Cantor lectures on photography and the spectroscope

w®m®m

Qllje TObrary

of ilic

JJaliib ;QunIap (DbscrViatoru

:]9resriiieu lig

Dr.. ..&,.. A.,. .Chant..

April.. 23 j. 19.35

SOCIETY FOE THE ENCOURAGEMENT

ARTS, MANUFACTURES, AND COMMERCE.

CANTOR LECTURES

PHOTOGRAPHY AND THE SPECTROSCOPE,

CAPT. AJ3NEY, F.R.S.

DELIVERED BEFORE THE SOCIETY OF ARTS, APRIL 20TH AND 2-th, 1885.

a*Js

LONDON :

PRINTED liV W. TROUJCCE, 10, GOTJGH SQUARE. FLEET STREET, E.C.

1885.
Prict One Shilling.

PHOTOGRAPHY AND THE SPECTROSCOPE

CAPT. ABNEY, F.R.S.

LECTURE I.— Delivered April io, 1885,

Every lecturer, when he begins, must have
a text of some description, and I propose to
make my text for the lectures a plate exposed
to the spectrum. You see before you a spectro-
scope comprising' a collimator, two prisms,
and a camera, with a lens of 13-inch focal
length, and in this slide is a sensitive collodion
plate.

The spectrum of the hot carbons of the
electric light is upon the focussing screen
which you see before you, and I will simply
expose this plate, and refer to it from time to
time as my lecture continues.

The plate is given six seconds' exposure to
the light of that spectrum, and now in the
subdued light coming from this lantern, whose
sides are covered with translucent orange
paper, I see the picture is coming out under the
action of the ferrous oxalate developer. After
fixing, we see that we have the photographed
spectrum on the plate.

This is the text on which I have to hang my
lectures. We have three things to consider.
First of all, we have got the light, then we
have the apparatus, and then the sensitive
material on which the spectrum is taken.
The white light from the carbon poles, in
passing through the apparatus, is spread out
into a coloured band, which we call the
spectrum ; and the spectrum has effected a
( hange in the sensitive salt of silver, as is
shown by the blackening on the application of
what is called a developer. The cause of the
change in the sensitive material is what I first
ess myself to.

To conceive a right notion of photographic
action we must first of all conceive, in the most
elementary manner, the structure of matter.
The structure is beyond our actual visual
acquaintance, but we may be able to visualise
it from the way it behaves ; we have to draw
our conclusion about it from evidence of an
experimental nature. What we want to get
is a mental picture of matter.

Physicists have come to the conclusion that
homogeneous matter is composed of molecules,
or small masses which are altogether similar
one to another, i.e., they have the same com-
position. In different matter these molecules,
have different weights. Further, it is believed
that the molecules, or the small particles of
matter, are themselves composed of atoms,
which we take to be the fundamental unit of
matter. Now, from experimental data, Sir
William Thomson and others have come to
conclusions as to the limits of the size of these
molecules, and also as to their distribution in
space. From the kinetic theory of gases, it is
concluded that the diameter of a molecule lies
somewhere between one twenty-five millionth
part of an inch and one two hundred and fifty
millionth part of an inch. Further, in gases
it is conceived that the molecules are free to
move in straight lines in any direction, the
direction being altered only when the molecules
collide ; that is to say, when they strike one
against another. In a liquid, the particles are
bonded much more closely together, and the
free path of the molecules is very much
shorter. That is to say, that they cannot go

4

from one place to another without very much
more frequently coming into contact with
other molecules ; and the molecules pass from
place to place at a very much slower rate than
they do in gases. A solid, such as is our silver
salt, is conceived to be such that the molecule
has no free path, but is confined in a limited
space in which it can oscillate, moving round
a mean centre. As to the distribution of these
molecules in liquids and solids, Sir William
Thomson has arrived at very definite con-
clusions also. In a lecture at the Royal
Institution, he said that he concludes that in
every ordinary liquid, or transparent solid, or
seemingly opaque solid, the mean distances
between contiguous particles is less than one
twelve millionth of an inch, and more than one
two thousand five hundred millionth part of
an inch. Those are big figures, but still the
distance apart is very small. " To form a con-
ception of this," he says, " imagine a globe
of water as large as an ordinary football to be
magnified to the size of the earth, each con-
stituent molecule being similarly magnified.
The magnified structure would be more coarse
grained than a heap of small shot, but probably
less coarse grained than a heap of footballs."
So you see that, by magnifying to this extent,
you have a coarse-grainedness which, of
course, is only relatively coarse grained after
such an enormous magnification. Or you may
put it in a different sense. If you magnify
eight thousand diameters by an ordinary
microscope — and that is about the limit to
which a microscope will magnify ; and if you
magnify that eight thousand diameters again
eight thousand times, you would be able to see
the molecular structure of water. So much,
then, for molecules.

We will now turn to the atoms. These will
not bear such a very large disproportion of
size to the molecules as do the molecules to the
smallest visible particles. We must, however,
I think, conceive that every atom (and this is
an important point) is charged with energy
very much in the same way that the magnet is
charged ; only, instead of two poles, as a
magnet has, each atom has only one pole. It
is unipolar.

Now, suppose that this energy is something
like electrical energy. We know that positive
repels positive, and that negative repels
negative. And further, we know that the
positive energy will attract what is called nega-
tive energy ; and if the two be exactly equal
when they combine, of course there will be a
neutral state. But in the case of the atoms of

matter, circumstantial evidence tells us that
the amount of electrical energy which is upon
a given atom of matter— if you like to put it
in that way — is never the same as it is upon
another atom of matter; that is to say, there
is always a surplus of one over the other. Thus
we may have an atom charged with what we
may call fi/tis 2 of energy, and another one
charged with minus 1 of energy. Those two
atoms, on coming together, give you a result of
energy of plus 1, and this would again be
capable of attracting another atom of matter
which was charged with a negative energy,
and so on. From chemical considerations, it
would appear that plus and minus energies of
different atoms, as I have said before, are
never exact multiples of one another, and that
when they are bonded together there is always
an excess one over the other. A good example
of the energy of the combination of atoms
together may be shown by the combination
between a gas, chlorine, which we have here,
and the metal antimony, and you will see
that when the latter, as powder, is thrown into
the former, the two combine with an evolution
of heat, showing that a vast amount of energy
is given out. The chlorine and the antimony
form chloride of antimony ; that is to say, five
atoms of chlorine and one of antimony. [A
small quantity of finely powdered antimony
was dropped into a jar of chlorine.] You see
the evolution of heat between those two ; so
much, in fact, that the chloride of antimony, as
it was formed, was at a perfectly bright white
heat. The case immediately before us is the
silver salt. Let us experiment with that in a
similar way. Into the chlorine I will throw
some powdered silver, and I wish you to notice
the difference between the results in the two
cases. [A small quantity of the powdered
silver was dropped into a jar of chlorine.]
You see that the combination between the
silver and the chlorine only produces a red
heat, whereas antimony produced a white
heat. In other words, the combination between
antimony and chlorine is much more vigorous
than the combination between silver and
chlorine. If you had to separate the atoms of
chlorine from those of the antimony, you would
have to use very much greater force than
if you had to separate the atoms of silver from
the atoms of chlorine.

When two electrified bodies attract one
another, they attract one another inversely as
to the square of the distance. That is to say,
if there is a distance of one foot between them,
they attract one another with a force of say —

1 . If they are two feet from one another, