA BY MEANS

OF THE RADIOMICROMETER

By
HARRY C. JONES and J. SAM GUY

VII

CHAPTER I.

INTRODUCTION.

An investigation of the effect of temperature on the absorption spectra
of certain solutions has already been carried out by Jones and Strong.1
The apparatus used was devised by Professor John A. Anderson,2 who
worked somewhat earlier with Jones on the absorption spectra of solutions.
The solutions were heated in an open vessel, and the temperature could,
of course, not be raised much above 100° F. It was found that, even over
this range of temperature, the effect of rising temperature was to cause the
general absorption of any salt in water to increase, and also to cause the
bands to broaden and become more diffuse. The results were entirely
unambiguous so far as they went, but were limited by the boiling-points of
the solutions in question. Indeed, it was not possible to work quite up to
the boiling-point of the solution, since the change in the concentration of
the solution resulting from boiling would have been too great, and there
would have been too much gas formed on the quartz windows through which
the light was to pass.

We wanted to study the effect of rise in temperature on the absorption
spectra of solutions to as high temperatures as it was possible to go. For
this purpose closed forms of apparatus devised by Anderson3 and by Strong4
were employed by Jones and Strong5 for nonaqueous solutions. The appa-
ratus consisted of a gold-plated steel tube, whose ends were closed with
glass windows. This worked very well with nonaqueous solvents up to
temperatures of approximately 200° C. Usually before this temperature
was reached a precipitate formed in the tube, which prevented work at
higher temperatures.

Some interesting results were obtained at the higher temperatures with
this apparatus. The general effect of rise in temperature is to deepen the
color of the solution of an inorganic salt. This is usually due to a widening
of the absorption bands. For details in reference to the effect of tempera-
ture on the absorption of light by nonaqueous solutions, reference must be
had to the Carnegie Institution of Washington monograph,6 where the
results in question are published in full.

The apparatus used by Jones and Strong for nonaqueous solutions did
not work satisfactorily for solutions in water as the solvent. The water
vapor, under the high pressure produced within the apparatus, worked its

^arn. Inst. Wash. Pub. 130. Amer. Chem. Journ., 43, 37, 97 (1910); 45, 1, 113 (1911).

2Carn. Inst. Wash. Pub. 110, p. 20. Amer. Chem. Journ., 41, 276 (1909).

3Carn. Inst. Wash. Pub. 160, p. 28. Amer. Chem. Journ., 47, 30 (1912).

4Carn. Inst. Wash. Pub. 160, p. 29. Amer. Chem. Journ., 47, 30 (1912).

6Carn. Inst. Wash. Pub. 160. Amer. Chem. Journ., 47, 27, 126 (1912).

6 Cam. Inst. Wash. Pub. 160.

1

2 ABSORPTION SPECTRA OF SOLUTIONS.

way through the layer of gold laid down on a layer of copper electrolytically,
rusted the steel, and caused the separation of the gold from the steel walls.
To avoid this, the apparatus, which was designed by Dr. Strong,1 was made
of brass and will be described in some detail in this monograph. It was
plated electrolytically with gold and this adhered firmly to the brass, even
when the aqueous solution contained in the apparatus was heated to 200° C.
We could work as satisfactorily with this apparatus with aqueous solutions
as with the former apparatus with nonaqueous solutions.

The work described in this monograph on absorption spectra of aqueous
solutions at high temperatures was all carried out in the gold-plated brass
apparatus. The results obtained and the bearing of these results on the
nature of solution will be discussed later in this monograph. Suffice it to
say here that up to 200° the effect of temperature on the absorption spectra
of aqueous and nonaqueous solutions has now been studied pretty exten-
sively on a large number of salts and a fairly large number of solvents.

The effect of dilution on the absorption spectra of solutions was taken
up with the following idea in mind: It was long a question as to what is
the nature of the absorber of light, say in aqueous solutions. It was at one
time supposed that chemical molecules were the absorbers, since these were
regarded as the ultimate units in solution. It was supposed that the mole-
cules were thrown into resonance by certain wave-lengths of light, and that
these were, consequently, stopped ; while the remaining wave-lengths passed
through the solution and gave to it its characteristic color.

When the theory of electrolytic dissociation was proposed in 1886, the
view as to the nature of solution of electrolytes underwent a serious change.
When electrolytes were dissolved in water, or in any other dissociating
solvent, they dissociated into charged parts or ions, and these were the
ultimate units in solution. If the solution was fairly concentrated we had
both ions and undissociated molecules in the solution, and the question in
such cases was, which is the absorber?

It was further recognized that a dilute solution of salt often has very
different color from a concentrated solution; and, moreover, solutions of
nonelectrolytes are often colored, i. e., have the power to absorb certain
wave-lengths of light and to allow others to pass on through. It was sup-
posed, then, that molecules have the power to absorb light, and ions also
have absorbing power. When a concentrated and a dilute solution of an
electrolyte had the same absorption spectrum — the same color — it was
supposed that the chemical molecule and the ions resulting from it had the
same absorption. When the dilute solution had a different color from the
concentrated solution, it was thought that the ions were the chief absorbers
of light. And since it frequently happens that a dilute solution of an elec-
trolyte has a very different color from a more concentrated solution, it
was supposed that in dilute solutions of electrolytes the ions are the chief
absorbers of light; since in very dilute solutions of electrolytes there are

1 Cam. Inst. Wash. Pub. 160. Amor. Chem. Journ., 47, 30 (1912).

INTRODUCTION. 3

mainly ions and practically no molecules present, it is obvious that in such
solutions it is not the molecules which are absorbing light. It must be the
ions, since these are the only units present, or something contained within
the ions. This was the view of absorption of light introduced by the theory
of electrolytic dissociation.

We have now gone much farther than this. We now know that the ions
are not the ultimate units in a solution of an electrolyte. The simplest ion
is very complex. It is made up of a large number of electrons, which are
unit negative charges of electricity. There is every reason to-day to believe
that the electrons are the real absorbers of light, are the units which are
thrown into resonance by the various wave-lengths of light. Granting this,
there is still a difference between an ion and the atom or atoms from which
it was formed. An ion contains one or more free electrons within it or on
it, i. e., one or more negative charges than would correspond to the positive
electricity within the atom. It would be interesting to know whether the
free electron or electrons upon the ion have anything to do with its power to
absorb light. This can be tested by studying the absorbing power of mole-
cules and then the absorption of light by the ions which are formed when
these molecules dissociate. It was with this idea in mind that the second
chapter of the work described in this monograph was undertaken.

A concentrated solution of a salt contains many molecules, and if the solu-
tion is sufficiently concentrated there are chiefly molecules and only a few
ions present. As the dilution is increased the dissociation increases; the
number of molecules becomes less and less and the number of ions greater
and greater. The problem, then, is to photograph the absorption spectrum
of a very concentrated solution of a salt, the layer being, say, 0.5 cm. deep.
Then take the spectrum of a more dilute solution of the same salt; if the
dilution is increased 100 times the depth of layer used would be 50 cm.
Under these conditions there would be the same number of parts of dis-
solved substance in the path of the beam of light; in the second case there
would be more ions and less molecules than in the first. By comparing the
two spectra we could see whether there is any difference between the
absorbing power of ions and molecules, i. e., whether the free electrons upon
the ions have anything to do with their power to absorb light. We then
took another step, increasing the dilution of the second solution five times
and also increasing the depth of the layer of the solution through which the
light passed five times, i. e., making the depth 250 cm. This second diluting
still further reduced the number of molecules present and increased the
number of ions. By comparing the three spectrograms we ought to be able
to say whether molecules and ions have the same or different resonance
with respect to light-waves; and, if it is different, to point out in what the
difference consists. This would then enable us to determine whether the
free electrons upon the ions played any part in the absorption of light.

We shall see that ions have somewhat different absorbing powers from
molecules, and in what this difference consists will appear later from the
text and from the plates.

4 ABSORPTION SPECTRA OF SOLUTIONS.

The work done on the absorption spectra of solutions by Jones and
Uhler,1 Jones and Anderson,2 and Jones and Strong,3 which extended over
five years, and in which some 6,000 solutions were studied, all involved the
photographing of the various spectra. In this way the positions of the
various absorption lines and bands were determined.

A question even more fundamental than the positions of the lines and
bands is their intensities, and the relative intensities of different parts of
the same band. The photographic method gave only a means of dealing
qualitatively with this problem. Some general idea could be gained of the
relative intensities of the various lines and bands on the photographic plate,
but these changed with the time of exposure, the intensity of the light used,
and with other conditions, so that we were able to learn very little about
the intensities of the various lines and bands by means of the photographic
method.

Further, the photographic plate is sensitive only between the wave-length
2,000 Angstrom units and 7,600 a.u.,4 which is a comparatively small part of
the spectrum. It is especially important to work also into the region of the
infra-red.

A method was used which dealt quantitatively with the intensities of the
various lines and bands. This same method, instead of being limited by
the above-named wave-lengths, could be used down into the infra-red to
wave-lengths as great as 20,000 a.u. to 30,000 a.u. Indeed, the method can
be used for even greater wave-lengths, if solvents can be found that are
transparent to the longer waves. This method involves the use of the radio-
micrometer.

The description of the instrument which we built, the method of work,
and the results thus far obtained, will be found on pp. 29 to 93.

^arn. Inst. Wash. Pub. 60. Amer. Chem. Journ., 37, 126, 207 (1907).

2 Cam. Inst. Wash. Pub. 110. Amer. Chem. Journ., loc. cit.

3Carn. Inst. Wash. Pub. 130 and 160. Amer. Chem. Journ., loc. cit. 0

throughout this paper we have employed this expression to designate Angstrom units.

CHAPTER II.

ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS AS AFFECTED

BY TEMPERATURE.

Jones and Strong1 studied the effect of temperature on the absorption
spectra of various nonaqueous solutions up to nearly 200°. The solutions
were usually heated until a precipitate formed, which cut off the light and
prevented work at still higher temperatures. Some work was also done by
Jones and Strong on the effect of temperature on the absorption spectra of
aqueous solutions. This was, however, not large in amount and did not
extend to very high temperatures.

The reason that the work with aqueous solutions was not pushed to higher
temperatures was that the form of apparatus then in use did not admit
of it. This consisted of a steel tube,2 lined on the inside with copper and
plated with gold on all of the inner surfaces. This worked very satisfac-
torily with nonaqueous solutions, the gold plate adhering firmly to the
copper, which, in turn, remained adherent to the steel. When an aqueous
solution was heated in the apparatus from 100° to 200° the result was
unsatisfactory. The water, under the high pressure, forced its way through
the copper and the gold and rusted the iron, as has already been stated.
The result was that the copper, with the gold, separated from the steel, and
the solutions, after heating for a time, gave the iron reaction. This appa-
ratus had the further disadvantage, that when a precipitate formed with
rising temperature it was necessary to open the entire apparatus and remove
the glass ends in order to clean them.

To overcome these difficulties the apparatus shown in fig. A was con-
structed by Jones and Strong and used to study the effect of rising tempera-
ture on the absorption spectra of aqueous solutions. The quartz ends are
fastened into the ends E' . The plunger P has guide grooves instead of
guide pins. A part of the plunger is provided with screw-threads for remov-
ing it. The entire cap is removed from tube T by unscrewing E', during
which the quartz end is untouched. When the ends are removed the quartz
window can be easily cleaned. Gold washers were inserted between T and
E' and between E' and U.

The general arrangement of the apparatus is also shown in fig. A. The
cell is kept in a horizontal position, so that any bubbles that may form will
rise in the side tube. The spectroscope, containing the grating G, photo-
graphic-plate holder C, and slit S, being kept vertically, a 45° quartz prism
was used to change the horizontal beam of light into a vertical beam, the
beam being totally reflected by the hypothenuse surface of O. The source
of light NG (Nernst glower) or SG (spark gap) was focused by the concave

^mer. Chem. Journ., 47, 27, 126 (1912). Hbid., 47, 30 (1912).

5

6

ABSORPTION SPECTRA OF SOLUTIONS

speculum mirror M on the slit S. A similar arrangement was used for the
fused silica cell. DTS is a double-throw switch, by means of which either
the Nernst glower or the spark gap may be thrown in circuit. B is a bal-
last. R is a variable resistance, by means of which the current in the Nernst
glower, as shown by the ammeter A, may be kept constant. OC is an oil-
condenser. IC isjm X-ray induction coil and R* is a resistance in the
primary circuit of this coil.

Fig. A.

In a recent paper Merton1 states that he has studied the effect of pressure
on the absorption spectra of solutions. This was studied here by Jones and
Strong and the results published in the American Chemical Journal.2 The
quotation of a paragraph from our earlier paper will show what was found :

Some preliminary tests were made with the cells at high pressures. The Cailletet
pump belonging to the Johns Hopkins University was used for this purpose, the cell
being made so as to fit into this pump. It was not at all difficult to obtain pressures of
200 atmospheres with water and alcohol solutions. Spectrograms were made of the
absorption spectra of neodymium solutions under pressures as high as 275 atmospheres.
No effect of pressure was detected. The work at high pressures is easier than at high
temperatures, on account of the fact that there is an expansion of the cell due to heating.

It should be stated that in all of the work on absorption spectra which has
been carried out in this laboratory for the past seven years, a grating spec-
troscope has been used. The arrangement of the heated cell, the grating,
photographic plate, etc., will now be discussed.

1 Proc. Roy. Soc, (A) 87, 146. 2 Amer. Chem. Journ., 47, 32, January 1912.

AS AFFECTED BY TEMPERATURE. 7

THE MAKING OF A SPECTROGRAM.

The apparatus used throughout the entire study of the effect of high
temperature has already been discussed. Two cells were used, one 10 cm.
and the other 1 cm. in length, both having the same general design and
differing only in length.

The cell, placed in a bath suitable for keeping the temperature constant,
was arranged as indicated in the diagram, and the source of light so located
that the rays, reflected from a concave mirror, passed longitudinally through
the cell and formed an image of the Nernst glower on the slit of the camera.
The position of the prism is so adjusted as to fill the grating uniformly with
light. Holding the eye directly above the grating, in a position later to be
occupied by the photographic plate, we could easily tell when the light was
falling properly upon the grating. When the cell was correctly adjusted
the lights were extinguished and the photographic plate inserted. With
the plate in position, the light was turned on and an exposure made at room
temperature. The position of the plate was then moved a given distance,
and the temperature of the cell raised very slowly, this process being repeated
at intervals of about 20° or 25°.

It is clear that, with such pressures as are developed by heating water to
200°, it is very difficult to obtain a tight joint between glass and metal.
This difficulty, however, has been partly overcome by the special form of
apparatus designed by Dr. Strong and described on page 6 of this mono-
graph. We were not able to secure a closing that would hold above 200°,
but once a good closing was secured it was not necessary to remove the ends
for several operations.

Great care had to be taken in heating the cell, on account of the difference
in expansion of the glass ends, and the metal in contact with them. When
the temperature was raised more than 40° an hour the glass ends usually
broke. At such high temperatures as we were employing the glass was
rapidly attacked by the water; later, when we were using the clear uviole
glass, a single heating rendered the glass ends almost opaque, especially if
they were allowed to stand for any length of time.

It was found that in many cases precipitates would appear in the cell
at a temperature slightly above 100°. This precipitate, however, formed
rapidly, once it began, and almost as quickly disappeared. By properly
regulating the intervals at which exposures were made, the effect of the
precipitate could be avoided; hence this effect does not appear on any of
the strips photographed.

It is probable that slight hydrolysis took place at first, as

2NdCl3 + 3H20 = 2Nd(OH)3 + 3HC1

The presence, then, of a slight excess of hydrochloric acid would hinder the
reaction in the direction indicated above by the arrow. Since most hydrox-
ides lose water at temperatures above 100°, it is possible that the following
reaction would take place:

2Nd(OH)8 = Nd8Oa + 3H80

8 ABSORPTION SPECTRA OF SOLUTIONS

The neodymium oxide, being heavy and very slightly soluble in water,
settles to the bottom of the cell, and the solution clears up. In this way
it is evident that the solution becomes slightly more dilute as the tempera-
ture is raised ; but this would lessen the number of absorbers in the path of
the beam of light, and thereby produce a narrowing of the bands and could
only decrease the effect indicated on the plates. This antagonistic influence
could certainly not cause a widening of the absorption bands, with rise in
temperature.

NEODYMIUM CHLORIDE IN WATER. (See Plate 1.)

The solution whose spectrum is given in section A was saturated, the
depth of absorbing layer being 1 cm. The temperatures, beginning with
the strip nearest the numbered scale, were 20°, 45°, 70°, 95°, 115°, 140°, and
165°, respectively. Absorption bands which are unchanged by the range of
temperature from 20° to 200° appear at X3800, X4025, X4200, X4325, X4440,
X4600, X4690, X4750 and X4820. The double band from X5050 to X5270 is
only slightly affected, if at all.

The two most interesting absorption bands are those whose centers are
near X4275 and X5800. The former of these in strip 1 is very sharp and
intense, though only a few a. it. wide. Both edges were well defined. As
the temperature is raised the violet edge remains very sharp, while a rapid
shading off of the red edge takes place. At a glance the band appears to be
less intense in the higher temperature strips, but on close examination it is
seen to be more diffuse, the red edge diffusing over a range of about 20 a.u.
at the highest temperature. This is exactly in accord with what Jones and
Anderson1 had found. They showed that when the number of molecules
in the path of light was kept constant, this band remained practically con-
stant ; while it has been shown by Jones and Anderson and by ourselves that
this band changes with dilution, being more intense in the most concen-
trated solution.

The X5800 band is affected most by temperature as well as by dilution.
In strip 1 this band is about 200 a.u. wide, the width increasing regularly
as the temperature is raised, until at the highest temperature it is over
250 a.u., or there is a total widening of 50 a.u. The violet edge remains
perfectly sharp, while the shading is toward the red end of the spectrum.

It occurred to us that whatever effect might be produeed by a rise in
temperature, if it was a true temperature effect, the reverse should happen
when the solution was allowed to cool.

With this in view B was made. The concentration of the solution and
the depth of layer photographed in section B were exactly the same as in A.
In fact, the same solution was used. As soon as the film A had been exposed
with rising temperature, it was removed from the camera and developed.
Without even allowing the cell to cool, another film was placed in the camera

2Carn. Inst. Wash. Pub. 110.

AS AFFECTED BY TEMPERATURE. 9

and section B made with falling temperature. In B the temperatures were
165°, 140°, 115°, 95°, and 70°, the highest temperature being nearest the
numbered scale, which is not accurately adjusted.

A study of the original film shows changes only in bands X4275 and X5800;
and this change is exactly the reverse of that shown by these same bands in A.
The X4275 band appears in strip 1, with a sharp violet edge and shading off
toward the red over a range of 15 or 20 a.u. As we pass to the succeeding
strips in the direction of falling temperature the red edge becomes sharper
and sharper, until in strip 5, which represents the lowest temperature, the
band assumes its normal sharp edge on the red side and covers less than 10
a.u. The X5800 band narrows uniformly from the red end as the tem-
perature falls, the total narrowing being about 40 a.u.

NEODYMIUM BROMIDE IN WATER. (See Plate 2.)

The concentration of the solution used in making the negative for A was
1.66 normal; the depth of cell, 1 cm. The temperatures, beginning with the
strip nearest the numbered scale, were 20°, 45°, 70°, 95°, 120°, 140°, 175°, and
190°, respectively. This plate seems to have had just the proper length of
exposure for the given concentration, and every known neodymium absorp-
tion band appears on the negative in excellent condition. With the bromide,
as with the chloride discussed in plate 1, only X4275 and X5800 show appre-
ciable changes with rise in temperature. The X4275 band, which has both
violet and red edges sharp in strip 1, feathers out toward the red end of the
spectrum as the temperature is raised.

The X5800 band widens toward the red as much as 60 a.u. The concen-
tration of solution used in making B was 0.166 normal, one-tenth that of A;
the depth of absorbing layer was 10 cm. The temperatures, beginning with
the strip nearest the numbered scale, were 20°, 45°, 70°, 95°, 115°, 135°,
1 55° and 190°, respectively. This is probably the best negative produced in
this part of the work, and the bands X4275 and X5800 show well the char-
acteristic changes spoken of above. The widening of band X5800, though
well marked, is not so great as in A, the total change being about 40 a.u.,
as compared with 60 a.u. in the former. If such a band be due to mole-
cules this is what we should expect, since, B being a more dilute solution, the
total number of molecules is less than in A. Hence, any change associated
with molecules would be more clearly apparent in A. This is in accord with
changes produced in this same band by dilution.

NEODYMIUM NITRATE IN WATER. (See Plates 3 and 4.)

The solution used in spectrogram A, plate 3, was saturated, the depth of
cell being 1 cm. The temperatures, beginning with the strip nearest to the
numbered scale, were 15°, 40°, 65°, 115°, 140°, and 165°, respectively.

The exposures were not as long here as in the previous plates, in order to
bring out more clearly the group of bands between X4200 and X4800. The
change in the X4275 band is here especially marked. At 15° this band i<5
very sharp and intense, while at 165° it has become broad and hazy, being

10 ABSORPTION SPECTRA OF SOLUTIONS

about 30 a.u. wide. The X4425 band shows a widening of about 15 a.u.
over the range shown in this plate.

The broad bands with the centers near X5125 and X5800 show most marked
changes. In each case the most marked change is almost entirely toward
the red end of the spectrum, the violet edge of the band remaining almost
unchanged. This is the case especially with the X5800 band.

The concentration of the solution used in B, plate 3, was one-tenth satu-
rated, the depth of the cell being 10 cm. The temperatures, beginning
next to the numbered scale, were 20°, 45°, 70°, 95°, 120°, and 145°.

Although the total number of absorbers in B are the same as in A, yet it is
seen that the change in the bands is far greater m A,i. e., where the concen-
tration is greatest. Only the X5800 band shows appreciable change in B,
and even this does not widen more than 40 a.u.

The concentration of the solution used in making the negatives of A,
plate 4, was one-tenth of saturation, the depth of absorbing layer 10 cm.
The temperatures, beginning with the strip nearest the numbered scale,
were 20°, 45°, 70°, 95°, 115°, 140°, 165°, and 190°.

Aside from the slight tendency of all the absorption bands to become a
little more diffuse at the higher temperatures, though not more intense,
there is no marked change in any band except X4275 and X5800. The former
of these, as we go toward the higher temperatures, remains perfectly sharp
and constant on its violet edge, while there is a regular shading toward the
red end of the spectrum. Again, the greatest change takes place in band
X5800, the violet end remaining fixed and the red edge widening between the
first and last strips to the extent of about 50 a.u. All the exposures of this
plate were made as the temperature of the cell was raised.

The identical solution used in A was photographed in B, plate 4, the cell,
intensity of light-source, and all of the apparatus remaining unchanged,
the only difference being that the exposures of B were made at regular
intervals as the temperature of the cell was lowered. The temperatures of
the successive strips in B were, beginning with the strip nearest the num-
bered scale, 190°, 165°, 140°, 115°, 95°, 70°, 45°, 20°.

The original films shGw A and B to be exactly the reverse of each other.
Just those changes produced in A by a rise in temperature are reversed by
the corresponding fall of temperature in B. Of course this is only qualita-
tive, since we can establish no definite quantitative relations from the photo-
graphic plates. In order to do this, energy measurements must be made,
not only on each band, but on different parts of the same band. Such work
is now in progress. This would be very difficult to do with a narrow
band like X4275, but should be comparatively simple with band X5800.

Band X4275, which in strip 1 appears broad and hazy on its red edge,
gradually acquires the characteristic sharp intense edges as the temperature
falls, until in strip 8 it is only about 8 a.u. wide. The total change in band
X5800 is a narrowing of about 60 a.u. There is no sudden or decided change
between any two successive strips, but, on the contrary, so far as the photo-
graphic plate is able to show, the change is a gradual one.

AS AFFECTED BY TEMPERATURE. 11

NEODYMIUM ACETATE IN WATER. (See Plates 5 and 6.)

In plate 5 we have photographed the change in the absorption bands of
neodymium acetate, produced by rise in temperature, section A, and by the
corresponding lowering of temperature, section B. The concentration of
the solution used for both negatives was one-tenth of saturation; the depth
of absorbing layer was 10 cm.

The temperatures of the strips in A, beginning with the strip nearest the
numbered scale, were 20°, 45°, 70°, 95°, 120°, 140°, 160°, 190°. This nega-
tive shows changes in bands X4275 and X5800; the former, as in the other
plates on the study of the effect of temperature, shows a marked shading
towards the red, while the remainder of this band virtually remains fixed.
The X5800 band widens rapidly toward the red, as the temperature is raised,
the total amount being about 80 a.u. All the absorption bands with the
acetate are more intense and broader than for the same concentration of any
of the other salts of neodymium studied. The acetate is not nearly so
soluble as the other salts, nor is the dissociation so great, yet we find in A,
which is the spectrogram of a one-tenth saturated solution of neodymium
acetate, greater changes than for the saturated solution of the chloride.
This is in accord with the results obtained from the effect of dilution; this, it
will be seen, was greatest with the acetate. This tends to strengthen the view
that the bands X4275 and X5800 are in some way associated with the molecules.

In B of this plate there is given the spectrogram of the same solution as
the temperature was lowered. The temperatures, beginning with the strip
nearest the spark spectrum, were 190°, 165°, 145°, 125°, 100°, 75°, 50°, 25°;
the cell and arrangement of apparatus were the same as in A . The nega-
tive shows changes the reverse of those discussed in section A. The X4275
band gradually assumes the sharply defined edges as the temperature falls,
and strip 8 of B corresponds exactly to strip 1 of A . In a word, there has
been no permanent change produced by heating the solution. This change
in the width of the absorption bands could not have been produced by any
substance dissolved from any parts of the apparatus, as there is no reason to
suppose that this should disappear as the solution was cooled. It seems,
then, that the broadening is solely a temperature phenomenon.

Plate 6 was made to show the relative effect of rise in temperature on a
solution of neodymium acetate, as compared with the same concentration
of neodymium chloride. The concentration in each case was one-tenth sat-
uration, the cell depth being 10 cm. The temperatures in A (neodymium
acetate), beginning with the strip nearest the numbered scale, were 20°,
40°, 60°, 80°, 100°, and 125°, respectively. The temperatures in B (neo-
dymium chloride), reading in the same order from the strip nearest the
spark lines, were 15°, 40°, 65°, 90°, 115°, 140°, 165°, and 190°, respectively.
A comparison of the two sections of this plate shows, first, that for the
same concentrations of the two salts the absorption bands are wider and
more pronounced with the acetate than with the chloride.

In each of these plates only the X4275 and X5800 bands show appreciable

12 ABSORPTION SPECTRA OF SOLUTIONS

change with rise in temperature. While the percentage change in the
former of these two bands is perhaps greater, this shows very poorly on
the prints from the original films. The X4275 band is very sharp at the
lower temperatures, but shades rapidly towards the red as the temperature
is raised, the X5800 band, which is most affected by temperature changes,
showing decidedly more widening with the acetate than with the chloride.
This is exactly what we should expect if this band were associated with the
undissociated molecules of the salt in question. The acetate, being a salt of
a very weak acid, is dissociated considerably less than the chloride, and con-
sequently the change is greater in the case of the acetate where there are
present a larger number of molecules.

The facts, then, are : The number of molecules in a given concentration of
neodymium acetate is greater than in the corresponding concentration of
neodymium chloride. Hydration decreases with rise in temperature. The
band X5800 is more marked in the acetate than in the chloride, and widen-
ing with rise in temperature indicates that it is in some way associated with
the hydrated molecules.

NEODYMIUM SULPHATE IN WATER AND COBALT CHLORIDE IN WATER.

(See Plate 7.)

On account of the slight solubility of neodymium sulphate in water, only
the saturated solution was studied. The depth of cell was 10 cm. and the
temperatures, beginning with the strip nearest the numbered scale, were 20°,
45°, 75°, 90°, 115°, and 140°, respectively.

It is seen that the first four strips of A (neodymium sulphate) show the
regular widening of X4275 and X5800. In strips 5 and 6 all of the bands
decrease in width. This is especially noticeable in bands X5800, X5100, and
X5225. This is no doubt due to the fact that some of the salt crystallized
out at this temperature, and the solution consequently became more dilute.
When the cell was opened it was found that nearly all of the salt had crys-
tallized out.

It is, however, obvious that the sulphate presents no exception to the
general rule that the bands widen with rise in temperature. This is cer-
tainly true up to 115°, at which temperature the crystals form rapidly, and
the effect of increase in dilution more than overcomes the counter effect of
rise in temperature.

B is the spectrogram of a solution of cobalt chloride, 1 cm. deep and 0.25
normal. The temperatures, beginning with the strip nearest the numbered
scale, were 12°, 32°, 52°, 76°, 92°, 112°, 132°, and 152°. This plate shows an
intense absorption in the violet up to X3600; also a broad, hazy band with its
center near X5100. On account of the haziness of the cobalt bands, it is
difficult to discuss them in detail. The change produced by rise in temper-
ature, however, is very slight. The cobalt salt was hydrolyzed very greatly
at the higher temperatures, and this also interfered, with the study of its
absorption.

AS AFFECTED BY TEMPERATURE. 13

PRASEODYMIUM CHLORIDE IN WATER. (See Plate 8.)

A represents the effect of rise in temperature on the absorption spectra of
a 2.56 normal solution of praseodymium chloride, the depth of cell being
1 cm. The temperatures, beginning with the strip nearest to the numbered
scale, were 20°, 50°, 80°, 100°, 120°, 140°, and 160°, respectively. The orig-
inal film shows general transmission from A3400 to X4350, with the sharply
defined absorption band extending from X4300 to X4750. There is faint
transmission near X4550. There is practically no change in either edge of
this band as the temperature of the solution is raised. There is a slight
widening of that band whose center is near X4825. The X5900 band changes
less than 25 a.u. over the entire range of temperature studied.

B is the absorption of a solution of the same salt, having a concentration
of 0.256 normal and a depth of layer of 10 cm. The temperatures, begin-
ning with the strip nearest the numbered scale, were 20°, 40°, 65°, 90°, 115°,
140°, 165°, and 190°, respectively. There are well-defined bands having
their centers near X4425, X4650, X4820, and X5900. None of these bands
shows any appreciable change with rise in temperature.

PRASEODYMIUM NITRATE IN WATER. (See Plate 9.)

The concentration of the solution used in making A was 2.6 normal; the
depth of cell, 1 cm. The temperatures, beginning with the strip nearest the
numbered scale, were 12°, 32°, 52°, 72°, 92°, 112°, 125°, and 145°, respectively.

In the ultra-violet the absorption extends to about X3500 in strip 1, but
rapidly increases as the temperature is raised, until in strip 8 there is com-
plete absorption as far as X3800.

There is a very intense double absorption band from X4350 to X4725 with
faint transmission near X4540. This transmission rapidly decreases as the
temperature is raised, and entirely disappears at a temperature slightly
above 100°. The X4650 band widens towards the red end about 25 a.u.
Band X4825 shows a total widening of about 30 a.u. over the range of tem-
perature studied. The orange band near X5900 shows a uniform total wid-
ening of about 25 a.u. From this plate it is seen that none of the praseo-
dymium bands shows very marked change with rise in temperature; at this
concentration all of them become slightly wider at the higher temperatures.

In section B of this plate is given the spectrogram of a 0.26 normal solu-
tion of the same salt, the depth of the absorbing layer being 10 cm. The
temperatures, beginning with the strip nearest the numbered scale, were
20°, 45°, 70°, 95°, 115°, 135°, and 165°, respectively. On this plate, bands
appear which have their centers near X4425, X4650, X4825, and X5900; the
ultra-violet absorption bands near X3500. None of these bands shows any
appreciable change over the range from 20° to 165°.

The plate which was used to study the effect of dilution upon this same
salt reveals the fact that only in the most concentrated solutions were the
bands affected at all, while in the dilute solutions all the bands remained
unchanged. Plate 9 shows that temperature also has a slight effect only in

14 ABSORPTION SPECTRA OF SOLUTIONS

the concentrated solutions, while in the dilute solutions the bands remain
unchanged. In a word, rise in temperature and decrease in dilution produce
the same effect upon solutions of praseodymium nitrate.

URANYL NITRATE IN WATER. (See Plate 10.)

The concentration of the solution used in making A was 0.2 normal, the
depth of layer being 1 cm. The temperatures, beginning with the strip
nearest the spark spectrum, were 20°, 40°, 60°, 80°, 100°, and 120°, respec-
tively. In every strip the exposure to the entire spectrum was made for 30
seconds, a screen cutting off all wave-lengths beyond X4500 was inserted,
and the ultra-violet end exposed an additional 8 minutes.

Since all the uranyl bands occur in the violet and ultra-violet end of the
spectrum, where general absorption is greatest, due to precipitates formed
by heating the solutions, etc., it was found very difficult to obtain satis-
factory results. So far as this plate shows, there is no decided change in any
particular band. The entire series seems to widen as the temperature is
raised, and at the same time the center of the band is slightly shifted toward
the red end of the spectrum. The general absorption, ending near X3500 in
strip 1, advances rapidly towards the red as the temperature is raised. The
broad diffuse edges of all the bands shade uniformly into each other, until
at the highest temperature they appear as one broad, hazy absorption band,
extending from X3800 to X4300. At least a part of this is due to general
absorption.

In section B is given the absorption of a 0.02 normal solution of uranyl
nitrate, the depth of absorbing layer being 10 cm. The red end of the spec-
trum, beyond X4500, was exposed 8 seconds, while the ultra-violet below
X4500 had an exposure of 3^ minutes to the same source of light. The tem-
peratures, beginning with the strip nearest the numbered scale, were 20°,
45°, 70°, 95°, 115°, 140°, and 165°, respectively. Eleven bands occur between
X3500 and X4600. As the temperature is raised, all the bands become more
diffuse and broader; the band whose center is near X4180 seems to be most
affected. The red edge of the band shades towards the red end of the spec-
trum as much as 25 a.u. The effect produced on this band by elevated tem-
peratures is more marked than in any of the other bands. There is very
broad and hazy absorption around X5100, X5600, and X6200. This increases
with rise in temperature.

It has been found very difficult to give an exact description of what takes
place in any uranyl band as the temperature is raised, since the edges of the
bands are so hazy and the general absorption so marked in the region of the
spectrum at which these bands occur. Only the general statement can be
made that all uranyl bands become more diffuse with rise in temperature,
and in the band X4165 there is a decided shading on the red edge.
URANYL SULPHATE IN WATER. (See Plate 11.)

The concentration of the solution used in making A was 0.166 normal and
the depth of cell 1 cm. The respective temperatures, beginning with the
strip nearest the numbered scale, were 20°, 45°, 70°, 90°, 115°, 135°, 155°,

AS AFFECTED BY TEMPERATURE. 15

and 185°. The part of the spectrogram above X4550 was exposed 40 seconds,
while below that wave-length the exposure was 10 minutes. The apparent
band extending entirely across the spectrogram near X4550 is the edge of the
screen used in making the long exposure on the violet end of the spectrogram
and must not be confused with an absorption band.

Absorption bands X4175 and X4325 have their centers shifted towards the
red end of the spectrum about 25 a.u. The red edges of bands X4325 and
X4550 shade rapidly towards the red. The well-marked band X4750 remains
unchanged throughout the spectrogram.

The encroachment of the general absorption in the ultra-violet towards
the red causes band X3625 to disappear above the fourth strip, while band
X3750 is scarcely visible above strip 5. All bands below X4500 become very
diffuse as the temperature is raised, and at the highest temperature are
hardly more than a single broad, hazy absorption band extending from
X4000 to X4400.

Section B is the spectrum of a 0.02 normal solution of uranyl sulphate,
the depth of absorbing layer being 10 cm. The respective temperatures,
beginning with the strip nearest the numbered scale, were 20°, 45°, 70°, 95°,
1 15°, 140°, and 165°. The exposures were 8 seconds in the visible part of the
spectrum and an additional exposure of 4 minutes to the ultra-violet. The
same changes described in A take place here, i. e., a strong general absorp-
tion in the ultra-violet beyond X3500, and increasing towards the red as the
temperature is raised. The most marked widening is in bands X4100,
X4200 and X4350; in each the center shifted slightly towards the red. Such
is also the case with the red edge of band X4600. The X4750 band remains
fixed throughout the spectrogram. The very broad, hazy bands around
X5100, X5600, and X6200 appear, and are not appreciably affected by changes
in temperature.

URANYL ACETATE IN WATER. (See Plate 12.)

In plate 12, A shows the effect of dilution, B of temperature. The con-
centrations of the solutions used in A, beginning with the strip farthest
removed from the numbered scale, were 0.25, 0.125, 0.062, 0.042, 0.0025,
and 0.0005 normal. So far as we can judge from this plate, none of the
absorption bands changes. Beer's law seems to hold to the dilution 0.0005
normal.

B shows the effect of rise in temperature on a 0.02 normal solution of
uranyl acetate. The temperatures, beginning with the strip nearest the
numbered scale, were 20°, 45°, 70°, 95°, 115°, and 140°. The exposures at
that part of the spectrum having a wave-length greater than X4500 was 8
seconds, while an additional exposure of 3 minutes was given to the ultra-
violet end. Every one of the nine bands shows a slight widening with rise in
temperature. While in strip 1 the bands are well marked, they appear
much more diffuse as the temperature is raised. The apparent change in
the band near X4475 is probably due to the screen used to cut off the visible
spectrum, while additional exposure was made to the ultra-violet region.

CHAPTER III.

EFFECT OF DILUTION ON THE ABSORPTION OF LIGHT

BY SOLUTIONS.

The question as to the effect of dilution on the power of solutions to
absorb light is an old one. This question became especially prominent at
the time the theory of electrolytic dissociation was proposed. In dilute
solutions of electrolytes there are practically only ions present, very few
molecules existing as such. All of the properties of such solutions are the
properties of the ions contained in them. Therefore, the power of these
solutions to absorb light must be due to the ions present in them. This was
the reasoning in vogue and the conclusion drawn. It was at the same time
freely recognized that molecules in solution have the power to absorb light.
This was shown by the fact that solutions of non-electrolytes, or completely
unionized substances, are often colored ; and color in solution means selective
absorption of light.

The result of the conclusion drawn from the theory of electrolytic disso-
ciation was that an enormous amount of work was done on the absorption
spectra of dilute solutions of both electrolytes and non-electrolytes. Ostwald
carried out an elaborate investigation on the relation between color and dis-
sociation, and published the work under the title "Uber die Farbe der
Ionen."1 A large number of salts were brought within the scope of this
investigation — salts of an acid having a colored anion, with colorless cations,
This is illustrated by the various permanganates, hydrogen, sodium, ammo-
nium, magnesium, zinc, cadmium, etc. Ostwald showed that these salts of
any given acid had essential^ the same spectra. In a similar manner, he
studied salts of fluorescein, eosin, iodoeosin, rosolic acid, diazoresorcinol, etc.
Ostwald then reversed the process and compared the salts of a given colored
base with colorless acids, thus studying the salts of p-rosaniline with acetic,
chloric, benzoic, hydrochloric, nitric, butyric, salicylic, lactic, etc., acids
and finding practically the same absorption spectra for all of these salts.

From the standpoint from which he undertook his investigation, Ostwald
may be said to have solved the problem of the role of ions in the absorption
of light, as far as that could be done with the prism spectroscope.

The problem that we studied was of a different nature. It had to do with
the absorption spectra of ions relative to that of the molecules from which
they were formed. Some earlier work of Jones and Anderson2 had shown
that if molecules have different action on light from ions, the difference is so
slight that there would be no hope of detecting it by ordinary means, even with
a grating spectroscope. This problem was attacked in the following manner :

1 Zeit, phya. Chem., 9, 579 (1892). 2 Cam. Inst, Wash. Pub. No. 110.

17

18 ABSORPTION SPECTRA OF SOLUTIONS.

MAKING A DILUTION SPECTROGRAM.

Before entering upon a detailed discussion of the spectrograms, it is
wise to state briefly the method used in making any given spectrogram.
Throughout all the work done on the effect of high dilution on absorption
spectra, under the conditions of Beer's law, only three exposures were made
for any given spectrogram, i. e., only three dilutions were compared. The
depths of cell in all cases were 0.5 cm., 50 cm., and 250 cm., the dilution being
increased 100 times between the first two solutions and 5 times between the
last two; or a total dilution of 500 times between the first and last solution.
Smaller depths of cell than 5 mm. were not used, on account of the large
percentage error in measuring such depths.

Much difficulty was experienced in getting sufficient light through the
longer cells to fill the grating completely; nor was this possible unless the
tube containing the solution was constantly moved backwards and forwards
so that the image of the source of light was moved along the slit of the camera.
By such a procedure the surface of the grating could be illuminated fairly
uniformily, and the exposures gave good results on the photographic plate,
as is shown by the spectrograms.

In order to insure complete illumination of the grating a uniform pro-
cedure was adopted. The longest cell, containing the most dilute solution,
was first placed in position, the light passed through, and the image of the
Nernst glower sharply focused on the slit of the camera in such a manner as
to throw as much light as possible on the grating. By holding the eye in the
position later to be occupied by the photographic plate, we could easily tell
when the grating was properly illuminated.

After everything was properly adjusted the lights were extinguished and
the plate inserted in the camera. Great care was taken not to move any
parts of the apparatus, the camera was closed, the source of light again
turned on, and the exposure made. It is clearly seen that in making any
spectrogram, using three cells differing in length so markedly, we virtually
had three different sources of light, and, consequently, the length of exposure
sufficient to give comparable results on the photographic plate had to be
determined by a long series of trials. In the case of the longest cell, expos-
ures as long as several minutes were made, while with the shortest cell only
a few seconds were necessary to give good clear spectrograms on the photo-
graphic plate.

The remaining procedure was essentially the same as that described by
Jones and Anderson1 and by other workers in this laboratory.

NEODYMIUM CHLORIDE IN WATER. (See Plate 13.)
The concentrations of the solutions used in making the negative for A,
beginning with the one whose spectrum is farthest from the spark spectrum,
were 2.05, 0.0205, and 0.00401 normal, respectively, the corresponding
depths of absorbing layer being 0.5 cm., 50 cm., and 250 cm.

iCarn. Inst, Wash. Pub. 110.

EFFECT OF DILUTION ON ABSORPTION OF LIGHT. 19

For B the concentrations used were 1.025, 0.01025, and 0.00205 normal.
The depths of layer were the same as used in A. It is seen that the dilu-
tions are just one-half those of the corresponding layers in A.

The concentrations of solutions used in making C were just half of those
in B, i. e., 0.512, 0.00512, and 0.00102 normal. In the entire plate, as in
all the dilution work, the most dilute solution is always nearest the spark
spectrum.

Since very much of the finer detail and several of the narrowest bands
are lost in reproducing and printing the films, our discussion is always based
upon the original photographic film. Lines will frequently be discussed
which do not appear on the printed plates, but which are very clear and dis-
tinct on the photographic film.

A study of A shows complete absorption in the violet up to X3350, then
slight transmission for about 50 a.u. The faint hazy band X3400 and the
well-defined band X3450-X3600 are not affected by the change in dilution.
Hazy bands appear at X3820, X4040, and X4200. Their intensities do not
seem to be affected by dilution. The beautiful sharp band X4275 is slightly
more intense in the most concentrated solution. The effect of dilution, if
any, on the bands X4325, X4440, X4600, X4690, X4750, X4820 is not measur-
able. On the original film they appear slightly broader, but not more
intense, on the third strip.

Bands which have their centers near X5100, X5200, and X5800 are decidedly
affected by dilution, the former two appearing distinctly as independent
bands in the most dilute solution, diffuse with a single broad band with the
center near X5150. There is the greatest change between the second and
third strips (in discussing any plate, strip 1 is always nearest the spark lines).
The broadening of these bands with increase in concentration, both of which
have rather hazy edges, is fairly uniform, i. e., they widen both towards the
red and violet ends of the spectrum.

The intense band which extends from X5690 to X5850 is affected very
markedly by concentration, the widening being almost entirely towards the
red end of the spectrum. The violet edge is hardly affected, while the wid-
ening towards the red is about 50 a.u. Here also the change in the width of
the band is greatest where the change in concentration of the solution is
greatest. There is a very faint band, X6225, which appears slightly more
diffuse in the most concentrated solution.

The concentrations of the solutions used in B are just one-half those of A,
and it is seen that some of the smaller bands are lost, while the broader ones
have split into two or more smaller bands. In this film, bands near X3425,
X3475, X3520, X3575, X4275, X4340, X4450, X4700, X4750,X4820,X5100,X5120,
show no change with dilution. The broad band X5700-X5825 shows a widen-
ing of about 25 a.u., being the only band which is changed by concentration.

C of this plate is the spectrogram of solutions twice as dilute as those of B.
No band on this plate shows any appreciable change produced by dilution,
except probably a slight widening of X5750.

20 ABSORPTION SPECTRA OF SOLUTIONS.

We then see, from a study of this plate, that in A bands X4270, X5100,
X5200, and X5750 narrow with dilution, the amount of change being in the
order given; that is, least in X5100 and greatest in X5700. In B there is an
appreciable change in only X5750, while in C none of the bands are affected
by dilution.

NEODYMIUM BROMIDE IN WATER. (See Plate 14.)

The concentrations of the solutions used in making negative A, beginning
with the solution whose spectrum is farthest from the scale, were 1.66, 0.0166,
and 0.0033 normal; the corresponding depths of absorption layer being 0.5
cm., 50 cm., and 250 cm., respectively.

The concentrations used in making B were half of those of A, and those of
C half those of B. The same range of cell depth was used in all three sec-
tions of this plate, viz, 0.5 cm., 50 cm., and 250 cm., respectively, beginning
with the strip farthest from the spark lines. In A , characteristic absorption
bands appear at X3400, X3525, X3800, X4275, X4450, X4700, X4750, X4800,
which are hardly affected by change in dilution, except for a slight increase
in the intensity of band X4275 in the most concentrated solution.

The three bands, X5090, X5120, and X52 10, narrow uniformly with dilution,
the greatest change being between strips 2 and 3, where the change in dilu-
tion is the greatest. With the bromide, as is seen in plate 14, the effect of
dilution is most pronounced in band X5750. The shading is almost exclu-
sively towards the red, the violet edge remaining practically unchanged.
This edge shows no change between strips 1 and 2, yet the red edge is
widened as much as 30 a.u.

In B, where the concentrations were 0.83, 0.0083, and 0.00166 normal,
respectively, the depths of absorbing layer were the same as used in A.
There is no measurable change in any of the absorption bands except the
band whose center is near X4800. This band shows the characteristic nar-
rowing with dilution, as the dilution is increased. The total change is not
greater than 20 a.u. Band X5200 is slightly more intense in the third strip.

When we reach the dilution used in C, which is four times that of A, any
change due to dilution has disappeared except a narrowing of probably 10
a.u. between the third and second strips.

Taking plate 14 as a whole we see, first, the narrowing due to increased
dilution is most marked in A, less in B, and least in C. This is seen to be
the same order as their respective concentrations. Considering an indi-
vidual section, we find the most pronounced narrowing where the change in
dilution is greatest — that is, between strips 2 and 3.

NEODYMIUM NITRATE IN WATER. (See Plate 15.)

The concentrations of neodymium nitrate used in making negative A of
this plate, beginning with the strip farthest from the numbered scale, were
2.15, 0.0215, and 0.00430 normal, the corresponding depths of absorbing
layer being 0.5 cm., 50 cm., and 250 cm., respectively.

EFFECT OF DILUTION ON ABSORPTION OF LIGHT. 21

111 discussing the absorption bands of this, as well as other plates through-
out this paper, we do not attempt to give the exact position of the band in
question, as has previously been done by many workers; but we simply indi-
cate the position of the band by selecting a wave-length near its center.
For instance, in speaking of band X5800, we mean that broad band extending
from X5700 to X5850. This is not confusing and saves space and time in
the description of any plate.

This is probably the best plate we have illustrating the effect caused by
dilution. Bands which are hardly affected over the range of dilution given
in A are located at X3525, X3820, X4440, X4620, X4750, X4830. In strip 3
the well-defined band X4275 is more diffuse, though probably not so intense.
This is in keeping with the behavior of this same band as shown by other
salts of neodymium, though probably a little more marked. There is faint
transmission at X5100 for about 10 a.u. In strip 3, representing the most
concentrated solution, the bands X5090 and X5125 have so broadened that
they coalesce. The X5220 band widens uniformly towards the red and
violet as the solution becomes more concentrated.

Band X5800, which is most affected by dilution, shows a total change of
probably as much as 70 a.u., the shading being largely towards the red.
In strips 1 and 2 the violet edge is hardly changed, while in strip 3 it is prob-
ably shifted 20 a.u. On the original film these three strips show abso-
lutely the same development, hence are directly comparable.

Section B represents the absorption of neodymium nitrate. Beginning
with the strip farthest from the numbered scale, the concentrations are
1.075, 0.01075, and 0.00215 normal, the corresponding depths of cell being
the same as in A. In this section only a few bands need be discussed.
Bands X5090 and X5125, which appear as distinct bands in strips 1 and 2,
have slightly broadened so as to form a single hazy band whose center is
near X5120. The X5750 band narrows as much as 40 a.u., almost the entire
change being between strips 2 and 3.

In C, the X5750 band alone is noticeably changed, narrowing about 20
a.u. from strip 3 to strip 2, but is not changed in the last dilution, i. e., from
strips 2 to 1.

NEODYMIUM SULPHATE IN WATER. (See Plate 16.)

A gives the absorption spectra of a solution of neodymium acetate. The
concentrations of the solutions used, beginning with the strip farthest
removed from the numbered scale, were 0.5, 0.01, and 0.002 normal. The
corresponding depths of cell were 1, 50, and 250 cm.

This additional plate of neodymium acetate was made to study the effect
of exposure on the apparent widening of the bands with concentration.
Strip 2 was more exposed than strip 1, and strip 3 had a longer exposure
than strip 2. Nevertheless the X5800 band has widened as much as 50 a.u.
between the first and third exposures. In view of the unequal exposures of
these strips, it is not thought advisable to discuss the other bands. These

22 ABSORPTION SPECTRA OF SOLUTIONS.

results show that difference in exposure can not account for the changes in
the widths of the bands in question.

Spectrograms B and C of this plate give the absorption of solutions of
neodymium sulphate. On account of the slight solubility of this salt, we
observe only slight changes in any of its absorption bands.

The concentrations in B were 0.1, 0.004, and 0.0008 normal; the corre-
sponding depths of cell being 2, 50, and 250 cm. The concentrations in C
were 0.1, 0.001, and 0.0002 normal, and the corresponding depths of cell
were 0.5, 50, and 250 cm.

In both B and C, those bands having their center near X3500 appear well
defined and remain unchanged both in position and intensity as the dilution is
changed. The band A5750 widens with concentration in A as much as 25 a.u. ;
it remains practically unchanged in B, where the solutions are more dilute.

The plate brings out the fact already mentioned, that only the more con-
centrated solutions show marked change, either with change in temperature
or with change in dilution.

NEODYMIUM ACETATE IN WATER. (See Plate 17.)

The concentrations of solutions used in making A of this plate were sat-
urated, one-hundredth saturated, and five-hundredth saturated, the corre-
sponding depths of cell being 0.5 cm., 50 cm., and 250 cm., respectively.
The most dilute solution is nearest the numbered scale.

This plate was made with very long exposures, to see if the apparent
widening of the bands could be due to the difference in the amounts of light
falling upon the photographic plate. In such a procedure the most concen-
trated solution was given the longest exposure and yet had the broader
bands. It is possible to narrow any given absorption band by lengthening
the time of exposure, but this can not account for so large a difference as is
shown by strip 3 of section A. Even in this section it is seen that the third
strip has wider bands than either of the other two strips of this section, not-
withstanding the fact that the actual exposure of the strip is greater. Thus
we see that the difference of exposure can not account for the changes in
the width of bands such as we have noted.

In section A the violet group of bands in the region X3500 came out beau-
tifully. Such is only the case when quite a long exposure is made. Indeed,
in order to show these lines clearly, the exposure must be long enough to
destroy those fine, sharp lines in the region of X3800 to X4600. Hence, in
this plate the latter group of lines do not appear distinctly, though traces of
them can be seen on the original film.

The hazy bands X3300 and X3400 appear on this plate and remain un-
changed by dilution.

The three bands, X3460, X3500 and X3540, remain perfectly constant
throughout the section. Band X5120, which appears broad and diffuse,
shows no change. Band X5210 narrows about 10 a.u. from the third to the
second strip, and remains unchanged with the next dilution. The broad

EFFECT OF DILUTION ON ABSORPTION OF LIGHT. 23

band X5750 narrows about 40 a.u. from the third to the second strips, and
about 15 a.u. with the next dilution.

The concentrations used in B were again just half those in A, the most
concentrated solution being one-half saturated, with succeeding dilutions of
100 and 5 times, respectively; the depths of cell, beginning with the strip
farthest removed from the scale, were 0.5 cm., 50 cm. and 250 cm.

Again, only bands X5220 and X5750 are changed, but with the acetate the
change extends farther with the more dilute solutions. In a word, the nar-
rowing of the bands with dilution is more marked in B and C than in the case
of the chloride, bromide, and nitrate. The group of bands near X3500 is not
altered with dilution.

The concentrations used in C are again half of those in B, and the corre-
sponding depths of cell the same as used throughout this plate. Their respec-
tive sequence is the same. In this spectrogram only band X5750 changes,
and, indeed, this is the only salt of neodymium with which a change with
dilution has been noted in so dilute a solution. This band narrows about 20 a.u.

Neodymium acetate, being the salt of a weak acid, is of course less dis-
sociated at any given concentration than a salt of a strong acid. This salt
approaches complete dissociation much more slowly than the others studied.
Spectrograms A, B, and C show also that the changes caused by dilution are
more marked and extend into more dilute solutions than with the chloride,
bromide, or nitrate. In a word, the changes in the absorption bands due to
dilution seem to follow the change in dissociation; that is, they are a direct
function of the number of molecules present.

PRASEODYMIUM CHLORIDE IN WATER. (See Plate 18.)

The concentrations of solutions used in making A, beginning with the
strip farthest removed from the numbered scale, were 2.56, 0.0256, and
0.00512 normal, the corresponding depths of absorbing layer being 0.5 cm.,
50 cm., and 250 cm., respectively.

The absorption is complete in the ultra-violet up to about X3100. The
bands of praseodymium are for the most part broad and have well-defined,
sharp edges. The violet edge of band X4450 is very sharp and unchanged
by dilution, while the hazy red edge is hardly affected. Band X4675 narrows
towards the violet about 20 a.u., while band X4830 is entirely unchanged.
The broad band X5900, with slightly hazy edges, shows a total narrowing of
about 25 a.u.

The concentrations of B and C, beginning with the strips farthest from
the numbered scale, are 1.28, 0.0128, and 0.00256 normal and 0.64, 0.0064,
and 0.00128 normal, respectively. The corresponding depths of absorbing
layer were the same as in A. None of the bands is affected by dilution,
and we may say that Beer's law holds very well for praseodymium chloride,
except for bands X4675 and X5900 in the most concentrated solutions. Even
here the change is very slight and not to be compared with corresponding
changes with dilution in salts of neodymium.

24 ABSORPTION SPECTRA OF SOLUTIONS.

PRASEODYMIUM NITRATE IN WATER, (See Plate 19.)

The concentrations of the solutions used in A, beginning with the strip
farthest removed from the numbered scale, were 2.6, 0.026, and 0.0052 nor-
mal; the corresponding depths of cell being 0.5, 50, and 250 cm.

The concentrations in B were just half of those in A, and those in C were
half of those in B. None of the absorption bands shows airy change with
dilution in either B or C. In A there is a slight change in the X4450 and
X4650 bands. Each of these bands widens about 20 a.u. with concentration
over the range of concentration studied.

With praseodymium nitrate, as with the chloride previously discussed,
there is only very slight change in the absorption with change in dilution.

URANYL CHLORIDE IN WATER. (See Plate 20.)

The concentrations of solutions used, beginning with the strip farthest
removed from the numbered scale, were 1.363, 0.682, 0.341, 0.227, 0.01363,
and 0.00272 normal; the corresponding depths of cell being 0.5, 1, 2, 3, 50,
and 250 cm. In making this spectrogram, no additional exposure was made
in the ultra-violet. The last four strips, one being nearest the numbered
scale, were each exposed 30 seconds to the Nernst glower. The first two
strips, on account of the length of cell used, had to be exposed a much longer
time. In this, as in all other cases, the length of exposure was governed
solely by the time required to give a clear print on the plate.

There is complete absorption of all the light having wave-lengths shorter
than X4500, well-defined bands with rather hazy edges appearing near X4700
and X4900. There is also rather diffuse absorption near X5500 and X6100.
The two last-named bands are too ill-defined for detailed discussion.

The X4700 band shows marked widening with increase in concentration,
the change being greater towards the red end of the spectrum. The entire
band is about 50 a.u. The X4900 band shades off rapidly towards the red
end of the spectrum, but not so much as the band X4700. The greatest
change is between strips 5 and 6, i. c, where the change in dilution is greatest.

The concentrations in B were just half of those in A. Starting with the
strip away from the scale, they were 0.685, 0.340, 0.170, 0.1135, 0.00685, and
0.00136 normal. The depths of the cell were the same as in A, 0.5, 1, 2, 3,
50, and 250 cm.

The changes produced by dilution, as shown in B, are much less marked
than in A . Indeed, this would be expected, since the concentrations of the
solutions were less. There is, however, a gradual widening of both X4600
and X4700 bands as the solution becomes more concentrated. The greatest
change is in strips 5 and 6.

URANYL BROMIDE IN WATER. (See Plate 21.)

The concentrations of the solutions used in making A, beginning with the
strip farthest removed from the numbered scale, were 1.365, 0.682, 0.341,
0.227, 0.01365, and 0.0027 normal, the corresponding depths of layer being
0.5, 1, 2, 3, 50, and 250 cm.

EFFECT OF DILUTION ON ABSORPTION OF LIGHT. 25

This spectrogram shows complete absorption in the violet to about X4500,
with a well-defined band near X4700. The latter widens uniformly with
increase in concentration. It is scarcely visible in strip 1, but becomes about
75 a.u. wide in the top strip. The most concentrated solution (strip 6)
shows decidedly more absorption towards the red. The change in both of
these bands is decidedly the most pronounced between strips 5 and 6, i. e.,
where the percentage change in dilution is greatest.

B contains the absorption spectra of a series of solutions whose respective
concentrations are just half of those in A, the corresponding depths of
absorbing layer being the same as in A.

There is faint transmission near X3800, with complete absorption of all of
the wave-lengths from this region to X4400. None of the bands shows any
change with dilution. In a word, Beer's law seems to hold perfectly for
these dilutions.

URANYL NITRATE IN WATER. (See Plate 22.)

The concentrations of A, beginning with the strip farthest removed from
the numbered scale, were 1.55, 0.775, 0.387, 0.269, 0.0155, and 0.0031 nor-
mal, the corresponding depths of cell being 0.5, 1, 2, 3, 50, and 250 cm.

In B, the concentrations were just half of those in A, the same depths of
cell being employed. The negative of A shows the bands with special
clearness.

In strip 1 there is complete absorption of the violet to X4500. This
gradually recedes towards the red, with increase in concentration amount-
ing to as much as 100 a.u. The X4700 band widens about 20 a.u. There is
a sharp band, X4878, which widens slightly with increase in concentration.

B shows faint transmission around X3750, with broad, intense absorption
to X4350. Absorption bands, which are unchanged by change in dilution,
appear at X4550, X4700, and X4850. The only change in the bands on this
plate is a slight encroachment on the red of the broad violet absorption in
strip 6.

The results recorded on this plate are in complete accord with those of
plates 20 and 21, which are the corresponding absorptions of uranyl chloride
and uranyl bromide. A , in all three of these plates, represents the most con-
centrated solutions, while B represents half the concentrations in A . Most
of the bands in A show well-marked widening with increase in the concen-
tration, while in B the change is scarcely detectable.

DESCRIPTION OF PLATES.

Plate

1. A. Neodymium Chloride in Aqueous Solution. Concentration, saturated. Depth

of layer, 1 cm. Respective temperatures, 20°, 45°, 70°, 95°, 115°, 140°, and
165°, with lowest temperatures nearest spark lines. Exposures made on
rising temperature.
B. The same solution used in A, with exposures made as cell cooled. Depth of layer
and concentration the same as in A. Temperatures, 165°, 140°, 115°, 95°,
70°, respectively. Highest temperatures nearest spark lines.

2. A. Neodymium Bromide in Aqueous Solution. Concentration, 1.66 normal. Depth

of cell, 1 cm. Respective temperatures, 20°, 45°, 70°, 95°, 120°, 140°, 175°,
190°. Lowest temperature nearest spark spectra.
B. Neodymium Bromide in Aqueous Solution. Concentration, 0.166 normal.
Depth of cell, 10 cm. Respective temperatures, 20°, 45°, 70°, 95°, 115°,
135°, 155°, and 190°. Highest temperature nearest spark spectra.

3. A. Neodymium Nitrate in Water. Concentration saturated; depth of cell, 1 cm.

The temperatures, beginning with the strip nearest the numbered scale,
were 15°, 40°, 65°, 115°, 140°, and 165°, respectively.
B. Neodymium Nitrate in Water. Concentration, one-tenth saturation, depth of
cell, 10 cm.; temperatures, 20°, 45°, 70°, 95°, 120°, and 145°, respectively.
Lowest temperature nearest the numbered scale.

4. A. Neodymium Nitrate in Aqueous Solution. Concentration, 0.1 saturated. Depth

of layer, 10 cm. Temperatures, 20°, 45°. 70°, 95°, 115°, 140°, 165°, 190°.
Exposures made as temperature was raised. Lowest temperature nearest
spark lines.
B. The same solution of neodymium nitrate as used in A . Concentration and depth
of layer identical with A. Temperatures, 190°, 165°, 140°, 115°, 95°, 70°,
45°, and 20°. Exposures made on falling temperatures. Highest tempera-
tures nearest spark lines.

5. A. Neodymium Acetate in Water. Concentration, 0.1 saturated. Depth of cell,

10 cm. Trace of acetic acid added to prevent precipitation. Tempera-
tures, 20°, 45°, 70°, 95°, 120°, 140°, 160°, and 190°. Exposures made on
rising temperature. Lowest temperature nearest spark line.
B. Solution, depth of cell and concentration the same as A. Temperatures, 190°,
165°, 145°, 125°, 100°, 75°, 50°, and 25°. Exposures made on falling
temperatures.

6. A. Neodymium Acetate in Water. Concentration, one-tenth saturation; depth of

absorbing layer, 10 cm. The temperatures, beginning with strip nearest
the numbered scale, were 20°, 40°, 60°, 80°, 100°, and 125°, respectively.
B. Neodymium Chloride in Water. Concentration one-tenth saturation; depth of
cell, 10 cm. The temperatures, beginning with the strip nearest the
numbered scale, were 15°, 40°, 65°, 90°, 115°, 140°, 165°, and 190°, respec-
tively.

7. A. Neodymium Sulphate in Water. Concentration was saturation, cell depth, 10 cm.

Temperatures, beginning with the strip nearest the numbered scale, were

20°, 45°, 75°, 90°, 115°, and 140°.
B. Cobalt Chloride in Water. Concentration, 0.25 normal; depth of cell 1 cm.

Temperatures, 12°, 32°, 52°, 76°, 92°, 112°, 132°, and 152°, respectively.

The lowest temperature was nearest the numbered scale.
S. A. Praseodymium Chloride in Water. Concentration, 2.56 normal; depth of

absorbing layer 1 cm. Temperatures, beginning with strip nearest the

numbered scale, were 20°, 50°, 80°, 100°, 120°, 140°, and 160°, respectively.
B. Praseodymium chloride in water. Concentration, 0.256 normal; depth of cell,

10 cm. Beginning with the strip nearest numbered scale, the temperatures

were 20°, 40°, 65°, 90°, 115°, 140°, 165°, and 190°, respectively.
9. A. Praseodymium Nitrate in Water. Concentration, 2.6 normal. Cell depth, 1 cm.

Temperatures, 12°, 32°, 52°, 72°, 92°, 112°, 125°, and 145°. Lowest

temperature nearest spark lines.
B. Praseodymium Nitrate in Water. Concentration, 0.26 normal. Cell depth,

10 cm. Temperatures, 20°, 45°, 70°, 95°, 115°, 135°, and 165°. Lowest

temperature nearest spark lines,
26

DESCRIPTION OF PLATES. 27

10. A. Uranyl Nitrate in Water. Concentration, 0.2 normal. Cell depth, 1 cm.

Temperatures, starting with strip nearest spark lines, were 20°, 40°, 60°,
80°, 100°, and 120°, respectively.
B. Uranyl Nitrate in Water. Concentration, 0.02 normal. Depth of cell, 10 cm.
Temperatures, starting with exposure nearest spark lines, were 20°, 45°,
70°, 95°, 115°, 140°, and 165°, respectively.

11. A. Uranyl Sulphate in Water. Concentration, 0.166 normal. Depth of cell, 1 cm.

' Temperatures, beginning nearest spark lines, 20°, 45°, 70°, 90°, 115°, 135°,
155°, and 185°, respectively.
B. Uranyl Sulphate in Water. Concentration, 0.02 normal. Cell depth, 10 cm.
Respective temperatures, beginning nearest spark lines, were 20°, 45°. 70°,
95°, 115°, 140°, and 165°.

12. A. Uranyl Acetate in Water. The concentrations, beginning with the strip most

removed from the numbered scale, were 0.25, 0.125, 0.062, 0.042, 0.0025,
0.0005 normal, respectively; the corresponding depths of absorbing layer
were 0.5, 1, 2, 3, 50, and 250 cm.
B. Uranyl Acetate in Water. Concentration, 0.02 normal. Depth of cell, 10 cm.
The temperatures, beginning with the strip nearest the numbered scale,
were 20°, 45°, 70°, 95°, 115°, and 140°, respectively.

13. A. Neodymium Chloride in Water. Concentrations, 2.05, 0.0205, and 0.00401 nor-

mal. Respective depths of cell, 0.5, 50, and 250 cm. Most dilute solution
nearest spark lines.

B. Neodymium Chloride in Water. Concentrations, 1.025, 0.01025, and 0.00205

normal. Depths of cell, starting with strip farthest from spark lines,
were 0.5, 50 and 250 cm., respectively.

C. Neodymium Chloride in Water. Concentrations, 0.512, 0.00512, and 0.00102 nor-

mal. Depths of cell, beginning with strip farthest removed from spark
lines, were 0.5, 50, and 250 cm., respectively.

14. A. Neodymium Bromide in Water. Concentrations, 1.66, 0.0166, and 0.0033 normal.

Corresponding depths of cell, 0.5, 50, and 250 cm., respectively. Most
dilute solution nearest spark lines.

B. Neodymium Bromide in Water. Concentrations, 0.83, 0.0083, and 0.00166

normal. Corresponding depths of cell, 0.5, 50, and 250 cm., respectively.
Most dilute solution nearest spark lines.

C. Neodymium Bromide in Water. Concentrations, 0.415, 0.00415, and 0.000S03

normal. Corresponding depths of cell, 0.5, 50, and 250 cm., respectively.

15. A. Neodymium Nitrate in Water. Concentrations, 2.15, 0.0215, and 0.00430 normal,

respectively. Corresponding depths of cell were 0.5, 50, and 250 cm.,
respectively. Most dilute solutions nearest spark lines.

B. Neodymium Nitrate in Water. Concentrations, 1.075, 0.01075, and 0.00215

normal. Corresponding cell depths were 0.5, 50, and 250 cm., respectively.

C. Neodymium Nitrate in Water. Concentrations, 0.537, 0.00537, and 0.00107.

Corresponding depths of cell were 0.5, 50, and 250 cm., respectively. Most
dilute solutions nearest spark lines.

16. A. Neodymium Acetate in Water. Concentrations, 0.5, 0.01, and 0.002 normal;

the corresponding depths of absorbing layers being 0.1, 50, and 250 cm.,
respectively.

B. Neodymium Sulphate in WTater. Concentrations 0.1, 0.004, and 0.0008 normal,

the corresponding depths of cell being 2, 50, and 250 cm., respectively.

C. Neodymium Sulphate in Water. Concentrations, 0.1, 0.001, and 0.0002 normal.

Depths of cell, 0.5, 50, and 250 cm., respectively. In each case, the most
dilute solution is nearest the numbered scale.

17. A. Neodymium Acetate in Water. Concentrations, saturated, 0.01s and 0.002s,

where s represents a saturated solution of the salt in water at 25°. Corre-
sponding depths cf cell were 0.5, 50, and 250 cm., respectively.

B. Neodymium Acetate in Water. Concentrations, 0.50s, 0.005s, and 0.001s (s being

a saturated solution as above). Corresponding depths of cell were 0.5,
50, and 250 cm.

C. Neodymium Acetate in Water. Concentrations, 0.25s, 0.0025s, and 0.0005s.

Corresponding depths of cell were 0.5, 50, and 250 cm., respectively.

18. A. Praseodymium Chloride in Water. Concentrations, 2.56, 0.0256, and 0.00512

normal, respectively. Corresponding depths of cell were 0.5, 50, and 250 cm.

B. Praseodymium Chloride in Water. Concentrations, 1.28, 0.0128, and 0.00256 nor-

mal. Corresponding depths of cell were 0.5, 50, and 250 cm., respectively.

C. Praseodymium Chloride in Water. Concentrations, 0.64, 0.0064, and 0.00128 nor-

mal. Corresponding depths of cell were 0.5, 50, and 250 cm., respectively.

28 ABSORPTION SPECTRA OF SOLUTIONS.

19. A . Praseodymium Nitrate in Water. Concentrations, 2.6, 0.026, and 0.0052 normal,

resp3Ctively. Corresponding depths of cell, 0.5, 50, and 250 cm., respec-
tively; most dilute solution nearest numbered scale.

B. Praseodymium Nitrate in Water. Concentration, 1.3, 0.013, and 0.0026 normal;

cell depths, 0.5, 50, and 250 cm., respectively.

C. Praseodymium Nitrate in Water. Concentration, 0.65, 0.0085, and 0.0013 nor-

mal; cell depths, 0.5, 50, and 250 cm. In each case the most dilute solu-
tion is nearest the numbered scale.

20. A. Uranyl Chloride in Water. Concentrations, 1.363, 0.682, 0.341, 0.227, 0.01363,

and 0.00272 normal Depths of cell, 0.5, 1, 2, 3, 50, and 250 cm., respec-
tively.
B. Uranyl Chloride in Water. Concentrations, 0.685, 0.340, 0.170, 0.1135, 0.00685
and 0.00136 normal, corresponding depths of absorbing layers being 0.5,
1, 2, 3. 50, and 250 cm. The most dilute solution in each case is nearest
the numbered scale.

21. A. Uranyl Bromide in Water. Concentrations, 1.365, 0.682, 0.341, 0.227, 0.01365,

and 0.00273 normal. Corresponding depths of cell, 0.5, 1, 2, 3, 50, and
250 cm. Most dilute solution nearest scale.
B. Uranyl Bromide in Water. Concentrations, 0.682, 0.341, 0.171, 0.113, 0.00682,
and 0.00136 normal. Corresponding depths of cell, 0.5, 1, 2, 3, 50, and
250 cm.

22. A. Uranyl Nitrate in Water. Concentrations, 1.55, 0.775, 0.387, 0.269, 0.0155, and

0.0031 normal. Corresponding depths of cell, 0.5, 1, 2, 3, 50, and 250 cm.,
respectively.
B. Uranyl Nitrate in Water. Concentrations, 0.775, 0.387, 0.193, 0.134, 0.00775,
and 0.0015 normal, respectively. Corresponding depths of cell, 0.5, 1, 2,
3, 50, and 250 cm., respectively.

PLATF 1

to

■2

•-0

o

in

o
in

CO

to

T

o

CM

00

10

C

o
in

•i
■r

QQ

PLATE 3

o

00

10

10

<o

1-

10

10

O

CO

ic
ID

IT)

IN

o

in

00

1

z\

10
T

*

^n i

TJ-

o

00

to

■

t
O

CM

CO
CM

I

10

cq

rE 4

QQ

PLATE

*

c

a

'T

'I
-Hi

03

PLATE 6

o

CO

w

QQ

PLATE V

03

PLATE 8

*

CQ

PLATE 9

o
vo

to

in

o

OB

*

O

<o

CO

CO

o

CO

QQ

PLATE 10

o

to

rr,

i_

j§-

=~

=_

=

rg

S_

m

§-

~r~

g

=_

«

i-

00

f-

OJ

i-

»

03

PLATE 1 1

CQ

o

•V

ID

o

ID

3!

S

03

PLATE 13

5?

O)

_ <c

oi

1M

»-*

Q

h*

I 00

[ <s

*

<M
\0

0
IB

00

in

U>

10

T

in

in
a

lD
CD

t

M

■<»•

O
00

m

10

tn

•*
ro

CM
CO

a
«o

00

u

■

QQ

U

PLATE 14

' |

s

s

CQ

%

PLATE 1 5

00

1 r'^^H

*

to

CM
10

1 °

is

00

in

Cfi

10

■"t
in

cm

o

OS

IS

■3-

<*

J CM

4 °

11 00

I J 10

CM

00
CM

CM

s

00
10

' M>

a

<o

O

vo

00

in

If)

1 f
in

CM

10

o
m

09

3

^
T

CM

Mr

0

00

m

CM

o

CO

oo

IM

CM

Q
If

1 00
10

1 *

9

3

1 °

VO

09

IB

1 v

10

m

CM

10

o
in

09

!|

•*

] CM

1 °

OS

10

CM

g

CO

OJ
CM

u

CM

03

00
10

16

O
<0

00

in

K)

-r

IT,

o
in

CD

is

t

IN

QQ

s

*s

03

u

PLAT!

0

(0

o
no

QO

QQ

PLATE 19

o

00

■o-

10

2

o

00

in

[

in

in

o
4 in

03

IS

<*
*

O

00

m

(0

gj^ ^^^^J

O

"0

00

*■"

0Q

U

PLAT

'XI

<x>

o
IS

ID

in

in

o
10

*

O

■<*■

oa

1 a

1 ■■•

00

(0

o

00
ID

IB

ID

10

o

in

03

B

■*
t

Si

1 o

1

00

10

■ i—

CM
CO

a

■t

QQ

QQ

CHAPTER IV.

ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS OF CERTAIN
SALTS OF NEODYMIUM AS STUDIED BY MEANS OF
THE RADIOMICROMETER.

The radiomicrometer is simply a thermoelement attached to a loop of
thin copper wire suspended in a magnetic field. One of the greatest diffi-
culties in constructing this element is to obtain copper wire free from all
magnetic metals. If perfectly pure copper wire could be found, an instru-
ment could be constructed of almost any desired sensibility.

A very good specimen of small copper wire was furnished us by Leeds and
Northrup, of Philadelphia. This wire was dipped in dilute nitric acid and
the exterior dissolved away until the wire was of proper size. The removal of
the outside coating of the wire removed practically all of the magnetic mate-
rial from it, this material probably being iron from the dies through which
the wire was drawn.

It was not a simple matter to construct a satisfactory thermo-electric
junction. The alloys used in making this junction were 90 parts bismuth
and 10 parts tin, and 97 parts bismuth and 3 parts antimony. The method
of making the thermo-electric junction and of soldering it on to the ends of
the loop of copper wire we owe to Professor A. H. Pfund.1 Fine strips of
the alloys were obtained in the following manner:

A few grams of the alloy in question were fused in a vessel free from all
magnetic material, and then thrown tangentially upon a clean and smooth
glass plate. In this way strips of the metal were obtained of almost any
desired thickness. Some were too thin to handle, those used being about
1 mm. wide, 0.01 mm. thick, and about 5 mm. in length.

The thermo-element was made by soldering an end of a strip of one of the
above-named alloys to an end of a strip of the other, the whole having the
form of a letter V. The two free ends of the V were soldered to the two ends
of the loop of copper wire. The soldered surfaces were blackened to absorb
the energy more completely. At the end of the loop of copper wire opposite
the thermo-element a light glass rod is fastened. This carries the mirror
and is suspended from above by a quartz fiber. The mirror employed was
about 4 sq. mm. This entire system, consisting of thermo-element, loop of
copper wire, and mirror, weighed about 20 mg. It was suspended by means
of a quartz fiber so that the loop hung between the poles of a strong magnet.

1 Plws. Rev., 34, 22S (1012). Phys. Zeit., IS, S70 (1912).

29

30 ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS OF CERTAIN SALTS

This entire system was suspended in the interior of a glass tube, the tube
being closed by a ground-glass stopper, and provided with suitable windows
for exposing the junction and observing the mirror. The upright tube was
provided with a side tube for evacuation, and by a method devised by Pro-
fessor Pfund a very high vacuum could be obtained and maintained for any
desired length of time. By suitably turning the ground-glass stopper in the
top of the glass tube, the loop of copper wire, mirror, and, indeed, the whole
system, could be made to occupy any position relative to the magnets, even
after the entire system had been evacuated. The whole apparatus was
supported upon a leveling stand and packed in cotton to protect it from
external radiation, the thermal junction alone being exposed to the radia-
tion in question.

The sensibility of the instrument used can be seen from the following
data : It had a full period of 8 seconds, and with a candle at a distance of a
meter gave a deflection of 15 cm. when the light was allowed to fall on the
junction after passing through a glass window.

When the apparatus was pumped out and the radiomicrometer thus sus-
pended in a vacuum, the deflection for a candle at a distance of a meter was
50 cm. Since glass absorbs just about half the energy emitted by a candle,
our radiomicrometer, when provided with a rock-salt window and exposed
to a candle at a distance of a meter, would give a deflection of about 100 cm.

How our instrument compared with the radiomicrometers constructed
and used by other investigators can be seen from the following table, taken
in part from the paper by Coblentz :l

Table 1.

Investigator.

Whole
period in
seconds.

Deflection in cm. per sq. mm.;
candle 1 m. distance.

Boys, Phil. Trans. (A) 180, 159 (1889) . . .
Paschen, Wied. Ann., 48, 275 (1893)
Lewis, Astrophys. Journ., 2, 1 (1895) . . .
Coblentz, Bull. Bur. Stand., (1 Sept. 1907)
Coblentz, Bull. Bur. Stand., (1 Sept. 1907)
Jones and Guy

10
40
20
40
25
8

0.9
3.0
1.3

3.6

6.0

8.0
25.0 (vacuo)
50.0 (vacuo) rock-salt window.

Jones and Guy

7
7

Jones and Guy

The magnetic control due to small amounts of magnetic impurities in the
copper wire was, of course, greater the more sensitive the instrument. For
this reason the radiomicrometer was not used in a vacuum. The length of
the quartz fiber was so chosen that a candle at a distance of a meter gave a
deflection of 16 cm. The half period was 4 seconds. This sensibility was
found to be quite sufficient for work in the red and infra-red, and even for
wave-lengths as short as 4,500 a.u. The measurement could be carried
out quickly and the magnetic disturbance was practically negligible.

1 BulUBur. Standards 4, No. 3.

AS STUDIED BY MEANS OF THE RADIOMICROMETER. 31

When the thermo-j unction was exposed to the radiation and the source of
energy removed, the instrument returned to its original zero position to
within 0.5 mm. In most cases several readings were made for a given
amount of radiation, and these usually agreed to within 1 per cent. The
source of energy was a Nernst glower attached to a storage battery, the
amperage being 1.2 and the voltage 110. This was found to be very con-
stant, successive readings in the same position of the spectra agreeing well
with one another.

The vessels used for holding the solutions were made of brass and gold
plated. They were about 4 cm. in diameter and of the desired thickness.
The ends were made of the best optical glass. Vessels of the same thickness
gave practically the same deflection both when empty and when filled with
water.

METHOD OF PROCEDURE.

The light from a Nernst glower was rendered parallel by a lens, then
passed through the vessel containing the solution, and allowed to fall on the
slit of a Hilger spectroscope. The solution was first inserted into the path of
the light, and then the pure solvent, this being done without disturbing the
adjustment. By means of a movable framework, first the vessel containing
the solution and then that containing the solvent were interposed in the path
of the beam. A metal screen interposed between the Nernst glower and the
vessel containing the solution allowed the light to pass through the solution
only when an observation was being made. By this means the thermo-
electric junction was exposed to the radiation only long enough to read the
deflection of the mirror.

The light, after passing through the solution and the slit of the spectro-
scope, fell upon the prism of the Hilger spectroscope. A second slit was
inserted in the spectroscope instead of the eye-piece. The light passed from
the prism through this second slit, and was then focused on the thermal
junction of the radiomicrometer.

The Hilger spectroscope contained a milled head, graduated so that the
wave-lengths could be read off directly. By suitably turning this head
any desired wave-length could be thrown upon the junction of the radio-
micrometer.

The width of the slit used in the visible part of the spectrum was 0.4 mm.
In the infra-red, where there is far more energy, the slit width was cut down
to 0.22 mm. A series of readings was carried out as follows: The vessel
containing the solvent was first placed in the path of the beam of light, the
screen removed, and the deflection of the mirror noted. Then the vessel
containing the solution was put in the same place that was formerly occupied
by the vessel containing the solvent, the screen removed, and the deflection
again noted. The prism was then turned slightly by means of the gradu-
ated and calibrated head, and a new wave-length allowed to fall on the junc-

32 ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS OF CERTAIN SALTS

tion. By repeating this procedure any wave-length could be allowed to fall
on the junction. If we represent by I the deflection with the solution in the
path of the beam of light, and by Jo the deflection with the solvent in the path
of the light beam, the percentage of light which passed through the solution
would be represented by I/Iq. In tables 2 to 5 we have the ratio of I/h.

Table 2. — Observed Transmission of Neodymium Chloride Solutions.

X

N.=3.43.

N.=3.43.

N. =0.857.

N. =0.427.

X

N.=3.43.

N.=3.43.

N. =0.857.

N.=0.427.

D. = 2.5mm.

D. = 5mm.

D. = 10mm.

D.=20mm.

D.=2.5mm.

D.=5mm.

D. = 10mm.

D.=20mm.

486

72

67

80

70

676

71

62

63

57

492

83

72

82

74

678

78

67

61

495

67

61

61

66

681

85

75

74

76

499

38

25

40

41

685

89

91

97

89

501

26

13

29

30

691

92

90

99

90

503

15

8

13

16

699

88

74

97

89

505

9

2

7

12

706

65

37

91

84

506

10

1

11

11

710

49

20

80

75

509

13

2

13

14

714

33

8

47

57

513

13

4

10

17

719

18

1

29

27

515

10

6

2

8

, 722

7

1

8

5

516

6

2

4

5

724

4

0

1

0

518

10

4

13

14

729

6

3

0

0

520

23

11

31

29

733

14

14

0

0

522

36

18

45

42

737

26

26

5

3

525

52

23

62

56

741

38

41

17

16

530

83

78

92

84

745

50

51

42

36

535

96

95

94

89

750

57

51

64

59

544

92

95

91

87

755

59

41

79

75

550

97

98

80

79

760

54

32

84

79

556

69

59

69

70

765

45

16

75

73

559

56

41

50

53

769

33

6

57

55

563

36

22

26

33

772

21

0

34

30

565

20

7

9

12

776

12

0

14

15

567

7

4

1

2

781

9

0

4

4

572

2

0

0

0

786

10

4

0

0

577

1

1

1

1

792

17

16

2

0

579

2

2

3

2

798

28

29

6

29

583

6

2

11

9

802

44

48

16

12

585

12

3

23

20

806

54

64

34

29

587

27

8

42

37

811

63

78

59

51

589

41

26

63

55

818

76

83

79

75

592

59

45

79

69

823

78

81

91

87

595

75

62

90

81

833

70

52

95

91

597

85

77

92

86

836

61

34

90

89

600

91

90

95

89

840

51

23

80

80

609

93

92

94

88

845

45

14

62

64

611

93

92

92

87

850

37

11

37

39

614

92

90

93

87

855

34

9

20

18

622

90

79

91

82

860

33

12

16

13

629

95

88

94

84

866

38

26

23

26

638

96

99

98

91

872

42

37

41

38

643

96

99

98

92

878

59

53

43

46

650

97

99

99

92

883

67

73

54

51

657

86

82

99

90

889

73

83

65

59

660

77

60

95

87

894

81

90

76

65

662

70

49

91

80

900

92

98

90

87

666

62

37

75

68

927

96

100

100

96

670

62

36

60

54

958

100

100

100

100

672

62

60

57

48

AS STUDIED BY MEANS OF THE KADIOMICKOMETEK. 33

Table 3. — Observed Transmissions of Neodymium Bromide Solutions.

X

N. = 1.66.

N. = 1.66.

N.=0.415.

N. =0.208.

X

N. = 1.66.

N. = 1.66.

N. = 0.415.

, N. =0.208.

D. = 2.5mm.

D. = 5mm.

D. = 10mm.

D.=20mm.

D.=2.5mm

, D.=5mm.

D. = 10mm.

D.=20mm.

486

64

44

84

83

672

71

52

09

67

492

64

46

81

84

676

67

57

68

70

495

60

35

73

76

678

67

73

75

76

499

38

22

54

53

681

71

77

83

80

501

29

11

45

43

685

82

89

87

90

503

2

9

33

26

691

95

92

92

93

505

17

7

21

25

699

82

80

95

92

507

15

4

26

25

706

72

60

93

91

509

17

9

30

29

710

55

45

88

S3

513

18

6

28

25

714

38

28

72

67

515

13

6

21

14

719

21

11

48

39

516

9

4

11

10

722

12

6

19

8

518

16

7

28

25

724

7

2

6

1

520

28

12

39

41

729

10

5

1

0

522

43

19

55

51

733

19

11

2

1

525

53

32

66

73

737

32

21

8

12

530

71

54

89

86

741

45

30

26

25

535

79

57

100

93

745

59

40

49

49

544

84

65

100

94

750

67

49

73

69

555

77

48

95

93

755

67

50

92

80

556

69

47

86

77

760

62

47

100

83

559

57

37

73

62

765

51

37

95

79

563

42

23

51

39

769

40

25

82

65

565

21

12

26

17

772

28

16

59

44

567

9

3

10

5

776

18

9

37

24

572

3

2

2

1

781

18

6

17

8

577

3

1

3

3

786

15

7

5

1

579

6

2

9

10

792

21

12

3

1

583

6

5

17

21

798

33

21

10

9

585

24

11

29

36

802

46

27

23

24

587

39

23

47

51

806

57

35

44

42

589

55

34

63

67

811

69

45

64

63

592

69

53

72

77

818

76

62

83

82

595

75

61

86

81

823

80

66

98

91

597

78

65

91

89

833

77

58

94

94

600

83

73

93

95

836

68

53

91

91

605

87

73

92

94

840

58

44

82

82

614

86

74

94

92

845

50

34

69

70

622

88

76

85

84

850

47

28

48

46

629

88

78

90

86

855

43

24

33

25

638

91

83

92

93

866

47

27

35

38

643

92

98

94

94

872

53

34

48

52

650

91

90

94

92

878

61

42

56

60

657

88

83

93

94

883

69

48

63

64

660

81

72

90

91

889

75

57

70

71

662

79

62

87

85

900

87

69

87

92

666

74

56

78

78

958

97

84

98

100

670

68

50

70

65

34 ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS OF CERTAIN SALTS
Table 4. — Observed Transmissions of Neodymium Nitrate Solutions.

X

N.=2.95.

N.=2.95.

N. =0.736.

N.=0.308.

X

N.=2.95.

N.=2.95.

N.=0.736.

N. =0.368.

D.=2.5mm.

D.=5mm.

D. = 10mm.

D.=20mm.

D. = 2.5mm.

D.=5mm.

D. = 10mm.

D.=20mm.

486
492
495
499

80
71
61
39

95

93
74
49

80

76
60
45

672
676
678
681

67

75

78
85

53
51

71

77

61
60
65

76

61
61
64

77

25

501

16

23

38

37

685

88

82

86

84

503

11

14

25

17

691

88

87

92

90

505

4

7

25

7

699

81

78

94

91

507

9

4

19

11

706

64

51

86

86

509

17

4

23

14

710

36

33

80

78

513

11

4

21

14

714

26

16

65

65

515

13

2

16

6

719

11

8

35

39

516

7

3

10

4

722

6

2

14

10

518

13

4

12

8

724

4

1

2

3

520

28

5

27

26

729

8

1

0

0

522

41

7

43

37

733

17

4

0

0

525

53

15

55

54

737

33

11

3

3

530

83

39

83

79

741

42

21

11

12

535

86

72

91

92

745

52

25

29

30

544

87

85

88

87

750

56

40

50

50

555

75

82

78

78

755

56

42

68

66

556

58

67

69

71

760

45

34

78

75

559

35

54

54

55

765

35

26

76

74

563

15

38

32

35

769

23

15

61

61

565

3

25

14

16

772

13

7

40

39

567

1

12

5

3

776

8

3

20

23

572

0

6

2

0

781

7

2

8

15

577

0

2

3

0

786

12

3

1

2

579

3

2

3

1

792

19

7

1

1

583

4

0

7

6

798

33

14

3

5

585

12

1

13

16

802

48

26

12

11

587

26

1

27

26

806

59

41

27

27

589

44

7

43

44

811

61

54

47

46

592

64

14

62

62

818

75

70

71

65

595

81

27

75

77

823

74

74

88

81

597

89

42

84

85

833

63

64

98

92

600

92

58

89

92

836

52

54

98

89

605

98

82

88

91

840

42

40

82

75

614

85

87

88

91

845

39

29

76

69

622

85

86

87

84

850

36

22

48

46

629

86

85

91

86

855

39

18

28

23

638

89

97

94

866

53

28

30

25

643

94

92

98

96

872

59

37

41

36

650

93

90

95

95

878

66

45

52

45

657

78

85

95

100

883

75

58

59

53

660

71

85

90

91

889

82

67

68

62

662

67

74

84

83

900

88

81

91

83

666

62

67

73

74

958

95

100

100

100

670

64

60

61

61

AS STUDIED BY MEANS OF THE RADIOMICROMETER. 35

Table 5. — Transmission of Neodymium Acetate; Transmission of Neodymium Sulphate.

N.=0.84.

N.=0.84.

N.=0.118.

A

N.=0.84.

N.=0.84.

N. =0.118.

A

D.=2.5mm.

D.=5 mm.

D. = 10mm.

D. =2.5 mm.

D. = 5 mm.

D. = 10mm.

486

86

93

89

672

87

83

91

492

88

90

95

676

88

84

92

495

93

85

98

678

90

88

94

499

S3

65

97

681

94

93

95

501

73

50

S6

685

96

90

95

503

67

39

78

691

98

94

96

505

55

29

72

699

93

88

97

507

57

31

72

706

89

67

98

509

53

40

73

710

72

47

97

513

57

35

SO

714

52

27

94

515

52

26

62

719

35

12

83

516

39

12

55

722

21

3

58

518

36

19

67

724

14

4

42

520

44

33

79

729

22

14

31

522

59

53

89

733

34

28

31

525

72

70

92

737

49

42

40

530

83

91

100

741

64

60

59

535

96

96

100

745

79

74

78

544

98

96

100

750

86

78

89

555

94

93

100

755

88

74

94

556

88

81

100

760

87

64

96

559

85

69

95

765

76

47

97

563

66

39

85

769

59

32

94

565

46

21

63

772

42

17

86

567

27

5

38

776

28

8

75

572

12

0

2S

781

22

9

50

577

4

0

29

786

25

17

29

579

3

1

42

792

26

29

25

583

8

3

55

798

51

43

40

585

19

12

69

802

63

59

53

587

32

23

81

806

77

72

74

589

48

40

93

811

86

83

84

592

62

50

97

818

91

88

92

595

76

73

100

823

95

89

95

597

85

90

100

833

94

87

96

600

91

98

100

836

95

65

96

605

97

97

98

840

77

53

95

614

99

96

97

845

68

46

94

622

99

96

94

850

63

46

83

629

94

95

97

855

65

47

71

638

98

93

100

866

73

63

71

643

96

92

99

872

78

74

81

650

99

93

99

878

82

82

85

657

93

87

98

883

90

87

88

660

91

80

98

8S9

93

90

89

662

89

80

98

900

96

96

90

666

88

82

93

958

98

100

99

670

86

80

93

36 ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS OF CERTAIN SALTS

When we first began to investigate any given salt we made a preliminary
survey of its spectrum, noting the approximate positions of the absorption
lines and bands. We then made our observations very close together over
the regions in which the preliminary survey had indicated the presence of
lines and bands. The number of absorption lines and bands, as is well
known, is very great in the case of neodymium compounds, and these lines
and bands frequently have very sharp edges. This made the work with this
substance very difficult. The proper width of slit and position had to be
chosen or a considerable error would result. Given a slit width which was
approximately the same as that of an absorption line, a very slight move-
ment of the slit or prism would change very greatly the total amount of
energy falling on the thermal junction.

Take the neodymium band X4275. which is very intense but narrow. On
both sides of this band there is a region of almost perfect transparency. If
the slit width necessary to give the desired deflection was greater than the
width of this band, light would pass through around the edges of the band,
and an error, which might be of very considerable magnitude, would result.
With substances which did not contain such fine lines and bands the work
would be much simpler.

The entire spectrum from wave-lengths X4000 to X20000 was observed at
intervals of from 20 a.u. to 50 a.u., except in the regions where the pre-
liminary survey indicated the absence of absorption lines and bands.

An examination of table 2 will show at X486 a transparency of 72 per
cent, which rapidly decreases, reaching the first minimum at X505. There
the transparency amounts to only 2 per cent. The transparency then
increases a little and quickly drops to 6 per cent at X515. The transparency
then increases, becoming nearly complete at X535. We have here, then, a
double band with greater absorption on the red side. Other minima appear
at X572, X730, X786, and X860. Bands X730, X786, and X860 do not appear
on the photographic plate, and the last two seem never to have been detected
before. The above wave-lengths are given as in the tables.

The salts of neodymium were studied as far as X20000, but beyond 1/x
there seems to be complete transparency. The absorption of water is, as is
well known, very great in the region X12000 to X20000.

DISCUSSION OF THE RESULTS.

The results are plotted in figs. 1 to 11. The abscissae are percentage trans-
parencies, the ordinates are wave-lengths. These curves, since they repre-
sent the transparencies of the solutions in question, are called transmission
curves.

Figs. 1, 2, and 3 represent the transparency of solutions of neodymium
chloride expressed in terms of Beer's law. If we represent the concentration
by N and the depth of layer by d,

Nd= constant

AS STUDIED BY MEANS OF THE RADIOMICROMETER. 37

The concentration represented in fig. I1 is 3.43 normal, in iig. 2 it is 0.857
normal, and in fig. 3 it is 0.427 normal. The depth of layer represented by
fig. 1 is 2.5 mm., by fig. 2 it is 10 mm., and by fig. 3 it is 20 mm. The con-
centration and depth of layer were thus varied so as to keep Nd constant.

If the solvent plays no role in the absorption, the three sets of curves must
fall directly over one another, i. c, be identical, since the number of absorb-
ing parts in the path of the beam of light is kept constant. A comparison of
the curves shows that, in general, the more concentrated the solution the less
the transparency and the broader the absorption bands. In the more dilute
solution the intensity of the bands is greater. This comes out very clearly
in the red and infra-red region, where there is greater accuracy of measure-
ment.

Take the three absorption bands, X730, X785, and X860. In curve 1
the minima of these bands are approximately 4, 9, and 33 per cent, while
the minima in curve 2 are much less. In fig. 2 the bands X730 and X785
reach the abscissa, which means that there is no transmission. At this dilu-
tion the band X860 has still considerable transparency, as will be seen by
the fact that it remains a considerable distance above the abscissa. The
band X860 does not reach the abscissa even at the dilution represented in
fig. 3.

All of the bands manifest the above phenomena, the change in intensity
being greatest where the change in dilution is greatest, i. e.,from curve 1 to
curve 2. With increase in dilution the position of the middle of the band is
displaced toward the region of greater wave-length.

Similar results were obtained with neodymium bromide, and these are
plotted in curves 4, 5, and 6. The concentrations and depths of layer were
varied so that the product of the two remained constant. The work with
the bromide was, therefore, done in terms of Beer's law. The concentra-
tions used were 1.66 normal, 0.415 normal, and 0.208 normal, the correspond-
ing depths of the solution being 2.5 mm., 10 mm., and 20 mm. We find here
the same general changes in the intensities of the bands as with the chloride.
The more dilute the solution the more intense and the narrower the band.

This is shown by comparing figs. 4, 5, and 6. In fig. 4, which represents
the most concentrated solution of the three, the bands are the least intense.
In fig. 5 the opacity of two of the bands has become complete, shown by the
fact that these touch the abscissa.

Neodymium nitrate was also studied and the results are plotted in curves
7, 8, and 9. The concentrations used were 2.95, 0.736, and 0.368 normal.
The depths of layer were 2.5 mm., 10 mm., and 20 mm.

Band X570, curve 7, appears to be an exception to the general relation
pointed out above, connecting intensity and width of band with dilution.
This was the first band studied by means of the radiomicrometer, and com-
paratively small deflections were observed in this region of the spectrum.

1 Our attention was drawn to the existence of these bands in the infra-red by Pfund,
who had already mapped them radiometrically for neodymium nitrate.

38 ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS OP CERTAIN SALTS

The remaining bands of neodyniium nitrate, however, show the same rela-
tions that have been pointed out for the chloride and bromide ; with increas-
ing dilution the intensities of the bands increase and the centers seem to be
displaced somewhat towards the longer wave-lengths.

We then have three salts, neodymium chloride, neodymium bromide, and
neodymium nitrate, all of which show a marked increase in the intensity of
the absorption bands with increase in dilution, when the product of con-
centration and depth of layer is kept constant, i. e., when the conditions
demanded by Beer's law are fulfilled.

POSSIBLE EXPLANATION.

It is well known that a resonator vibrates more strongly if excited by the
vibrations from one single vibrating resonator of the same pitch than when
set into vibration by a large number of resonators, one of which has the same
period as its own, and the others slightly different periods. In other words,
if several vibrators are near one another, every one exerts a certain influence
on its neighbors. The result is that no one of them has exactly the same
period as the original resonator.

The presence of one vibrator seems to exercise a damping influence on the
other, and causes it to vibrate with a period slightly different from its normal
period. We thus have less perfect resonance.

The absorption of light by solutions appears to be a resonance phenomenon.
In a concentrated solution the vibrators are relatively close to one another
and mutually affect one another. The result is an imperfect resonance, and
consequently the absorption bands are less intense in the more concentrated
solution.

The vibrators are farther removed from one another in the more dilute
solutions, and in most cases are probably surrounded by large amounts of
water of hydration. The damping effect would not be so pronounced, and
a resonator would have greater freedom to vibrate in its own period. In
such cases we would have a more nearly perfect resonance, and the resulting
absorption bands would be more intense. This tentative explanation seems
to account for the observed facts. Subsequent work has shown that a part
of this effect can be explained as due to the fact that the slit width was not
infinitesimal. Fig. 10 is plotted from the results for neodymium sulphate,
and fig. 11 from those for neodymium acetate. The concentration of the
sulphate is 0.118 normal, and of the acetate 0.84 normal. The length of the
solution of the sulphate is 10 mm., and of the acetate 2.5 mm.

The absorption of the acetate, for a given concentration, is much greater
than that of any other neodymium salt thus far studied. This agrees with
the results obtained photographically.

The absorption of water beyond lju is very great, as has already been
stated. If we are working with very concentrated solutions and use a
"water" vessel of the same thickness as the "solution" vessel, it is obvious
that the results would not be comparable. Take the 3.43 normal solution

AS STUDIED BY MEANS OF THE RADIOMICROMETER.

39

of neodymium chloride; it contains, for a given thickness, only about 90 per
cent as much water as the same thickness of pure water. It is, then, obvious
that in the longer wave-lengths a correction term must be introduced for this
difference. This was practically negligible with salts of neodymium, since
these do not seem to have any bands in the region where water has appre-
ciable absorption.

Salts of praseodymium have bands in the infra-red, at least as far as 2ju.
In such cases the above correction must be introduced. This correction
can be introduced in either of two ways. We can take the specific gravity
of the solution and from the concentration calculate the amount of water
present. We can then use a "water" vessel of suitable thickness. For
example, if the very concentrated solution in question contains only 90 per
cent of water, and we use a vessel for the solution which is 10 mm. thick, we
must use a vessel for the water which is only 9 mm. thick. In this way the
beam of light is made to pass through the same amount of water both in the
case of the solution and of the solvent, and the absorption due to water is,
therefore, the same in the two cases.

The second method of procedure is to allow the "water" vessel and the
"solution" vessel to be of the same thickness, and to apply mathematically
the proper correction to the results obtained.

OAji

0.6//. Q.I/jl

Fig. 1.

i.cy

40 ABSORPTION SPECTRA OF AQUEOUS SOLUTIONS OF CERTAIN SALTS

0
0.4/*

0.6/i 0.7/i

Fia. 5.

OAju

OSju

0.6/x 0.7/i

Fig. 6.

AS STUDIED BY MEANS OF THE KADJLOMlCUOMETEK.

41

0.4// 0.5/ti

0.6yU 0.7//

Fig. 9.

Q.S/U.

CHAPTER V.

THE ABSORPTION OF LIGHT BY WATER CHANGED IN THE
PRESENCE OF STRONGLY HYDRATED SALTS, AS SHOWN BY
THE RADIOMICROMETER— NEW EVIDENCE FOR THE SOLVATE
THEORY OF SOLUTION.

The use of the radiomicrometer in studying the absorption spectra of cer-
tain substances has already been discussed by Jones and Guy.1 The radio-
micrometer was used for studying the absorption spectra of solutions
rather than the grating spectrograph and the photographic plate, because it
enabled us to measure not only the positions of the different lines and bands,
but also to study quantitatively their intensity; and also because it made
possible the study of the absorption spectra of solutions over a much greater
range of wave-lengths than the photographic method.

In building a radiomicrometer adapted to this work — that is, with suffi-
cient sensibility and with a short period — one of the greatest difficulties, as
already mentioned, was to obtain copper wire free from iron. This was a
necessity, since the presence of an appreciable quantity of iron in the copper
gave rise to a "magnetic control" which rendered the instrument unstable
and the zero-point inconstant. This difficulty was for the most part over-
come, due to the kindness of Messrs. Leeds and Northrup of Philadelphia
and of R. W. Paul of London. Both of these houses furnished us with
copper wire so free from iron that the "magnetic control" could easily be
regulated. By means of this wire and the thermo-electric junction already
described, a most sensitive radiomicrometer was built, which at the same
time had a very short period, and with this instrument work was done with
salts of neodymium and praseodymium, the results of which were recorded
in the Physikalische Zeitschrift.2

ABSORPTION OF FREE AND COMBINED WATER.

At the beginning of the academic year 1912-13 the absorption spectra
of solutions of a large number of salts of different metals were mapped out
and compared with the absorption of water, using the same depths of water
as the water in the various solutions. The depth of water in the solution
was determined from the concentration of the solution and from its specific
gravity. It was soon found that the absorption of the solution was less, and
in many cases very much less than that of the layer of water having a depth
equal to the depth of the water in the solution.

The above result is directly at variance with everything that was known
at the time. The dissolved substance could not have less than no absorp-
tion of light, the assumption having been made up to this time that in an

1 Phys. Zeit., 13, 649 (1912). 2 Ibid., 13, 651 (1912).

43

44 ABSORPTION OF LIGHT BY WATER CHANGED

aqueous solution the water present absorbs just as much as pure, unconi-
bined water.

It became at once obvious that we could not measure the absorption
spectrum of a solution, subtract from it the absorption due to water, and
conclude that the remainder was the absorption due to the dissolved sub-
stance ; since the water in the solution has very different absorption from an
equal amount of pure, uncombined water.

We then carried out a number of experiments in cells whose depths could
be easily and accurately adjusted, with different substances, in the following
manner: The absorption spectra of a number of different substances were
first measured , then the absorption spectra of water having the same depths
of layer as the water in the solutions. For certain substances the pure
water was more opaque than the solutions, and for other substances the
water was more transparent. The percentage transmission — that is, the
deflection of the racliomicrometer for the solution, divided by the deflection
for water — for the first-named substances amounted to more than 100 per
cent. Pure water had a different absorption from an equal depth of water
in the solution, and since this difference varied from one dissolved substance
to another, it is obvious that this method was not the one to be followed.
It would be very difficult, not to say impossible, to interpret the results
obtained by dividing the radiomicrometer deflections for the solution by
those for pure water. We should simply be obtaining the transmission of
the solution in terms of pure water, which was not what was desired.

What we want to know is the actual absorption or transmission of the solu-
tion, and then that of pure water having a depth of layer that was just equal
to that of the water in the solution. These two sets of results could then be
compared with one another.

HYDRATED AND NONHYDRATED SUBSTANCES.

In this earlier work we had, however, noted that solutions of those sub-
stances which are largely hydrated are more transparent than pure water
having the depths of the water in the solutions in question. Solutions of
nonhydrated substances, or of only slightly hydrated substances, provided
the substances themselves do not absorb light, are not more transparent
than pure water having the same depths as the water in the solution. It
would seem from this observation that water combined with the dissolved
substance had less absorption of light than pure, uncombined water. To
test this quantitatively the following procedure was adopted.

METHOD OF PROCEDURE.

A solution of the substance in question was prepared of known concen-
tration and its specific gravity determined. This solution was placed in one
cell set to a depth of say 21 mm. Some of the same solution was then
placed in another cell set to a depth of say 1 mm. Light of a known wave-
length was then passed through the one solution and the deflection noted.
Light of this same wave-length was then passed at once through the other

IN THE PRESENCE OF STRONGLY HYDRATED SALTS. 45

solution and the deflection in this case also noted. The deflection produced
when the deeper solution was in the path of the beam of light was then
divided by the deflection produced by the shallower solution, and this gave
the absolute transmission of the solution of the substance in question of
known concentration, having a depth of layer of 20 mm.

This process was repeated for the different parts of the spectrum, chang-
ing the wave-length of light from reading to reading b}r only a small amount.
The object of using the two depths of the same solution, and then dividing
the deflection produced by the deeper layer by that obtained when the more
shallow layer was in the path of the beam of light, was to eliminate any
effect of reflection from the glass ends closing the cells containing the solu-
tions, and also to eliminate any changes in the total amounts of energy sent
through the solution, due to slight changes in the intensity of the Nernst
glower. From the specific gravity of the solution and its known concentra-
tion, the amount of water in a layer of the solution, say 21 mm. in depth,
could easily be calculated. Similarly, the amount of water in a layer of the
solution which was 1 mm. deep, could also be calculated. Water was then
introduced into two cells, and the cells so adjusted that the difference in
depths was exactly equal to the depth of the water in the layer of the solu-
tion, which was 20 mm. deep.

The deflection for the water in the deeper cell was then read for any given
wave-length of light, and then, at once, the deflection when the light was
passed through the more shallow layer of water. The deflection for the
deeper layer was divided by the deflection for the shallower layer. The
result was the absolute transmission for water with a depth of layer just
equal to the depth of water in the solution in question.

RESULTS.

The above results for the solution are plotted as one curve and those for
water having the same depth as the water in the solution as another curve,
wave-lengths being abscissae and transmission ordinates. A comparison
of the two curves shows at once whether water in the free, uncombined con-
dition or the same depth of water in the solution in question is the more
transparent.

The data obtained by dividing the deflections produced by the deeper
solutions by those for the shallower, and, similarly, by those for water, are
also given in tables 6 to 10. These are the data from which the accom-
panying curves were plotted.

The substances studied were chosen from the standpoint of their power to
solvate or to combine with the solvent in which they were dissolved. In all
of the work recorded in this paper the solvent used was water. We were
practically limited, in this phase of the work, to those substances which
themselves have little or no power to absorb light, and which are both color-
less in the visible part of the spectrum, and have little or no absorption in
the regions in which the absorption bands of water occur.

46

ABSORPTION OF LIGHT BY WATER CHANGED

We selected for these substances with little or no hydrating power, salts
of potassium and ammonium. The potassium salts studied were the chlo-
ride and nitrate. Ammonium chloride and nitrate were also investigated.
For the salts with large hydrating power, calcium chloride, magnesium
chloride, and aluminium sulphate were used. These salts were shown, from
our earlier work, to be among the most strongly hydrated substances with
which we are familiar. Two depths of layer of each solution of every

Table 6.

X

KC1, 4N.
Hh

H20.

NH4C1,4N.

I/h

H20.

NH4NO3,

3.12 N.

Ilh

H20.

711

97

97

92

98

95

98

724

96

95

91

96

97

98

741

95

95

90

92

96

96

760

93

95

85

92

91

95

776

92

95

85

88

92

94

798

94

95

88

95

91

95

818

92

95

87

95

92

96

836

94

93

87

95

91

94

855

91

90

86

93

89

92

878

92

90

86

91

90

93

900

90

89

84

87

89

88

922

87

86

82

86

85

90

947

82

84

79

82

82

83

958

77

78

73

72

73

79

964

75

73

69

70

70

71

969

65

65

64

63

65

67

974

58

56

57

58

58

59

979

51

50

52

52

50

54

982

47

46

45

46

44

46

985

41

45

43

43

40

44

991

39

43

41

43

39

43

1,007

39

43

41

42

39

44

1,013

40

42

39

44

41

44

1,019

42

46

40

46

42

44

1,025

41

42

44

49

45

48

1,032

49

49

44

49

48

49

1,037

53

52

56

55

52

53

1,042

56

56

53

58

55

55

1,046

59

60

57

57

57

60

1,059

63

62

58

65

60

65

1,065

68

67

62

67

64

65

1,072

71

68

64

68

66

69

1,078

74

72

67

67

66

72

1,085

75

73

67

66

68

74

1,100

77

75

68

72

69

78

1,113

76

76

69

72

69

78

1,138

75

72

68

70

68

72

1,148

70

69

64

65

64

71

1,158

64

63

62

64

59

64

1,165

58

59

58

58

56

60

1,172

52

51

50

50

53

52

1,179

42

40

40

40

38

41

1,186

29

28

29

26

29

30

1,193

18

19

19

19

18

19

1,200

13

16

14

13

12

17

1,206

10

12

12

13

9

13

1,213

10

11

10

12

9

13

1,220

10

11

10

11

9

12

1,227

10

11

10

11

9

12

IN THE PRESENCE OF STRONGLY HYDRATED SALTS.

47

substance investigated were employed, in order to bring out the two most
important water-bands in the region of the spectrum used. This could
not be done by studying only one depth of solution, since the depth which
was necessary and sufficient to bring out clearly one of these water-bands
would not bring the other out in the way desired. By using the two depths
of solution, and studying them in the manner above described — that is,
by the differential method — we were able to study both of the water-bands
as produced, on the one hand by the pure solvent, and on the other by the
solution.

In tables 6 to 10, under X, are given the wave-lengths of light that were
passed through the solution; and under I/Iq the percentage of transmission,
on the one hand, of the solution; and on the other, of water having a depth
exactly equal to that of the water in the solution.

Table 7.

X

KC1, 4 N.

I/h

H20.

NH4C1,4N.
Ilh

H20.

NH4NO3,
3.12 N.

Ilh

H20.

1,085

85

86

79

87

81

88

1,100

87

88

80

92

81

93

1,113

86

87

79

86

84

86

1,138

81

85

79

84

81

84

1,148

79

82

77

84

78

84

1,158

80

79

74

81

76

81

1,165

76

77

71

77

72

77

1,172

72

71

66

70

67

70

1,179

64

62

59

62

61

62

1,186

51

51

52

50

50

50

1,193

41

43

42

44

40

44

1,200

35

38

37

40

34

40

1,206

37

36

35

37

32

37

1,213

30

34

30

36

32

36

1,220

30

35

29

35

34

35

1,227

30

34

30

35

32

35

1,233

30

33

28

35

31

35

1,241

30

34

29

34

31

34

1,248

31

34

28

33

31

33

1,250

33

34

30

34

32

34

1,255

34

35

30

36

32

36

1,268

34

35

31

37

32

37

1,270

37

37

38

38

38

38

1,285

38

38

33

38

33

38

1,295

39

38

32

38

33

38

1,300

41

38

32

39

34

39

1,308

42

39

32

41

35

41

1,316

41

39

32

40

34

40

1,323

42

37

32

38

34

38

1,330

40

37

32

37

33

37

1,338

40

35

30

35

31

35

1,346

36

33

28

36

30

36

1,352

34

29

26

29

27

29

1,358

29

26

23

27

25

27

1,365

25

22

21

23

21

23

1,372

20

18

17

20

17

20

1,387

13

12

12

11

10

11

1,404

7

7

7

8

7

8

1,418

3

4

4

3

3

3

1,430

2

2

3

2

2

2

1,445

0

0

0

1

1

1

48

ABSORPTION OF LIGHT BY WATER CHANGED

In table 6 the depth of layer of all the solutions was the difference between
21 and 1, i. e., 20 mm. The depth of water was in every case the same as
that of the water in the solution in question.

The depth of layer of the solutions given in table 7 was the difference

between 11 mm. and 1 mm., i. e., 10 mm., and was only half of that in

table 6. The object of this was to bring out more prominently the second

water-band. The depth of water used was in every case the same as that

of the water in the solution.

Table 8.

CaCU,

MgCU,

Al»(S04)s,

X

5.38 N.
IJh

H20.

4.96 N.
Ilh

H20.

1.012 N.
Ilh

H20.

710

94

98

95

98

95

93

724

92

98

98

98

95

95

741

90

95

95

98

94

93

760

87

94

94

98

92

93

776

88

93

92

97

93

95

798

91

96

93

94

92

90

818

93

99

90

90

93

92

836

92

97

92

95

92

92

855

90

93

91

90

90

91

878

90

90

91

93

91

90

900

89

92

88

92

89

90

922

86

91

88

91

85

86

947

87

84

84

86

82

81

958

78

79

76

78

76

73

961

75

73

82

76

72

66

969

70

68

75

69

68

61

974

65

62

68

65

64

55

979

59

53

61

56

58

48

982

51

49

48

51

53

42

985

48

49

54

45

51

40

991

44

46

48

49

47

39

1,007

42

46

46

48

46

38

1,013

42

46

45

50

46

39

1.019

43

49

44

51

44

40

1,025

47

50

46

44

46

43

1,032

52

53

51

54

46

45

1,037

55

55

52

56

52

50

1,042

58

59

56

58

53

53

1,046

62

62

59

65

55

55

1,059

66

65

63

67

55

58

1,065

71

70

69

70

62

63

1.072

74

72

71

75

60

65

1,078

75

74

71

76

64

69

1,085

78

76

76

79

65

70

1,100

80

77

78

79

67

72

1,113

79

78

80

81

67

74

1,138

77

75

77

78

64

67

1,148

74

71

75

77

60

65

1,158

69

65

73

73

57

55

1,165

66

62

65

65

55

53

1.172

61

52

61

58

50

43

1,179

54

41

52

44

45

34

1,186

42

30

43

32

34

22

1,193

32

21

32

24

25

15

1,200

22

17

23

18

20

12

1,206

16

16

18

17

16

10

1,213

13

15

16

18

14

9

1,220

12

13

14

15

11

10

1,227

12

13

14

16

12

8

TN THE PRESENCE OF STRONGLY HYDRATED SALTS.

49

The depth of layer oi' the solutions given in tabic 8 was the difference
between 21 mm. and 1 mm., i. e., 20 mm. The depth of water used in every
case was the same as that of the water in the solution.

In table 9 the depth of layer used was the difference between 11 and 1 mm.,
i. e., 10 mm. The object of using the smaller depth of the solution was to
bring out more clearly in the case of hydrated salts the second water-band.

When salts wrhich are strongly hydrated in aqueous solution are not very
concentrated, the difference between the transparency of the salt solution
and that of wTater of the same depth of layer as the water in the solution is
not so pronounced. This is what would be expected, since the total amount
of water combined with the dissolved salt increases with the concentration of
the solution. The data given in table 10 bring out this fact.

Table 9.

CaCl2,

MgCl2,

! A12(S04)3,

i

X

5.38 N.
I/h

H20.

4.96 N.
Ilh

H20.

1.02 N.
Ilh

H20.

1,085

84

84

82

84

79

82

1,100

84

84

83

84

78

81

1,113

84

86

83

84

78

84

1,138

86

85

82

83

77

83

1,148

82

83

79

80

75

80

1,158

80

79

77

77

73

77

1,165

78

76

77

75

70

73

1,172

76

72

74

70

66

69

1,179

72

65

71

64

63

58

1,186

63

55

62

52

55

48

1,193

54

45

56

46

49

39

1,200

45

40

48

42

43

34

1,206

38

39

42

40

38

30

1,213

35

39

39

38

36

29

1,220

33

36

37

38

34

28

1,227

32

36

36

38

32

28

1,233

31

35

34

37

32

28

1,241

32

35

34

34

31

28

1,248

32

35

34

37

31

28

1,250

33

35

34

38

31

28

1,255

33

37

34

38

31

28

1,268

35

38

33

38

30

29

1,270

37

39

34

39

30

30

1,285

38

40

35

40

30

31

1,295

40

40

35

40

30

32

1,300

42

41

36

41

30

32

1,308

42

41

37

42

30

33

1,316

45

41

39

42

30

33

1,323

47

40

39

41

28

33

1,330

46

39

40

40

27

32

1,338

45

37

38

38

27

30

1,346

42

35

38

36

24

27

1,352

40

32

34

33

22

24

1,358

37

29

33

30

20

21

1,365

33

25

29

26

18

19

1,372

29

21

25

22

15

15

1,387

19

13

18

15

10

9

1,404

12

10

12

11

7

5

1,418

7

6

7

9

4

3

1,430

3

3

4

3

2

2

1,445

2

1

1

1

1

1

50

ABSORPTION OF LIGHT BY WATER CHANGED

The depth of layer of the different solutions for which the results are
recorded in table 10 was the difference between 21 and 1 mm., i. e., 20 mm.
The results are, therefore, comparable with those recorded in table 8, the
difference being a difference in the concentrations of the solutions used.
The difference between the transmission of the solution and that of water at
the same depth as the water in the solution is very much less for the more

Table 10.

CaCl,

MgCli

A12(S04)S

X

2.69 N.
I/Io

H20.

2.48 N.
I/h

H20.

0.508 N.
I/h

H20.

710

96

94

95

95

97

96

724

95

96

93

96

98

96

741

95

95

90

95

95

93

760

94

96

92

95

95

95

776

93

97

93

95

95

95

798

90

98

91

95

96

96

818

93

97

91

93

95

96

836

91

96

89

93

93

95

855

91

92

88

92

92

92

878

90

92

84

90

90

91

900

88

90

84

88

89

86

922

89

92

81

86

82

85

947

82

86

78

83

78

80

958

75

79

72

76

73

75

964

70

74

70

73

68

69

969

65

69

62

64

62

62

974

58

61

58

58

57

54

979

50

52

50

51

50

46

982

44

47

46

46

46

42

985

40

43

42

43

43

40

991

39

41

40

41

41

39

1,007

38

40

41

42

40

40

1,013

39

42

40

44

40

40

1,019

40

43

40

44

41

41

1,025

43

45

44

41

43

43

1,032

45

47

47

44

45

46

1,037

48

50

50

46

47

48

1,042

51

62

52

48

49

49

1,046

56

56

56

54

53

54

1,059

61

59

58

55

60

59

1,065

65

64

64

62

59

62

1,072

69

67

67

64

63

65

1,078

70

69

69

67

65

69

1,085

72

72

72

68

68

72

1,100

73

73

73

71

69

73

1,113

72

74

74

72

68

74

1,138

72

74

74

70

67

72

1,148

66

69

69

67

64

67

1,158

67

62

62

60

58

62

1,165

57

58

58

58

54

54

1,172

52

51

53

52

47

46

1,179

46

39

42

42

39

35

1,186

30

27

31

27

28

25

1,193

20

19

21

20

20

16

1,200

13

14

15

15

14

12

1,206

12

11

12

13

12

10

1,203

11

11

12

12

11

10

1,220

11

10

11

11

11

10

1,227

10

9

10

10

10

9

IN THE PRESENCE OF STRONGLY HYDRATED SALTS.

51

dilute than for the more concentrated solutions; this is what would be
expected in terms of the solvate theory applied to the phenomenon in
question.

Considerable work was done in comparing directly the transmission of a
solution and that of water having the same depth as the water in the solu-
tion in question. The deflection of the radiomicrometer as given by the
solution is in the column marked " Deflection of solution," and the deflection
as given by water having the same depth as water in the solution is given
in column "Deflection of water."

Table 11.

X

Deflection
of solution

of
Ah (S04)3.

Deflection
of water.

Deflection

of solution

of KC1.

Deflection
of water.

X

Deflection
of solution

of
Ah (S0,)3.

Deflection
of water.

Deflection

of solution

of KC1.

Deflection
of water.

710

50

51

53

53

1,037

91

84

112

108

724

58

58

56

56

1,042

92

92

119

116

741

62

63

67

68

1,046

99

99

125

120

760

72

72

77

77

1,059

105

105

141

136

776

75

76

88

90

1,065

109

112

150

145

798

83

83

98

99

1,072

114

119

159

153

818

82

82

108

109

1,078

118

125

164

158

836

93

94

116

116

1,085

122

132

168

164

855

97

97

124

124

1,100

128

140

176

172

878

105

105

129

130

1,113

129

142

178

175

900

105

105

140

138

1,138

127

142

174

170

922

112

112

140

140

1,148

123

131

164

162

947

113

113

142

142

1,158

112

118

161

159

958

109

106

136

136

1,165

108

111

157

154

964

107

100

129

125

1,172

99

94

132

126

969

104

93

118

116

1,179

87

74

107

100

974

98

83

108

106

1,186

68

49

73

66

979

93

73

92

92

1,193

54

35

50

48

982

82

66

83

83

1,200

42

26

34

36

985

80

64

78

80

1,206

35

23

29

32

991

78

62

78

80

1,213

30

21

25

30

1,007

78

65

78

81

1,220

28

20

24

29

1,013

74

65

81

85

1,227

26

19

24

28

1,019

77

68

84

88

1,241

24

19

23

26

1,025

80

75

96

96

1,255

23

18

25

27

1,032

84

77

100

101

The results obtained for aluminium sulphate having a concentration
1.02 N, and for potassium chloride 4 N are given in table 11. The depth
of solution used was 20 mm., and the depth of water that of the water in the
solutions in question.

Duplicate measurements were made with the radiomicrometer for nearly
every solution of all the substances worked with at the various wave-lengths
studied. It was found that readings for the different solutions of the same
substance having the same concentration were, for a given wave-length,
different from one another to the extent of somewhat less than 2 per cent.
From this it seems fair to assume that the error in our work was not greater
than 2 per cent.

52 ABSORPTION OF LIGHT BY WATER CHANGED

DISCUSSION OF THE RESULTS.

An examination of the tables of data for potassium chloride, ammonium
chloride, and ammonium nitrate — that is, for those substances which, in
aqueous solutions, combined with very little water, as was demonstrated by
the freezing-point method, shows that for all wave-lengths studied the solu-
tion, and water of the same depth as the water in the solution, have prac-
tically the same transmission. The dissolved substance does not combine
with the solvent water, and the water in the solution has almost exactly the
same effect upon light as so much pure water would have. This is exactly
what would be expected from our knowledge of the absorption of light by
dissolved substances and by the solvent. When we began this work we
supposed, as others had done, that the water in the solution, whether it was
combined with the dissolved substance or not, would have the same power
to absorb light as so much pure solvent water. We shall now see that such
is not the case.

The results for the above-named substances were not plotted in the form
of curves, since the curve for water and for the solution would practically
coincide with one another, the dissolved substance having very little absorp-
tion over the region of wave-lengths studied in this investigation.

When we turn to the data in tables 8 and 9 very different relations mani-
fest themselves. These are the data for calcium chloride, magnesium chlo-
ride, and aluminium sulphate, that is, for salts which, in aqueous solution,
are strongly hydrated, as was shown by the earlier work in this laboratory.1
The solution in these cases is often more transparent than the same amount
of water that is contained in the solution.

That these relations may appear the more clearly, the results obtained for
the above-named salts are plotted as curves in figs. 12 to 17. Fig. 12 is the
curve for calcium chloride having a depth of 20 mm. This was obtained by
dividing the deflection produced by 21 mm. of the solution by that pro-
duced by 1 mm. of the solution. On the same sheet we have the curve for
water having a depth equal to that of the water in the calcium chloride.
This curve for water was also obtained by the "differential" method, i. e., by
dividing the deflections produced by the deeper solution by those obtained
with the more shallow solution, the difference in the depths of water in the
two cases being just equal to the depth of water in 20 mm. of the solution in
question. Fig. 13 is the curve for calcium chloride with a depth of layer of
10 mm. (11 — 1). The data from which the curve was plotted are contained
in table 9. The smaller depth of solution was used, so that the water-band
between 1.2/x and 1.3jU would come out more distinctly. The results for this
solution, like those for all the others, are compared with the absorption of
a depth of water equal to that of the water in the solution. The absorption
of the water, in this as in all other cases, was obtained by the "differential"
method.

1 Cam. Inst. Wash. Pub. 60.

IN THE PRESENCE OF STRONGLY HYDKATED SALTS. 53

Fig. 14 is the curve for magnesium chloride having a depth of 21 — 1 =
20 mm., and the corresponding water-curve. The data from which these
curves are plotted are given in table 8.

Fig. 15 is the curve for magnesium chloride having a depth of 1 cm., also
obtained by the "differential" method. These data are taken from table 9.

Fig. 16 is the curve for aluminium sulphate having a depth of 21 — 1 =
20 mm., and the corresponding absorption curve for water.

Fig. 17 is the curve for aluminium sulphate having a depth of 11 — 1 =
10 mm., and the corresponding water-curve.

Fig. 12 shows the relative absorption of water and of the solution of cal-
cium chloride having a concentration of 5,38 normal and a depth of 20 mm.
The corresponding water-curve is marked throughout by the symbol H20.
The solution is the more transparent from 0.9/* to nearly 1/*. The water then
becomes the more transparent over a short region of wave-lengths. From
1.05,1* to 1.2/* the solution is the more transparent. In this region the solu-
tion becomes as much as 25 per cent more transparent than the pure water,
as can be seen by comparing the points on the "water" curve with the corre-
sponding points on the curve for the solution which are vertically above the
points on the water-curve. The water becomes appreciably more trans-
parent only at and near the bottom of the "water-band" having a wave-
length of approximately 1/*. This is the effect that we would expect to get
if the dissolved substance exerted a "damping" effect on the absorption of
light by water.

It will be recalled that the salts which do not form hydrates show, in
aqueous solution, practically the same absorption as the corresponding
amount of water. It would, therefore, seem reasonable to account for the
differences in the case of nonhydrating and strongly hydrating salts as due
to the water of hydration, or the water that, in this case, is combined with
the calcium chloride.

The curves in fig. 13 are for a smaller depth of the same solution of cal-
cium chloride. This figure brings out the same general relations as was
shown in fig. 12. The water-curve in the region 1.25/* is above that of the
solution, showing that water in this region for the shallower depths of solu-
tion is more transparent than the solution. The additional feature brought
out by this figure is the water-band in the region 1.4 to 1.5/*-. After the first-
named water-band is passed the solution becomes more transparent than the
water and remains so until the wave-length 1.42 is reached. Here both the
solution and the water are practically opaque, as is shown by both the curves
approaching the abscissas.

The curve for magnesium chloride having a depth of 20 mm. is almost
exactly a duplicate of that for calcium chloride having the same depth.
Practically the only difference worthy of mention is in the region from 1.0/*
to 1.1/*. In the case of magnesium chloride the water remains the more
transparent over this region of wave-lengths. In the case of calcium chlo-
ride the solution is the more transparent over this region. The difference

54 ABSORPTION OF LIGHT BY WATER CHANGED

in the transparency of the water and the solution throughout this region is,
however, not very great. From 1.1/j towards the longer wave-lengths, as
we come down the descending arm of the curve towards the second water-
band, the water in the case of the magnesium chloride (as in the case of cal-
cium chloride) becomes much more opaque than the solution, the differences
here being of the same order of magnitude as those with calcium chloride.

Fig. 15 gives the results for magnesium chloride with a depth of layer of
1 cm., and the same relations hold as in fig. 14, for the relative transparency
of the water and of the solution. The water becomes the more transparent
from 1.22/z to 1.34/z. For the longer wave-lengths the solution becomes
more transparent until the region 1 .41/j is passed. For wave-lengths longer
than 1.41/* the transmission of both solution and water is practically zero —
that is, they both become opaque to the longer wave-lengths.

The results in fig. 16 bring out some new features of interest and impor-
tance. These are the results that were obtained with aluminium sulphate.
The new feature shown by the curve for aluminium sulphate, as compared
with those for calcium chloride and magnesium chloride, is that at the
minimum of the curve corresponding to wave-length 1/j the solution is
more transparent than the corresponding water. Beyond the wave-length
1.04ju the water becomes the more transparent with aluminium sulphate as
with magnesium chloride. Beyond the wave-length 1.17/z the solution
becomes more transparent in this case as with magnesium chloride and
calcium chloride.

If we turn to fig. 17 the relations are as follows. In the region of 1.2ju the
water is the more opaque. From 1 .29/* to 1 .36/j the water becomes the more
transparent. From 1.36// to the longest wave-length studied, the solution
again becomes more transparent than the corresponding layer of water.

An examination of all the results thus far obtained bearing on this prob-
lem leads us to conclude that the greater transparency of the solution as
compared with the water in the solution must be due to some action of the
dissolved substance on the solvent water. The question remains, what is
this action?

EXPLANATION OF THE RESULTS.

We have seen from our earlier work on the absorption spectra of solutions,
which has been in progress in this laboratory continuously for the past eight
years, that the solvent can have a marked effect on the power of the dis-
solved substance to absorb light. This was first shown by Jones and
Anderson,1 and a large number of examples of this effect have since been
found by Jones and Strong.2 We interpreted the effect of the solvent on the
power of the dissolved substance to absorb light as due to a combination
between a part of the liquid present and the dissolved substance. This
enabled us to explain a large number of facts which were brought to light for
the first time by our investigations of the absorption spectra of solutions.
Many of the phenomena which were thus explained, it seemed, could not be

1 Cam. Inst. Wash. Pub. 110. 2 Cam. Inst. Wash. Pubs. 130 and 160.

IN THE PRESENCE OF STRONGLY HYDRATED SALTS. 55

explained in terms of any other suggestion that has thus far been made.
In a word, the solvate theory of solution as proposed by Jones about a dozen
years ago,1 to supplement the theory of electrolytic dissociation in order that
we might have a theory of the real solutions which we use in the laboratory,
and not simply a theory of ideal solutions as the theory of electrolytic dis-
sociation alone must be regarded, has served good purpose in explaining the
phenomena that have been previously observed in connection with the
absorption of light by solutions of dissolved substances.

We are inclined to explain the phenomena recorded in this paper by means
of the same theory. For solutions of those substances which have been
shown by entirely different methods not to hydrate to any appreciable
extent, the absorption of light by the solution and by a layer of water equal
in depth to that of the water in the solution, is the same almost to within the
limit of experimental error.

For those substances which have been shown to form complex hydrates,
however, the absorption of light by their solutions and by a layer of water
equal in depth to that of the water in the solution is very different. The
water in these solutions is usually more opaque to light than the solution —
or, in other words, a solution is more transparent than the water that is
present in the solution.

The most rational explanation of this phenomenon appears to be that the
part of the water that is combined with the dissolved substance has a smaller
power to absorb light than pure, free, uncombined water. The fact that
we are able to detect the difference between the water in the solution and
pure water, by its action on light, we regard as good evidence that water in
the solution is different from pure, free water. This difference, it seems to
us, can be readily accounted for by the theory that a part of the water
present in the solution is in combination with the dissolved substance.

We have carried out similar investigations with aluminium nitrate, but the
concentration of the strongest solution that could be obtained was not suffi-
ciently great to show the phenomenon in question. We therefore do not
incorporate the results obtained with this substance. That the solutions
must be very concentrated to show clearly the phenomenon with which we
are dealing is seen from the results given in table 10. Here the solutions of
the three salts in question that were used are more dilute than those for
which the results are tabulated in tables 8 and 9. An examination of table
10 will show that the phenomenon in question does not manifest itself to
anything like the same extent as with the more concentrated solutions.
This is exactly what we would expect in terms of the solvate theory of solu-
tions. The more concentrated the solution the larger the total amount of
the water present combined with the dissolved substance. If combination
between water and the dissolved substance explains the facts recorded in this
paper, then the larger the amount of water present that is combined with
the dissolved substance the more pronounced the phenomenon in question.

1 Amer. Chem. Journ., 23, 89 (1900).

56 ABSORPTION OF LIGHT BY WATER CHANGED

The results obtained with aluminium sulphate bring out the same facts
shown by calcium chloride and magnesium chloride, and also that water is
more transparent in the region l.lju and more opaque at lju. That the sul-
phate should not agree throughout with the chlorides is really not surprising,
since the sulphates show abnormal results in almost every particular. This
is probably due, in part at least, to the large amount of polymerization
which the sulphate molecules in general undergo in the presence of even
water as a solvent. It should also be remembered in the present connection
that while calcium chloride and magnesium chloride crystallize with only 6
molecules of water, and are therefore only largely hydrated, aluminium sul-
phate crystallizes with 18 molecules of water and is therefore very largely
hydrated.

The results in table 11 are the radiomicrometer deflections for a solution
of aluminium sulphate and those for water having the same depth as the
water in the solution in question, and the corresponding data for potassium
chloride. A comparison of the two columns for potassium chloride and its
corresponding water shows that the two are almost equally transparent to
all the wave-lengths studied.

A comparison of the aluminium sulphate with its corresponding water
brings out the phenomenon that we are now discussing in a very pronounced
manner.

One other relation of a general character should be pointed out. The
curves (figs. 12 to 17) show that the addition of salt to water shifts the
absorption towards the longer wave-lengths. This is analogous to what had
already been found by Jones and Uhler,1 Jones and Anderson,2 Jones and

80^ H20

70

60

50

40

30 i

201

CaCI2,5.38N.
Depth 2 cm.

0.9 1.0 11 12

Fig. 12.

' Cam. Inst. Wash. Pub. 60. 2Carn. Inst. Wash. Pub. 110.

IN THE PRESENCE OF STRONGLY HYDRATED SALTS. 57

Strong,1 and Guy and Jones,2 when the absorption of salts as affected by the
water present was studied. It was found that rise in temperature and
increase in the concentration of the solution both tended to shift the ab-
sorption of the salt towards the longer wave-lengths. The effect of rise in
temperature and the increase in the concentration of the solution tended to
simplify the hydrates in combination with the particles of the salt. The
resonator within this simplified system seems to vibrate so as to shift the
absorption bands towards the red.

The effect of the salt on the absorption of the water is the same as that of
rise of temperature and increase of concentration on the absorption of the
dissolved substance. We would naturally look for a similar explanation of
the two sets of phenomena. It has been suggested by Dr. Guy, that the
effect of the salt on the absorption of light by water may be due to the
breaking down of the associated molecules of water by the dissolved sub-
stance. This would be in keeping with the fact established by Jones and
Murray,3 that one associated substance when dissolved in another associated
substance diminishes its association.

In terms of this explanation, however, it is a little difficult to see why non-
hydrated salts, such as were used in this work, do not also diminish the asso-
ciation of water and cause a shifting of its absorption bands towards the
longer wave-lengths. It may be that the effect of the dissolved substance
in breaking down the association of the water is pronounced only in the case
of water of hydration or the water that is combined with the dissolved sub-
stance, and that the explanation offered above is fundamentally correct.

70-

60--

50-

40-

30

20

10

CaCI2,5.38N
Depth 1cm.

1.15 12 1.3 1.4 1.5

1 Cam. Inst. Wash. Pubs. 130 and 160. 3 Amer. Chem. Journ., 30, 193 (1903).

2 Amcr. Chem. Journ., 49, 1 (1913).

58

ABSORPTION OF LIGHT BY WATER CHANGED

H;0

70 h

60

50-

40

30

20

MgCI2,4.96N
Depth 2cm.

0.9

1.0

Fig. 14.

1.2

80

70

60-

50-

40

30-

20

10

H20

MgCI2,4.96N
Depth 1cm.

1.15

1.2

1.3
Fig. 15,

IN THE PRESENCE OF STRONGLY HYDRATED SALTS.

59

SO

70-

60

50-

40 -

30-

20

10

hUO

AI2(S04>3 ,1.017 N

Depth 2cm.

0.9

1.0

Fig. 16.

1.2

20-

10- ■

1.15

AI2(S04>3, 1.017 N.

Oepthlcm.

1.2

H?0

1.5

Fig. 17,

CHAPTER VI.

ABSORPTION SPECTRA OF A NUMBER OF SALTS AS MEASURED
BY MEANS OF THE RADIOMICROMETER.

The results tabulated and discussed in Chapters IV and V, which are con-
cerned with the energy measurements of the absorption spectra of solutions
by means of the radiomicrometer, were made by comparing the intensity of a
given source of light (after passing through the solution) with the intensity
of the same source of light after passing through an equal depth of water.
In a word, the depths of cells in each case were the same. As has already
been stated, a cell whose depth was 1 cm. was filled with the solution and
placed in the path of the beam of light and the deflection of the instrument
noted; then a cell of the same depth was filled with the solvent and interposed
in exactly the same position as the former cell, and the deflection of the
instrument again noted. Denoting the former by I and the latter by 70 we
get the ratio I/h, which represents the percentage transmission of the solu-
tion as compared with water. Such a procedure was repeated at frequent
intervals throughout the spectrum, locating a series of points through
which the transmission curves could be drawn.

Certain phenomena presented themselves throughout the course of this
investigation, which suggested a more careful study of some of the absorp-
tion bands located in the infra-red portion of the spectrum ; and at the same
time it was thought advisable to map the absorption spectra of some of the
more common salts of cobalt, nickel, etc., in terms of Beer's law; since up to
the time of this investigation no satisfactory quantitative study of the
infra-red spectrum of these salts had appeared.

In order to make a careful study of the exact intensity of the various por-
tions of any given bands, it is clear that we are dealing with a much more
complex and intricate problem than simply with the location of the band;
and on this account it was necessary to improve our apparatus and at the
same time to exert more care, if possible, in carrying out any given operation.

It was early found that if we desired to study that region of the infra-red
spectrum in which water had considerable absorption, we must not compare
our solutions with an equal depth of layer of water, as noted above ; but with
a depth of layer equal to the water in the solution, which in the most con-
centrated solutions was much less than the actual depth of the cell containing
the solution — a part of the cell's depth being occupied by the dissolved sub-
stance. Even when such a correction was made, it was found that for a
given wave-length, in the water absorption bands, the solution gave greater
deflections than did the solvent, i. e., that in such regions the solution was
actually more transparent than an equivalent depth of water.

61

62 ABSORPTION SPECTRA OF A NUMBER OF SALTS

Remembering that the solutions with which we were then working, i. e.,
solutions of salts of neodymium and praseodymium, were strongly hydrated,
it was thought that in view of the fact that at least a part, and in the con-
centrated solutions a considerable part of the water present was there as
water of hydration, it would be advisable to study the effect of colorless
hydrated salts upon the absorption of water.

This chapter of our work has been sufficiently discussed elsewhere in this
monograph, and will be taken up here only to state that these experiments
showed clearly that there were many variables to be considered. We have,
first, the effect of the solvent on the absorption of the solute; and, secondly,
the effect of the solute upon the absorption of the solvent. In addition to
these, there was, of course, the absorption of the solvent and the solute inde-
pendently. Such being the case, we would not be obtaining comparable
results for various dilutions of any solutions in terms of Beer's law, even if
we did compare each dilution with an equivalent amount of water. It is
clear that by so doing we would not be getting comparable ratios, since the
solvent and the solute were mutually affecting each other's absorption ; and
this effect would not be the same for the different dilutions of the same salt.

MODE OF PROCEDURE.

It is, however, possible to get the exact transmission of a given depth of
solution by a method of differentiation. If we placed in cell A 11 mm. of a
solution and in cell B 1 mm. of the same solution, the ratio representing the
respective deflections of the instrument, when these cells are alternately
placed in the path of the beam of light, should give the absorption or trans-
mission of (11 — 1) or 10 mm. of the solution.

Since, if we let A be the percentage absorption of a unit's depth of layer of
the solution, and 70 the initial intensity of the light impinging upon the sur-
face, we get

AIq — amount of light absorbed by first unit layer of the solution.
Then,

I0 — I0A = io(l — A) = light incident upon surface of second unit layer.
Denoting this by 7i, we get

71 = 7o-/oA=70(l-A)or^ = l-^

Considering again the third unit layer, we get, by similar reasoning,

Ii—IiA = amount of light incident upon its surface.

Denoting this by 72, we get

h=I1-IlA=I1(l-A)

but71 = 70(l-/l); therefore, 72 = 70(1-A)2; hence 7„ = I0(1-A)n. We can
then, by this process, obtain transmissions for given depths of solution
and for varying concentrations. This was the method adopted throughout
this chapter of the work.

AS MEASURED BY MEANS OF THE RADIOMICROMETER. 63

DESCRIPTION OF CELLS USED.

In all cases where we were dealing with different depths of layer, it was
necessary to use cells adjustable in length. A very satisfactory form of cell
was devised and used throughout the latter part of this work. It consisted
essentially of two brass cylinders telescoping neatly into each other. The
external diameter of the outside cylinder was about 2| inches, and the thick-
ness of the walls was in every case about 2 mm., which was sufficient to with-
stand handling without danger of changing the shape of the cell. Into the
ends of each cylinder there was sealed, by means of Wood's metal, a glass
plate about 1 mm. thick, made of the very best optical glass. In all cases
the glass plates were so nearly parallel as to show interference fringes ; and
both cells gave the same deflections, either when empty or filled with the
same solution and placed in the path of the light before the radiomicrometer.

After adjusting the glass ends and fixing them securely by means of Wood's
metal, the entire cell was first plated with silver, being taken out of the
plating-bath from time to time and polished to a bright surface with the
finest crocus paper. On top of this silver coating a heavy plating of gold
was deposited. The distance between the glass plates fastened to the ends
of the telescoping cylinders, which determined the depth of layer of solution
used, was in all cases fixed by gold-plated washers, whose thickness had been
accurately measured to 0.001 inch by means of a vernier caliper.

Before any series of readings was made, the positions of the two cells was
so adjusted in the sliding carriage as to give equal deflections, when alter-
nately placed in the same position before the radiomicrometer, in that part
of the spectrum where neither the solute nor solvent had any absorption;
and from time to time throughout the experiment duplicate readings were
made on this point to see that the cells had not changed their relative
positions. In case any change was noted, a duplicate series of readings was
always made. Such readings upon the same cell usually agreed to about
one division of the scale, which corresponded to about 1 to 2 per cent,
depending upon the throw of the instrument. In the midst of the very
intense absorption bands,where the deflections of the instrument were small,
reaching zero at many points, the error resulting from any drift in the instru-
ment or reading of the scale was greater than the mean error given above.

In nearly all cases new solutions were made up and the results duplicated,
so that the tables and curves below represent a mean of several series of
readings. In most cases the agreement was very satisfactory, usually the
difference not being over 3 per cent.

Since any change in the position of the prism was a determining factor in
the portion of the spectrum which fell upon the thermo-j unction, and since
in the very intense, sharp bands of the neodymium salts any slight change in
the position of the prism would make a great difference in the final results,
great care had to be exerted in setting the head reading of the spectroscope.
Such difficulties were not met with in solutions where the absorption bands
were broad and diffuse, as in salts of cobalt, nickel, etc.

64 ABSORPTION SPECTRA OP A NUMBER OF SALTS

In studying the changes which might occur in any band, it is of course
necessary that all conditions be as nearly as possible the same. One of the
most important factors here is that of the width of the slits of the spectro-
scope. With those solutions whose absorption bands are broad and diffuse,
not having such well-defined edges as with the salts of neodymium and
praseodymium, this is not such a determining factor. Should the band be
very narrow — say approaching that of the width of the slit necessary to be
used in order to secure reasonable deflections of the instrument — it is seen
that any slight change in the slit will make a large difference in the amount
of light falling on the thermal-junction.

Considering a concrete example, let us suppose that the slit-width is just
equal to that of the absorption band, under a given dispersion. If, now, the
band and the slit exactly coincide, it is evident that no light will be falling
upon the junction, this being indicated by zero deflection of the instrument.
If, on the other hand, the slit is slightly wider than the band, some light will
enter around the edges of the band ; and, though the narrow band may act-
ually have complete absorption at a given point, it would not be indicated by
the instrument, since some light is entering around the edges of the band.

Denoting the deflection of the instrument for a cell of 2 mm. depth of a
solution of x concentration by A, and the same for 1 mm. of the same solution
by B, we get, by the differential method discussed above, the ratio A/B for
the intensity of the light transmitted by (2 — 1) or 1 mm. of the solution in
question.

By a similar reasoning we get the ratio A'/B' for the value of the trans-
it
mission of a solution of concentration — , using absorbing layers 21 mm. and

1 mm., respectively. While such a method is theoretically and mathe-
matically correct for infinitely narrow slit-widths, and practically so for
bands which are wide in comparison with the necessary slit-widths, yet in
the case of very sharp, narrow neodymium bands it has been found not to
give comparable results. The reason for this is clearly seen in the light of
the facts discussed above.

Let us consider the ratios A/B and A'/B'. In the first case we are deal-
ing with concentrated solutions, where the absorption bands are broad;
hence B is small, and, in case the slit- width is comparable with the width of
the absorption band, B will be very much smaller than B', since B' is only

x
1 mm. of an — concentration solution. In a word, B, which is 1 mm. of the

more concentrated solution, has 20 times the number of absorbers as has an

equal depth represented by B' , and a decrease in the denominator of the

fraction means an increase in its value.

While the ratios A/B and A'/B' give the transmissions for 1 mm. of a

x
solution of concentration x, and 20 mm. of a solution of concentration —

respectively, provided the slits are narrow; yet in the visible part of the

AS MEASURED BY MEANS OF THE RADIOMICROMETER. 65

spectrum, where such wide slits had to be used on account of the small
amounts of energy in this region, these ratios are not comparable.

For this reason we have confined the larger part of our work on neo-
dymium compounds almost entirely to wave-lengths greater than0.7^. In
all the following tables and curves representing these data, constant slit-
widths of 0.2 mm. have been used. This was the minimum width which
could be employed, in order to get reasonable deflections throughout the
spectrum from 0.7// to 1/x. Experiments have shown that any error result-
ing from slit-widths would not amount to more than 3 or 4 per cent through-
out this region.

The source of light was, as in the previous chapters, a Nernst glower carry-
ing about 1.2 amperes, and the current kept constant by means of an adjust-
able slide-wire resistance. The source of current was a large number of
storage cells, and this was never allowed to vary over 0.01 ampere. Great
care was exerted in keeping the current constant while obtaining a single
ratio, since this is really the only time in which a slight change in current
density was dangerous.

DISCUSSION OF TABLES AND CURVES.
NEODYMIUM CHLORIDE IN WATER.

Table 12 gives the observed transmissions of solutions of neodymium
chloride in water. In all the tables the following four dilutions have been
studied, the depths of cell being, generally, 2.5, 5, 10, and 20 mm., respec-
tively; and the concentrations being made so as to keep nXd = k. In
column 1 of each table there is given X, taken at such intervals as the solu-
tion in question required. In those portions of the spectrum where the
transmission was complete, or very nearly so, these intervals were greater
than in those regions where there were absorption bands.

Reading from left to right in this table, beginning with column 2, there are
given the absorptions for solutions of the following concentrations: 2.141,
1.07, 0.535, and 0.267 normal, respectively; the corresponding depths of
absorbing layer being 2.5, 5, 10, and 20 mm., respectively. In every case

x-\-\
the transmission was obtained from the ratio — - — , where x is 2.5, 5, 10,

and 20 mm., respectively.

In all cases the concentrated mother-solution was carefully made up, its
concentration determined by a gravimetric precipitation of the metal, and the
succeeding solutions made by diluting measured parts of the mother-solution.

Observations are given here over only that portion of the infra-red spec-
trum from X6800 to X10000. It is in this region that the most pronounced
neodymium bands occur. It was thought advisable not to go further into
the infra-red, since bej'ond lfx the general absorption due to the solvent is
very marked. This would, of course, interfere with a quantitative study of
any band occurring in this region, since it is impossible to separate the two
absorptions, previous work having shown that they are not additive.

66

ABSORPTION SPECTRA OF A NUMBER OF SALTS

The work in the visible region of the spectrum was limited by the slit-
widths necessary to be used, which has been mentioned and discussed above.
We have rather chosen a limited portion of the infra-red, over which we
could work without altering either the intensity of the light or the slit-width,
which was in all cases 0.2 mm.

The curves representing table 12 are given in figs. 18 to 21 inclusive.
The percentages of transmission are plotted as ordinates, while the wave-
lengths are given as abscissse. An examination of these curves shows three
pronounced minima, representing the three absorption bands, with their
centers near X7300,1 X7950, and X8700, and less-marked bands near X7150
and X9000. The latter of these small bands is possibly due in part to the

Table 12. — Percentage Ti

ansmission of Neodymium Chloride Solutions.

X

D.=2.5mm.

D. = 5mm.

D. = 10mm.

D.=20mm.

X

D.=2.5mm.

D. = 5mm.

D. = 10mm.

D.=20mm.

C.=2.141N.

C.=1.071.

C. =0.535.

C. =0.207.

C.=2.141N.

C. = 1.071.

C.=0.535.

C. =0.267.

686

93

88

88

86

800

15

10

8

6

693

95

95

95

94

805

24

23

22

18

698

96

96

96

94

809

40

39

38

37

704

96

96

98

94

814

58

58

58

53

708

96

93

95

92

819

80

78

78

76

712

92

93

95

88

825

89

91

91

82

716

88

89

88

85

830

93

92

93

88

720

81

78

81

81

834

94

95

95

88

723

64

62 63

56

839

93

93

93

87

726

32

31

25

23

845

91

92

91

86

730

7

7

5

6

850

87

86

87

78

733

0

0

0

0

856

75

73

71

66

737

0

0

0

0

861

54

46

43

40

741

5

2

1

2

867

29

21

18

16

746

18

14

12

5

872

28

24

23

18

751

36

39 29

28

877

40

39

40

34

755

54

55 52

49

882

53

48

52

47

759

75

74 72

68

888

00

59

60

53

763

85

85

83

81

894

(il

59

01

56

767

84

86

85

83

900

67

66

66

59

770

79

81

80

76

906

80

78

79

71

774

67

73

65

61

912

90

92

89

79

779

47

45 45

40

917

96

93

94

86

783

29

27 26

24

923

98

97

96

86

787

22

10 10

7

928

98

96

96

84

791

0

0 0

0

933

98

96

96

81

796

0

0 0

0

938

i

98

94

90

75

absorption of the solvent; but since its intensity does not increase markedly
with dilution, it is more probably a doublet. Considering the curves repre-
senting all four dilutions, we see that the X7300 and X7900 bands show com-
plete absorption over a considerable range of wave-lengths, and any change
in intensity could not be very noticeable. The X8700 band, however, has
its minima gradually lowered as we pass from curve 18 to curve 21, i, e., in
the direction of increasing dilution. This phenomenon has been noted else-
where in this monograph, and a possible explanation of it based upon a
theory of resonance suggested. A closer and more exact study has shown
that, although the phenomenon is a real one, yet it is probable that it may

1 The wave-lengths in the above and following tables are given, in general, to only
three places.

AS MEASURED IJY MEANS OF THE RADIOMICROMETER.

67

in part be due to the combined effect of the slight water absorption and,
even a more important factor, the slit-widths, as discussed above.

The regions of maximum transmission occur near X7600 and X8400, and
solutions of neodymium chloride become almost completely transparent
beyond 1/jl, except for the general absorption of the solvent. Slight absorp-
tion bands occur in this region, one near 1.5/x, but they are so masked by the
intense water absorption that it was found impossible to make a quanti-
tative study of them.

75

2
O
1/5
i/}

5
i/)

2

UJ

a

u
<r

LJ
Q.

50

25

Neodymium Chloride

Cell Depth 2.5mm.
Concentration 2.141 N.

0.65//

0.7a

0.75,7

Fig.

0.8^
18.

0.85^

0.9//

\.M

100

75

2

o

50

u

a:

UJ

25

0 65//

Neodymium Chloride

Cell Depth 5mm.
Concentration

071 N.

0.8//
Fig. 19.

0 95//

From a comprehensive study of the four curves representing the absorp-
tion spectra of concentrated solutions of neodymium chloride, it seems
probable that Beer's law holds quantitatively for the infra-red region, except
for such slight changes as have been fully discussed above.

68

ABSORPTION SPECTRA OF A NUMBER OF SALTS

Since, as mentioned above, the absorption bands with their centers near
X7300 and X7950, in solutions of such concentrations as are given in table 12,
reach complete absorption, and as in such cases it would not be easy to

100

z
Q 75

5

UJ

53 50

CL

25

0.65^

Neodymiurn Chloride

Cell Depth 10mm.
Concentration 0.535N.

0.75a 0.8m

Fig. 20.

0.85m

0.9m

0.95^

100

y 75

<

z

UJ

o

IT

50

25

0 65m

Neodymiurn Chloride

Cell Depth 20mm.
Concentration 0.2G7N.

0.7m

0.75m 0.8m

Fig. 21.

0.85m

0.9m

0.95m

detect any change in their intensity, it was thought advisable to make a
study of a series of more dilute solutions of the same salt.

The results of this experiment are given in table 13. Reading from left to
right, the concentrations are 0.536, 0.267, 0.133, and 0.067 normal, respec-

AS MEASURED BY MEANS OF THE KADIOMICROMETER.

69

tively, the corresponding depths of absorbing layers being 2.5, 5, 10, and
20 mm., respectively. Here again it is seen that the conditions of Beer's
law are adhered to. The results are graphically represented in figs. 22 to 25,
inclusive. These curves show minima in about the same positions as did the

100

z
o

l/>
I 75

LU

o

a 50

25

Neodymium Chloride

Cell Depth 2.5mm.
Concentration 0.536 N.

065//

0.7//

0.75// 0.8//

Fig. 22.

085//

09//

0.95//

100

75-

2
O

50

Id

LU

a.

25

Neodymium Chloride

Cell Depth 5mm.
Concentration 0 267N

0.65//

0.7//

0.75// 0.8//

Fig. 23.

0.85//

0.9m

0.95//

curves representing the more concentrated solutions, but since the solutions
are more dilute, they are accordingly more transparent; hence the minima
in the curves are not so pronounced. The maximal absorption occurring
near X7300 and X7900 are at about 25 per cent.

70

ABSORPTION SPECTRA OF A NUMBER OF SALTS

It will be noticed that beyond 0.9ju all of the curves drop sharply with
dilution, which is due entirely to the increasing absorption of the water.

Figs. 22 to 25, inclusive, show just what might be anticipated from
figs. 18 to 21, a lowering of absorption maxima as we pass towards the
more dilute solutions. This change is most marked in the X8700 band, and

100

75

z
o

50

UJ

O

<

z

UJ

u

a:

UJ

CL

25

Neodymium Chloride

Cell Depth 10 mm.
Concentration 0I33N

0.65/u

0.7//

0.75/; 0.8//

Fig. 24.

0.85//

0.9/;

0.95/j

100

75

o

us

s

SO

UJ

u
tr

LU

?c>

Neodymium Chlondo

Cell Depth 20 mm.
Concentration 0.067 N

0.65//

0.7//

0.75/<

Fig,

08,(/

25.

0.85//

0.9//

0.95//

it is in this region that the absorption of the water is most pronounced,
although a 20 mm. layer of water in this region has at no point over 10 per
cent absorption. The change in the intensity of the absorption band is
greater than this amount, but it seems probable that this, together with the
added correction for the slit-widths, may account for the phenomenon, and
that Beer's law holds for the dilute solutions of neodymium chloride.

AS MEASURED BY MEANS OF THE RADIOMICROMETER. 71

Table 13. — Transmission of Neodymium Chloride Solutions (Dilute).

X

D.=2.5mm.

D. =5iiini.

D.=10mm.

D.=20mm

X
800

D. =2.5inni.

D.=5mm.

D. = 10mm.

D. = 20nnn.

C.=0.536N.

C.=0.207N.

C.=0.133N.

C. = 0.0G7N.

C. = 0.53(iN.

45

C.=0.267N.

C.=0.133N.

C.=0.067N.

686

92

95

94

92

39

38

39

693

93

95

95

97

805

64

59

57

57

698

97

95

95

97

809

78

72

71

70

704

98

95

96

97

814

85

84

82

82

708

98

95

97

98

819

92

91

88

88

712

100

98

96

96

825

100

94

94

94

716

100

98

98

92

830

100

96

94

93

720

96

94

89

89

834

100

94

94

92

723

85

84

83

84

839

100

96

93

91

726

59

54

58

53

845

98

95

94

91

730

42

41

40

34

850

97

94

90

89

733

31

29

26

25

856

92

87

84

83

737

25

24

20

22

861

78

77

74

70

741

28

29

29

29

867

70

64

62

60

746

47

44

51

44

872

64

60

59

56

751

69

65

65

60

877

74

71

68

66

755

85

82

82

78

882

84

80

76

75

759

93

86

90

86

888

89

84

82

78

763

97

90

92

92

894

90

84

83

79

767

97

95

90

92

900

91

88

86

81

770

97

93

90

89

906

95

90

87

83

774

91

86

87

84

912

98

95

90

84

779

79

76

78

74

917

100

96

92

85

783

69

61

59

56

923

100

96

93

84

787

47

39

38

34

928

100

95

92

84

791

23

24

20

18

933

100

94

94

83

796

28

25

23

23

938

100

94

95

78

Table 14. — Transmission of Neodymium Nitrate Solutions.

X

D. = 2.5mm.

D.=5mm.

D. = 10mm.

D.=20mm.

1
X

D.=2.5mm.

D. = 5mm.

D. = 10mm.

D.=20mm.

C.=2.010N.

C. = 1.05 N.

C.=0.525N.

C.=0.262N.

C.=2.010N.

C. = 1.05 N.

C.=0.525N.

C.=0.262N.

686

86

88

85

78

800

12

10

6

6

693

93

96

91

89

805

21

19

14

15

698

93

96

93

93

809

35

33

30

30

704

96

94

93

91

814

49

49

47

47

708

95

96

96

88

819

63

67

65

64

712

89

90

94

89

825

79

80

79

77

716

84

88

88

82

830

82

88

89

82

720

78

79

78

78

834

92

92

93

85

723

80

63

61

58

839

96

94

94

89

726

40

33

30

28

845

92

91

91

85

730

14

14

10

9

850

87

84

85

77

733

0

0

0

0

856

76

70

72

65

737

0

0

0

0

861

61

57

53

47

741

11

4

0

0

867

43

36

35

32

746

12

15

6

6

872

30

27

26

22

751

22

23

22

17

877

31

30

28

25

755

38

39

40

38

882

43

42

41

36

759

60

60

62

55

888

55

52

51

48

763

75

75

74

70

894

65

60

59

61

767

80

80

83

79

900

68

66

66

62

770

75

75

79

75

906

77

75

75

67

774

64

66

68

62

912

86

84

83

75

779

47

47

48

44

917

94

90

97

83

783

32

32

29

26

923

98

92

90

84

787

16

10

13

9

933

100

96

88

86

791

11

8

2

5

938

100

96

83

77

796

0

0

0

0

72

ABSORPTION SPECTRA OF A NUMBER OF SALTS
NEODYMIUM NITRATE.

Table 14 gives the percentage transmission for solutions of neodymium
nitrate. Column 1 gives the respective wave-lengths at such intervals as
the solutions required. Reading from left to right, we find the percentage
transmissions for the following concentrations: 2.010, 1.050, 0.525, and
0.262 normal, respectively, the corresponding depths of absorbing layers

100

75

z
o

10
!2
2

in

UJ

§

Z
UJ

o

cr

50

25

0.65//

100

75

z
o

to

2
i/>

z

s

t-

o

«.
I—

z

UJ

50

25

0.65//

Neodymium Nitrate

Cell Depth 2.5mm.
Concentration 2 01 N

0.7A

0.75// 0.8//

Fig. 26.

0.85//

0.9//

0.95a

Neodymium Nitrate

Cell Depth 5 mm.
Concentration 1.05 N

0.75// 0.8a

Fig. 27.

0.95//

being 2.5, 5, 10, and 20 mm. Figs. 26 to 29, inclusive, represent these
results, the abscissae being wave-lengths and the percentage transmissions
being given as ordinates. It is seen that the absorption bands in the nitrate
solutions, as with those of the chloride discussed above, show three minima
at X7300. X7950, and X8750. The nitrate bands are not as intense as those

AS MEASURED BY MEANS OF THE RADIOMICROMETEK.

73

of the concentrated solutions of the chloride given in figs. 18 to 21. This
is what we should expect, since the concentrations of the nitrate solutions
are not so great. However, two of the absorption bands reach zero trans-
mission. A comparative study of any of these bands in the succeeding
curves shows that, just as was found with the chloride bands, they become

100

7:.

z
o

z
<
cr

y~

UJ
U
<

z

UJ

u

tr.

UJ

Q.

■')

25

0.65a

Neodymium Nitrate

Cell Depth 10 mm.
Concentration 0.525 N

0.75/j. 0.8a

Fig. 28.

0.85/z

0.95a

too

75

uj 50
o

o

UJ

Ql

25

0.65a

Neodymium Nitrate

Cell Depth 20 mm.
Concentration 0.262 N

0.75/^ 0.8^

Fig. 29.

0.95a

more intense with dilution. The decided decrease in the transmission in the
regions of the spectrum beyond 1/u is undoubtedly due to the increasing
absorption of water as the solution becomes more dilute. The other slight
deviations from Beer's law are not greater than could be accounted for by
the corrections mentioned under the discussion of the chloride curves.

74

ABSORPTION SPECTRA OF A NUMBER OF SALTS
NEODYMIUM ACETATE.

Table 15 gives the results obtained for solutions of neodymium acetate in
water. The concentrations, reading from left to right, were 0.(517, 0.308,
0.154, and 0.077 normal, respectively, the corresponding depths of absorbing

100

o

m

75

UJ

a

| 50

UJ

Q.

25

0.65/;

Neodymium Acetate

Cell Depth 2.5 mm.
Concentration 0.6I7N

0.7a

0.75/7

0.8/7

Fig. 30.

0.85/7

0.9/7

0.95/7

100

75

o

50

UJ

o

UJ

0.

25

0.65/z

0.7/7

0.75//

Fig. 31.

Neodymium Acetate

Cell Depth 5 mm
Concentration 0308N

0.85,"

0.9/7

0.95/7

layer being 2.5, 5, 10, and 20 mm., respectively. The results in this table
are plotted in figs. 30 to 33, inclusive. The percentage transmission and
wave-lengths are represented, respectively, by the ordinates and abscissae of
the curves.

AS MEASURED BY MEANS OF THE RADIOMICROMETER.

75

The minima of transmission fall at approximately tho same positions as
with the chloride and nitrate solutions discussed above, i. e., at X7300, X7950,
and X8750. As indicated by the photographic method, the solutions of neo-
dymium acetate have greater absorbing powers for a given concentration
than either the chloride or nitrate.

100 r

75

z
o
in

50

o

UJ

(J

UJ

a.

25

Neodymiurn Acetate

Cell Depth 10 mm
Concentration 0.I54N

0.65a

0.7//

0.75a

0.8//
Fig. 32.

0.85a

0.9a

0.95a

100

75

o

ijj

UJ

u
or

0_

50-

25

Neodymiurn Acetate

Cell Depth 20mm.
Concentration 0.077 N

0.65a

0.7a

0.75// 0.8a

Fig. 33.

0.85//

0.9//

0.95a

The small band near X7000 appears slightly more intense with the acetate
than with equal concentrations of the other salts; and the more intense bands
X7300, X7950, and X8750 show the same general tendency to have their
minima lowered with increasing dilution. This set of curves shows in a
marked way the rapid increase in the absorption near X9500, which is due to
the water present in the solution and illustrates the difficulty that is met

76

ABSOKPTION SPECTRA OF A NUMBER OF SALTS

with when working with aqueous solutions at greater wave-lengths than l/x.
Even over the range of wave-lengths at which we have worked, it is seen
that the absorption due to water is a disturbing factor.

Table 15. — Trans?nissions of Neodymium Acetate Solutions.

X

D.=2.5mm.

D. = 5 mm.

D. = 10mm.

D. = 20mm. '

X

D. = 2.5mm.

D. = 5 mm.

D. = 10mm.

D.=20mm.

C.=0.617N.

C.=0.308N.

C. = 0.154N.

C.=0.077NJ

C.=0.617N.

C.=0.308N.

C.=0.154N.

C.=0.077N.

686

98

97

94

94

soo

27

28

27

28

693

92

91

93

95

805

57

46

44

46

698

97

97

96

93

809

69

63

62

64

704

98

100

92

93

814

74

76

73

75

708

96

98

92

96

819

90

83

83

82

712

100

96

94

96

825

92

89

86

87

716

100

94

90

94

830

92

93

92

91

720

94

90

90

94

834

98

98

92

91

723

87

87

79

76

839

96

97

94

90

726

69

64

60

57

845

97

97

90

88

730

44

36

32

30

850

95

96

90

87

733

22

22

18

21

856

95

94

84

83

737

24

17

14

15

861

91

83

77

71

741

30

23

22

25

867

72

69

65

67

746

41

39

34

35

872

62

60

56

53

751

52

55

52

55

877

66

67

58

57

755

71

71

65

67

882

71

72

66

66

759

81

84

81

82

888

83

S2

76

75

763

90

91

87

88

894

87

84

81

78

767

93

93

89

87

900

91

88

84

79

770

94

90

87

89

906

92

90

87

83

774

89

88

83

84

912

94

90

90

82

779

72

74

74

75

917

96

93

89

85

783

62

63

59

55

923

97

96

94

86

787

45

40

37

32

933

99

95

91

83

791

24

21

17

18

938

99

95

92

77

796

22

17

17

14

PRASEODYMIUM CHLORIDE.

Solutions of praseodymium salts are not of great interest from our stand-
point, in those regions beyond the visible part of the spectrum. It was
found that such solutions were transparent in the infra-red end of the spec-
trum as far as 1.5, except two very weak bands which fall just in the midst of
the intense water-bands. Since, at this point, a very thin layer of water is
almost completely opaque, it is evident that it would be impossible to study
aqueous solutions in this region, especially dilute solutions.

As shown by the photographic plate, praseodymium salts possess two
groups of bands in the visible spectrum, one in the green near X4600 and
another near X5900. Since the amount of energy at the former wave-length
is so very small, the width of slits necessary to be used was too large to give
satisfactory results. Such bands could, of course, be detected, but the
deflections of the instrument at this part of the spectrum are very small,
and, hence, relatively large errors would occur in making the readings.

For these reasons we have confined our attention to a careful study of the
one band which has its center near X5900. Table 16 gives the observed
transmissions for the four dilutions of solutions of praseodymium chloride.

AS MEASURED BY MEANS OF THE RADIOMICROMETER.

77

In column 1 there is given X taken at such intervals as the graduated head of
the spectroscope would permit. Reading from left to right, the respective
concentrations were 1.377, 0.688, 0.344, and 0.177 normal, the corresponding
depths of absorbing layer being 2.5, 5, 10, and 20 mm., respectively.

Table 16. — Percentage Transmissions of Praseodymium. Chloride Solutions.

X

D.=2.5mm.

D.=5 mm.

D.=10mm. ».

=20 mm.

X

D. = 2.5mm.

D. = 5 mm.

D. = 10 mm.

D. = 20 mm.

C. = 1.377N.

C.=0.688N.

C.=0.344N.|C.

=0.177N.

C. = 1.377N.

C. = 0.688N.

C.=0.344N.

C.=0.177N.

506

100

100

100

100

587

60

59

55

56

518

100

98

100

100

589

45

46

40

40

530

100

98

100

100

592

35

35

32

33

544

99

99

99

98

595

35

34

34

33

556

99

99

99

98

597

43

43

42

42

563

100

98

100

98

600

56

56

56

56

565

99

99

98

97

602

69

69

68

69

567

97

100

98

97

605

81

82

81

81

572

97

99

95

97

607

90

92

88

88

577

96

98

93

95

611

93

94

93

94

579

91

92

90

92

614

97

99

95

97

583

85

86

84

80

629

98

100

97

98

585

77

76

70

72

In terms of Beer's law, the curves representing these tables should be iden-
tical. Such curves are represented by fig. 34. Beginning with the curve
nearest the left and proceeding towards the right, the succeeding curves rep-
resent the four dilutions of praseodymium chloride as given in the preceding
paragraph.

100

o

I/}

i/>

Z
<
01

t—

L|J

<J

I—
z

UJ

<_>

IE

75

50

25

Cell Depth 2.5 mm.
Concentration I.377N

Cel, Depth 5 mm.
Concentration 0688N

Cell Depth 10 mm.
Concentration 0344 N

Cell Depth 20 mm.
Concentration 0.177 N

0.55/^

0.6/7 0.55u

0.6/7 0.55/y

Fig. 34.

0.6/7 0.55/^

0.6/7

The curve representing the most concentrated solution is nearest the left
of the figure. It is seen that these curves are identical to within the limits of
experimental error; the slight increase in the absorption with dilution is to
be attributed to the slit-width correction. The slit-width was in every case
0.4 mm. Water has no absorption in this region.

78

ABSORPTION SPECTRA OP A NUMBER <>F SALTS

The results recorded in these curves are in agreement with previous photo-
graphic results. The minimum in transmission, which in each case is about
30 per cent, occurs near X5900. The total deviation from Beer's law over
the dilution studied is not over 3 per cent, which is well within the experi-
mental error in this portion of the spectrum.

PRASEODYMIUM NITRATE.
Corresponding results for solutions of praseodymium nitrate are given in
table 17. The concentrations of the solutions, beginning on the left and
reading towards the right, were 1.282, 0.641, 0.320, and 0.160 normal, respec-
tively, the corresponding depths of absorbing layer being 2.5, 5, 10, and
20 mm., respectively.

Table 17. — Percentage Transmission of Praseodymium Nitrate Solutions.

X
506

D. = 2.5mm.

D.=5 mm.

D. = l0mm.

D.=20mm.

X

D.=2.5mm.

D.=5mm.

D. = 10mm.

D. = 20mm.

C. = 1.282.

C. =0.641.

C. =0.320.

C. =0.160.

C. = 1.282.

C. = 0.641.

C. =0.320.

C. =0.160.

100

100

99

100

587

62

60

60

58

518

98

98

100

100

589

48

46

46

45

530

100

100

99

99

592

40

40

38

36

544

99

98

100

100

595

37

36

36

36

556

99

98

100

98

597

44

44

44

42

563

99

99

98

97

600

56

56

56

54

565

96

100

98

96

602

68

67

68

67

567

98

100

98

97

605

79

82

79

80

572

95

97

96

95

607

87

89

89

88

577

93

92

94

93

611

91

92

94

92

579

90

92

90

91

614

95

95

97

96

583

83

86

85

83

629

98

98

99

98

585

73

74

72

72

100,

o

z

<

<

50

25

Cell Depth 2 5mm
Concentration I 282 N

Cell Depth 5mm.
Concentration 0.641 N

Cell Depth 10 mm.
Concentration 0 320N

Cell Depth 20mm.
Concentration 0 I60N

0.55^

0.6/.' 0.55a

0.6a 0.55a

Fig. 35.

0.6/i 0.55a

0.6a

The results are plotted in fig. 35. Reading from left to right, there is
shown the effect of increased dilution on solutions of praseodymium nitrate
and such concentrations and depths of absorbing layers as were mentioned
above, the most concentrated solution being nearest the left of the figure.

AS MEASURED BY MEANS OF THE RADlOMlCROMETEK.

70

As with the curves representing the solutions of praseodymium chloride,
these curves show that Beer's law holds quantitatively for solutions of the
nitrate. Neither the position nor the intensity of the band is altered more
than the limits of error of our work for the range of dilution studied. It
may be recalled that this is in exact agreement with the photographic results
recorded elsewhere in this monograph.

SALTS OF NICKEL.

Table 18 gives the percentage transmission of the nickel salts studied.
Beginning at the left of the table, column 1 gives X, taken at such intervals
as the solutions required, and reading towards the right are the results for
the following salts: nickel chloride, depth of cell 3 mm., concentration 2.74
normal; nickel nitrate, depth of cell 5 mm., concentration 1.68 normal; nickel
sulphate, depth of cell 5 mm., concentration 1.108 normal, respectively.

Curves representing these results are given in figs. 36, 37, and 38.

Table 18. — Percentage Transmissions of Solutions of Nickel Salts.

Nickel

Nickel

Nickel

Nickel

Nickel

Nickel

X

chloride.

nitrate.

sulphate.

X

chloride.

nitrate.

sulphate.

D.=3 mm.

D.=5 mm.

D.=5 mm.

D.=3 mm.

D.=5 mm.

D. = 5 mm.

C.=2.74 N.

C. = t.68N.

C. = 1.108N.

C.=2.74 N.

C. = 1.68 N.

C. = 1.108N.

544

71

69

81

796

16

20

33

556

64

69

79

805

18

24

40

563

62

65

76

814

22

31

44

565

60

61

73

825

28

36

49

577

56

57

68

834

32

40

55

583

52

50

66

845

36

46

58

587

47

46

61

856

41

48

64

592

40

40

60

867

46

51

66

597

34

33

40

877

49

52

68

602

26

28

43

888

51

56

69

607

22

22

37

900

53

56

68

614

17

17

33

912

54

56

67

618

12

12

26

923

52

51

64

625

10

9

23

933

50

48

62

632

8

6

18

938

47

44

59

638

5

6

15

955

42

38

52

643

3

4

12

966

36

30

44

650

0

4

12

978

30

24

38

662

0

3

10

990

26

20

33

676

0

2

10

1002

22

16

28

693

0

9

11

1012

20

14

24

755

7

8

10

1023

15

9

22

770

7

8

18

1035

13

8

20

779

11

13

20

1047

9

4

17

787

14

17

30

1060

6

4

14

Nickel Chloride.

Fig. 36, the curve for nickel chloride, shows an increasing absorption from
70 per cent transmission at X5200 to complete absorption at X6300. From
this point there is complete absorption to the regions X7200, and then a
gradual increase in transmission, reaching a maximum of 53 per cent near
X9000, then decreasing again to zero transmission at X10000.

As has been shown photographically, the visible spectrum of salts of
nickel consists of intense broad absorption bands both in blue and red,

80

ABSORPTION SPECTRA OF A NUMBER OF SALTS

showing a single region of transmission, extending to about X6500 in the red.
It is interesting to note that by means of the radiomicrometer we are able to
study another region of transmission which reaches a maximum near X9000.
From this point the absorption rapidly increases until the region of water
absorption is reached. It was not possible to determine whether the solu-
tion again became transparent beyond this point, on account of the water
absorption.

75

z
o

50

25

0.5^

Nickel Chloride

Cell Depth 3mm
Concentration 2 74 N

'0.7//

0.8//

Fig. 36.

1.18

Nickel Nitrate.

Fig. 37, which represents the second column of table 18, gives the trans-
mission, as observed, for a solution of nickel nitrate, concentration 1.68 nor-
mal and 5 mm. absorbing layer. This figure is almost exactly analogous to
the curve for nickel chloride, just discussed in the preceding paragraph, and
shows maxima at X5400 and X9000. There is complete absorption in the
region X7000 and beyond l.lfx. Figs. 36 and 37 represent approximately
equal amounts of the chloride and nitrate respectively, and are very similar.

75

o

§ 50

25

Nitrate

Cell Depth 5 mm
Concentration 1.68 N

0.5//

0.8//

Fig. 37.

1.18

Nickel Sulphate.

The curve representing the last column of this table is given in fig. 38.
The concentration of nickel sulphate was 1.108 normal, and the depth of
absorbing layer was 5 mm. It will be seen that this solution is slightly more
dilute than the other two solutions of nickel salts studied; and that in no

AS MEASURED BY MEANS OF THE RADIOMlCROMETER.

81

region does the curve representing the transmission reach complete absorp-
tion. Maxima in transmission appear near X5400 and X9000 and minima
of about 8 per cent at X6900 and XI 1000. Readings were not made at
greater wave-lengths with any of the nickel salts, on account of the intense
water absorption slightly beyond this region.

No striking or characteristic difference is noted in the absorption as shown
by the three curves representing the three salts of nickel studied. Each of
them shows maxima and minima in about the same region of the spectrum.

O.B,a

Fig. 38.

SALTS OF COBALT.

The photographic plate shows that in the visible region salts of cobalt
have a strong ultra-violet absorption, a band in the orange near X5000, and
increasing transmission from the center of this band toward the longer
wave-lengths. It was interesting to see whether solutions of cobalt salts
were completely transparent beyond the limit of sensibility of the photo-
graphic plate, or if such solutions again showed absorption bands in the
infra-red. It was also of interest to know whether equal concentrations of
the different salts showed the same or different absorption bands. With this
idea in view, the solutions of five salts of cobalt were studied. In each case
the concentration was 0.347 normal, and the depth of absorbing layer 10 mm.

The results are given in table 19. Beginning at the left are given, in their
respective order, the observed transmissions for 10 mm. of solution of the
following salts: cobalt chloride, cobalt bromide, cobalt nitrate, cobalt sul-
phate, and cobalt acetate.

The results of table 19 are plotted as transmission curves in figs. 39 to 43,
inclusive. A study of these curves shows that they are very similar, all
having maxima of transmission at the following points : X5950, X7800, X91 00,
and X10,600. In general, for all the salts of cobalt studied, the transmission
curves rise rapidly from X5000 to X5900, where the transmission reaches
about 65 per cent. The curves show a broad but slight absorption over the

82

ABSORPTION SPECTRA OF A NUMBER OF SALTS

region near X6500, and reach a maximum transmission over the region near
X7000 to X8000. The curves then descend, showing a series of small absorp-
tion regions near X8400, AS900, and A9800. The last of these absorption
bands shows a fairly sharp edge toward the shorter wave-lengths. Beyond
A10,500 the absorption increases rapidly until the region is reached where
water is practically opaque.

Table 19. — Transmissions of Cobalt Solutions.

Cobalt chloride.

Cobalt bromide.

Cobalt nitrate.

Cobalt sulphate.

Cobalt acetate.

X

D. = 10mm.

D. = 10 mm.

D. = 10mm.

D. = 10mm.

D. = 10 mm.

C. =0.347 N.

C.=0.347 N.

C. =0.347 N.

C. =0.347 N.

C. = 0.347 N.

544

12

21

20

14

15

556

19

33

31

26

31

572

39

50

35

48

53

583

48

61

64

57

59

589

59

68

68

62

69

592

63

74

73

67

72

005

68

74

77

67

72

614

70

74

77

68

74

622

72

74

79

68

75

632

72

75

84

69

76

640

72

75

76

70

77

650

73

76

79

71

76

660

73

78

81

72

80

667

74

78

82

74

83

678

79

82

85

77

84

686

80

84

88

79

85

693

81

85

90

88

88

708

84

87

91

81

89

720

84

89

91

82

92

730

85

87

92

83

89

741

88

88

88

82

91

755

87

89

89

83

91

767

87

92

89

82

92

779

88

89

89

84

90

791

90

89

89

83

90

805

91

89

87

84

89

819

8S

88

84

83

90

834

87

87

83

83

86

850

88

86

76

80

86

867

88

95

95

80

85

882

84

84

82

79

83

900

84

81

77

76

84

917

83

82

79

77

82

933

78

77

79

73

78

949

70

69

70

66

70

966

58

55

56

54

56

984

51

49

50

47

52

1,002

49

48

50

47

49

1,018

51

50

50

47

52

1,035

51

50

53

50

52

1,053

48

50

56

49

52

1,072

46

46

50

44

48

1,091

41

41

46

40

41

1,109

34

35

39

34

36

1,123

29

31

35

30

32

1,136

24

25

23

24

25

1,147

17

21

14

16

19

1,162

12

12

13

11

12

1,174

9

10

11

10

10

1,187

9

9

10

9

9

1,202

8

8

9

8

10

AS MEASURED BY MEANS OF THE RADIOMICROMETER.

83

100 r

73

50

Cobalt Chloride

Cell Depth 10 mm.
Concentration 0.347 N

0 5a

0 6^

07//

0.8a

Fig. 39.

0.9a

1.1/i

1.18

100 r

75

2

a

50

25l

Cobalt Bromide

Cell Depth 10mm.
Concentration 0.347 N

0 5a

0.6a

0.7a

0.8//

Fig. 40.

0.9a

W

1.1//

1.18

100, —

75

50

25

Cell Depth 10mm.
Concentration 0 347 N

0.5a

0.6//

0.7//

0.8//

Fig. II.

0.9a

1/'

1.1/.

1.18

84

ABSORPTION SPECTRA OF A NUMBER OF SALTS

Fig. 42, representing the curve of transmission of cobalt sulphate, appears
slightly different from the curves of the other salts of cobalt, having better-
defined bands in the region of X8000.

100

75

2
O

50

25

Cell Depth 10 mm.
Concentration 0.347 N

0.5//

0.6//

0.7/<

0.8//

Fig. 42.

0.9//

V«

1.1//

1.18

0.8// 0.9//

Fig. 43.

1.18

CHAPTER VII.

GENERAL SUMMARY OF RESULTS.

The work on the effect of temperature on the absorption spectra of .solutions
was extended to aqueous solutions, the range in temperature being from
ordinary temperatures up to about 200°. For this purpose a special form of
apparatus was constructed, made of brass and lined on the inside with gold.
This was for the purpose of preventing the hot vapor under high pressure
from coming in contact with any metal except gold.

With this apparatus the absorption spectra of aqueous solutions could be
studied up to 200°, just as well as the spectra of nonaqueous solutions in the
apparatus used by Jones and Strong, and described in Publication of the
Carnegie Institution of Washington No. 160.

With neodymium chloride the following bands remain unchanged by
temperature over the range from 20° to 200°: X3800, X4025, X4200, X4325,
X4440, X4600, X4690, X4750, X4820. The double band X5050 to X5270
changes very slightly. The bands X4275 and X5800 show marked changes,
the red edge widening and becoming more diffuse. The X5800 band widens
as much as 50 a.u. toward the red, the violet edge remaining sharp.

With the neodymium bromide, as with the chloride, only X4275 and X5800
show any marked changes with rise in temperature. The band X5800 widens
for the bromide 60 a.u. from 20° to 190°.

The X4275 band, for neodymium nitrate, shows a marked change, widen-
ing towards the red. The X4425 band widens about 15 a.u. from 15° to 165°.
The bands X5125 and X5800 show marked changes towards the red. The
change was greatest in the most concentrated solutions, although the total
number of absorbers in the path of light was kept constant.

The bands X4275 and X5800 for neodymium acetate show marked changes
on the red side, the latter widening as much as 80 a.u. The acetate bands,
for a given concentration of salt, are the most intense of all the neodymium
bands. When the solution was cooled down, the absorption spectra went
through exactly the reverse changes as when the temperature was raised.
Since the acetate band, X5800, is more intense for the same concentration
than for neodymium chloride or nitrate, and since the acetate is less dis-
sociated than the neodymium salts of the strong acids, it appears probable
that this band is in some way connected with the molecules.

The sulphate of neodymium shows the same temperature effect as the
other salts of this element.

The effect of temperature on the absorption spectra of cobalt chloride is
very slight.

There is a slight widening of the band of praseodymium chloride whose
center is near X4825. The X5900 band undergoes slight change with tem-

85

80 ABSORPTION SPECTliA OF SOLUTIONS.

perature, but from 20° to 160° it changes less than 25 a.u. There is no
appreciable change with temperature of the bands whose centers are near
X4425, X4650, and X4820.

The above-described temperature changes take place only in concentrated
solutions. In very concentrated solutions all of the praseodymium bands
show a slight widening with rise in temperature. We shall see that increase
in dilution affects the bands of praseodymium salts only when the solutions
are fairly concentrated. Thus, rise in temperature and increase in concen-
tration produce the same effect on the absorption spectra of solutions of
praseodymium nitrate.

The effect of rise in temperature on the absorption spectra of solutions of
uranyl nitrate is a general widening of the bands, with a slight shift of the
center towards the red. The general absorption ending near X3500 moves
rapidly towards the red with rise in temperature. All of the eleven bands
between X3500 and X4600 become more diffuse and broader with rise in
temperature, the X4180 band being most affected. The red edge of this
band shifts as much as 25 a.u. from 40° to 120°.

The uranyl sulphate bands X4175 and X4325 have their centers shifted
towards the red about 25 a.u. for a temperature range of from 20° to 185°.
The band X4750 remains unchanged, while the red edges of X4325 and X4550
shade rapidly towards the red. All bands below X4500 become very diffuse
as the temperature is raised, and at the highest temperatures are a single,
broad, hazy absorption band extending from X4000 to X4400.

The most marked widening is in the uranyl sulphate bands X4100, X4200,
and X4350, the center of each of these bands being slightly shifted towards
the red. The broad, hazy bands X5100, X5600, and X6200 are not appreci-
ably affected by changes in temperature.

None of the uranyl acetate bands seems to undergo change with dilution;
all of the nine bands on the plate undergo change with rise in temperature,
becoming more diffuse.

While some of the absorption bands of solutions are practically unaffected
by temperature, many of them widen as the temperature is raised. The
effect of rise in temperature is not to produce a symmetrical widening of
the bands, but most of the widening is towards the red. The violet edge of
the band usually remains pretty sharp. The red edge widens out, becoming
more hazy and diffuse.

The effect of dilution on the absorption of light by solutions was early
studied by Ostwald and others, especially in connection with the theory of
electrolytic dissociation. It was known that both molecules and ions in
solution absorb light, and the question was whether they have the same or
different absorption. It was not possible to answer this question satis-
factorily by means of the prism spectroscope. It has been possible to solve
this problem by means of the grating.

Jones and Anderson had shown that if molecules and ions absorb differ-
ently, the difference is slight. We therefore worked over a wide range in

GENERAL SUMMARY OF RESULTS. 87

dilution, comparing the absorption of a concentrated solution with one 500
times as dilute.

The neodymium chloride bands X3400, X3450 to X3600 are not affected by
change in dilution. The sharp hand X4275 is more intense in the most con-
centrated solution. The bands near X5100, X5200, and X5800 are markedly
affected by dilution, the former two appearing as distinct bands in the most
dilute solution. The broadening of these bands with concentration is fairly
uniform, both towards the red and the violet ends of the spectrum. The
intense band from X5690 to X5850 is greatly affected by concentration,
widening almost entirely towards the red. This widening is about 50 a.u.
When a more dilute solution was used to start with, the broad band X5700 to
X5825 is the only one which widens with increase in concentration, the widen-
ing being about 25 a.u. When a still more dilute solution is used as the
starting-point, there is no appreciable change in any of the bands with
increase in concentration.

With neodymium bromide, with increase in concentration there is a
slight increase in the intensity of X4275. The bands X5090, X5120, and X5210
narrow uniformly with dilution. The greatest change is in band X5750,
which widens towards the red as much as 30 a.u., the violet edge remaining
practically unchanged. When half the concentrations were used, the only
band affected by dilution is the one near X4800, which widens with the con-
centration as much as 20 a.u. When all of the dilutions were again doubled,
there was practically no difference between the absorption spectra of the
various dilutions.

The effect of dilution on the absorption spectra of neodymium nitrate is
probably greater than on any other neodymium salt. Then bands X5090 and
X5125, in the most concentrated solution, have so broadened as to become
one band. The band X5220 widens uniformly towards both the red and
violet with increase in concentration probably as much as 70 a.u. Starting
with a different concentration, the X5750 band widens as much as 40 a.u. as
the concentration is changed. When the original solution was still more
dilute, only the X5750 band changed appreciably.

The band X5750 of neodymium sulphate widens with concentration as
much as 25 a.u. It remains unchanged if the initial solution of neodymium
sulphate is more dilute.

A number of the bands of neodymium acetate change with the dilution.
The X5210 band narrows about 10 a.u. with the first change in dilution, and
then remains unchanged with further increase in dilution. The broad band
X5750 changes about 55 a.u. with the change in dilution studied. When
only half the initial concentrations were used, only the bands X5220 and
X5750 underwent change. With neodymium acetate further increase in
dilution produced still further narrowing of the absorption bands. When
one-fourth the initial concentration was used, the band X5750 underwent
change, narrowing about 20 a.u. This is the only salt of neodymium in
which a change in a band was noted at such a high dilution.

88 ABSORPTION SPECTRA OF SOLUTIONS.

The X4675 band of praseodymium chloride narrows towards the violet
about 20 a.u. with increase in dilution, while the broad band X5900, under
the same conditions, shows a narrowing of about 25 a.u. When more dilute
solutions are employed, none of the bands shows any change with dilution.
The changes in the two bands X4675 and X5900 with dilution are much less
than with the corresponding bands of neodymium.

The two bands X4450 and X4650 of praseodymium nitrate widen about 20
a.u. with increase in concentration in very concentrated solutions. In the
more dilute solutions there is no change.

The X4700 band of uranyl chloride shows marked widening with increase
in concentration, especially towards the red end of the spectrum. The
X4900 band also shades off rapidly towards the red end of the spectrum.
When more dilute solutions were used, both X4600 and X4700 gradually
widen with increase in concentration.

The X4700 band of uranyl bromide widens uniformly with increase in the
concentration. When a more dilute solution was employed as the starting-
point, none of the bands changed with dilution.

The absorption of concentrated solutions of uranyl nitrate is complete to
X4500. With increase in concentration this gradually recedes towards the
red, amounting to as much as 100 a.u. The X4700 band widens under the
conditions about 20 a.u. The sharp band X4875 widens slightly with
increase in concentration.

In the more concentrated solutions of uranyl salts many of the bands
change with change in the dilution , while in the more dilute solutions there
is scarcely any change at all.

The introduction of the radiomicrometer into this work converted it into a
quantitative study of the absorption spectra of solutions. The grating
spectroscope and photographic method were very efficient in locating the
positions of the absorption lines and bands from wave-lengths X2000 to
about X7600 ; and the photographic method gave some approximate idea as
to the intensities of the various lines and bands. This method is, however,
only roughly quantitative, and is very limited in the range of wave-lengths
to which it can be applied.

The radiomicrometer provides us with a quantitative method for studying
the intensities of the various lines and bands, and also greatly extends the
range of wave-lengths that can be studied. In the earlier work with the
radiomicrometer much time and labor were expended in perfecting the
instrument, especially in constructing a sensitive radiomicrometer with a
short period. Dr. Guy accomplished this very successfully.

The earlier work was practically limited to the study of the absorption
spectra of solutions of neodymium salts — neodymium chloride, bromide,
nitrate, and acetate.

The results were plotted in what is known as transmission curves, which
express the percentage transmission for the different wave-lengths. Solu-
tions of different concentrations of a given salt were studied, the depth of

GENERAL SUMMARY OF RESULTS. 89

layer of the solution varying inversely as its concentral ion. The product of
the depth of layer times the concentration is a constant. If the solvent
plays no role in the absorption, then the transmission curves lor the different
concentrations of any given salts must fall directly over one another — the
different curves would be the same curve.

We found that, in general, the more concentrated the solution the less t be
transparency and the broader the absorption bands; this is exactly what we
obtained with the grating spectroscope and the photographic method. But
in the more dilute solution the intensity of the bands was greater. We
observed further, that with increase in dilution the middle of the band is dis-
placed towards the longer wave-lengths.

The same general changes with dilution in the absorption spectra of solu-
tions of neodymium bromide were observed as with the chloride; the more
dilute the solution the narrower and more intense the bands.

The bands of neodymium nitrate, in general, show the same changes with
dilution as those of the chloride and bromide. With increase in dilution the
intensities of the bands increase, and their centers are displaced somewhat
towards the longer wave-lengths.

The three salts of neodymium, then, all show an increase in intensity with
dilution. A possible explanation of this phenomenon, based upon reson-
ance, has been offered. It is a well-known fact that a resonator, when excited
by vibrations from a single vibrating resonator having the same pitch,
vibrates more strongly than when set into vibration by a large number of
resonators, one of which has the same pitch as its own and the others slightly
different periods. In a word, if several vibrators are near together, every
one exerts a certain influence on the others. The result is that no one of
them has exactly the same period as the original resonator. Each resonator
damps the other and we have less perfect resonance.

In a concentrated solution the resonators are relatively close together and
mutually affect one another. The result is imperfect resonance and the
absorption bands are less intense in the more concentrated solution.

In the more dilute solution the vibrators are farther removed from ono
another and are surrounded by large amounts of water of hydration. The
damping effect would thus be diminished. In such cases we would have more
perfect resonance and the resulting absorption bands would be more intense.

Subsequent work has, however, shown that a part of this effect can possibly
be explained as due to the fact that the slit-width used was not infinitesimal.

It was found by the radiomicrometer, as with the grating spectroscope
and photographic plate, that for a given concentration the acetate absorbs
much more than any other salt of neodymium.

One of the most interesting facts thus far established by means of the
radiomicrometer is the effect of the dissolved substance on the absorption
spectra of water. We noted that aqueous solutions of hydrated salts
were often more transparent than pure water. This is obviously a very
remarkable fact, and we at once took up its careful study. We compared

90 ABSOEPTION SPECTRA OF SOLUTIONS.

the absorption of aqueous solutions of strongly hydrated salts with the
absorption of a layer of water equal in depth to the water in the solution
through which the light was passed. We then carried out similar experi-
ments with salts which, in the presence of water, combine with only a small
amount of it. In a word, we compared the absorption of light by water with
the absorption of an equal depth of water in aqueous solutions of strongly
hydrated salts, and the absorption of light by water with an equal depth of
water in aqueous solutions of salts which are scarcely hydrated at all.

The nonhydrated salts with which we worked were potassium chloride,
ammonium chloride, and ammonium nitrate. It was neeessarjr in all of this
work to choose salts which themselves have little or no absorption in the
region in which water absorbs, i. e., in the infra-red. It was found that
aqueous solutions of the above-mentioned compounds showed the same
absorption of light as water having the same depth as the water in the solu-
tions in question. This is exactly what would be expected. The dissolved
substance and the solvent do not combine with one another to any appre-
ciable extent, and it would be very difficult to see how either could appre-
ciably affect the absorbing power of the other.

When we turn to the strongly hydrating salts, very different relations
manifest themselves. The salts of this class that were studied were cal-
cium chloride, magnesium chloride, and aluminium sulphate.

In the case of a 5.3 normal solution of calcium chloride, the solution is the
more transparent from 0.9/* nearly to 1/*. The water then becomes the more
transparent for a short distance. From 1.05/* to 1.2/1 the solution is the
more transparent, becoming as much as 25 per cent more transparent than
the pure water. The water becomes more transparent than the solution
only at and near the bottom of the "water-bands" at approximately 1/*. This
is what we should expect if the solute exerts a damping effect on the absorb-
ing power of water. When a smaller depth of the solution of calcium
chloride is used, the water in the region 1.25/* is more transparent than
the solution. From this band on to the longer wave-lengths the solution
becomes more transparent than the water until 1.42/* is reached, when both
solution and water are practically opaque.

The results for magnesium chloride are essentially the same as those
obtained for calcium chloride. The main difference is that from 1.0/* to
1.1 /*, in the case of magnesium chloride, the water is more transparent; while
for calcium chloride in this region the solution is the more transparent.
The difference between water and the solution of magnesium chloride in this
region is, however, not great. For wave-lengths longer than 1.1/*, the solu-
tion of magnesium chloride, like the solution of calcium chloride, is more
transparent than the water, the difference for the two salts being of the same
order of magnitude.

When a smaller depth of layer of the solution was used, the water was the
more transparent from 1.22/* to 1.34/*. For the longer wave-lengths the
solution was the more transparent.

GENERAL SUMMARY OF RESULTS. 91

The curve for aluminium sulphate brings out this new feature; at 1/jl the
solution is more transparent than the water. Beyond 1.04/t the water is
transparent to 1.1 7/jl, beyond which the solution is the more transparent, as
with magnesium and calcium chlorides.

In the region 1.2/z water is the more opaque when a shallower layer of solu-
tion is used. From 1.29/j. to 1.36/*, water is the more transparent; beyond
1.38/1 the solution is the more transparent.

The explanation of these remarkable results is that they must be due to
some action of the dissolved substance on the solvent. That the solvent
can affect the absorption spectra of the solution was first shown by Jones and
Anderson;1 and a large number of examples of this same action has since
been found by Jones and Strong.2 The action was satisfactorily explained
as due to a combination of the solvent with the dissolved substance, and this
explanation accounted for many facts which could not be otherwise satis-
factorily explained. This theory of solvation in solution has aided us in
explaining many phenomena which the theory of electrolytic dissociation
alone could not account for, as has frequently been pointed out.

The same solvate theory of solution seems to aid us in explaining the facts
just discussed. Those substances that do not form hydrates when in the
presence of water show normal results as far as absorption spectra are con-
cerned. Their solutions have the same absorption as so much pure water,
the substance itself showing no absorption.

It is the hydrated salts, and only these, which give the abnormal results
herein recorded. The combined water seems to have less poiver to absorb light
than free or uncombined water. This would account for all of the facts
observed.

It should be noted that the presence of the salt shifts the absorption of the
water towards the longer wave-lengths. It was earlier observed that rise in
temperature and increase in concentration shifted the absorption of the salt
towards the longer wave-lengths. The effect of rise in temperature and of
increase in concentration is to simplify the hydrates existing in the solution.
This simplified resonator shifts the absorption towards the red. The effect
of the salt on the absorption of the water is the same as rise in temperature
and increase in the concentration of the solution on the absorption of the
dissolved substance. It may be that the dissolved substance diminishes the
association of the solvent and thus simplifies the resonator. This may be
true especially with the water of hydration or the water combined with the
dissolved substance.

This new line of spectroscopic evidence bearing on the solvate theory of
solution is regarded as probably the most direct that we have or can hope to
obtain in favor of the view that there is combination between solvent and
solute.

In studying the absorption spectra of salts, the intensity of the light after
passing through the solution of the salt in question, was compared with the

1 Cam. Inst. Wash. Pub. 110. 2 Ibid., 13) and 160.

92 ABSORPTION SPECTRA OF SOLUTIONS.

intensity of the light after passing through a depth of water equal to the
water in the solution in question.

The absorption spectrum of neodymium chloride shows three pronounced
minima representing the three absorption bands with their centers near
X7300, X7950, and X8700, and less pronounced bands near X7150 and X9000.
The latter may be due in part to the solvent; for all four dilutions studied
the X7300 and X7900 bands show complete absorption over a considerable
range of wave-lengths. The minimum of band X8700 is gradually lowered
with increasing dilution.

The maximum transmission of solutions of neodymium chloride occur
near X7600 and X8400, these solutions becoming almost completely trans-
parent beyond Ijjl, except for the absorption of the solvent of this region. It
seems that Beer's law holds, in general, for the infra-red absorption of solu-
tions of neodymium chloride.

The minima in the curves for the more dilute solutions of neodymium
chloride are in about the same positions as those in the more concentrated
solutions, but the solutions being more dilute are more transparent; hence
the minima are not so pronounced. The maximal absorption occurs near
X7300 and X7900. The X8700 band has a minimum transmission of 64 per
cent. There is complete transmission in the regions X7200, X7600, X8300,
and X9300. The drop in all of the curves beyond 0.9/i is due to the absorp-
tion of the water.

With further increase in the dilution of the solution, there is a lowering of
the maxima. The change is most pronounced in the X8700 band, and here
the absorption of the water is the most pronounced. The absorption of the
water together with the correction for slit- width may account for this change,
and Beer's law may hold for the dilute solutions of neodymium chloride
almost as well as for the more concentrated.

Neodymium nitrate shows three minima at X7300, X7950, and X8750.
The nitrate bands are not as intense as those of the chloride, the solution of
the nitrate not being as concentrated as that of the chloride. The nitrate
bands, like those of the chloride, become more intense with increase in
dilution. The absorption of water becomes more and more pronounced
beyond 1/x.

The minima for solutions of neodymium acetate fall at approximately the
same positions as with the chloride and nitrate.

Solutions of neodymium acetate, as indicated by the photographic
method, show greater absorbing power than those of either the chloride or
nitrate. The X7000 band is slightly more intense in the acetate. The
more intense bands X7300 and X7950 show the same tendency to have their
minima lowered with increasing dilution. There is a rapid increase in the
absorption near X9500, due to the water present in the solution.

Solutions of praseodymium salts are transparent in the infra-red as far as
1.5ju, except two very weak bands which fall in the midst of the intense water
bands. Praseodymium salts have two groups of bands, one in the green

GENERAL SUMMARY OF RESULTS. 93

near X4600 and another near X5900. We have limited our investigations to
the latter band, on account of the small amount of energy transmitted at
X4600. The curves representing the absorption of different concentrations
of praseodymium chloride are identical to within the limits of possible experi-
mental error. The results obtained with the radiomicrometer arc in accord
with those found by the grating and photographic plate. The minimum in
each case occurs near X5900. The total deviation from Beer's law, as shown
by solutions of praseodymium chloride, is within the limits of experimental
error.

The curves for praseodymium nitrate show that Beer's law holds here as
well as for the chloride.

Solutions of nickel chloride show an increasing absorption from X5200
to X6300, where it is complete. Complete absorption extends to X7200.
Transmission increases to X9000, then decreases to zero at X10000. The
visible spectrum of salts of nickel consists of intense broad absorption bands
in the blue and red, having a single region of transmission in the red, extend-
ing to about X6500. By means of the radiomicrometer we could study the
region of transmission near X9000. Beyond this we could not go because of
the absorption of the water.

The absorption of nickel nitrate closely resembles that of the chloride.
There is maximal absorption at X5400 and X9000. There is complete absorp-
tion in the region X7000 and beyond 1.1/z.

The solution of nickel sulphate studied is slightly more dilute than the
chloride and nitrate. In no region is there complete absorption. There is
maximal transmission near X5400 and X9000, and minima at X6900 and
XI 1000. Readings were not extended beyond this region on account of the
intense absorption of the water. The three salts of nickel studied have just
about the same absorption spectra, the curves showing maxima and minima
in just about the same regions of the spectrum.

Salts of cobalt in the visible region have a strong ultra-violet absorption
There is a band in the orange near X5000, and increasing transmission
towards the red. The infra-red absorption of solutions of cobalt salts was
studied, and the absorption of the chloride, bromide, nitrate, sulphate, and
acetate compared. The transmission curves for all of these salts have
maxima at X5950, X7800, X9100, and X10600.

The transmission curves for all of the salts of cobalt studied rise rapidly
from X5000 to X5900. The curves show a broad, slight absorption over the
region near X6500, and reach a maximum transmission from X7000 to X8000.
There is a series of small absorption regions near X8400, X8900, and X9800.
Beyond X10500 the absorption increases rapidly to the region where water
is practically opaque.

The curves for cobalt sulphate are slightly better defined than those for
the other cobalt salts.

iffl

WW/ZStim

■ ■'■■'•'<•:■
.■.■■■.■■■.■■J:

■

tt&KKMCfc

■'•■'■■■■

mssi

Ml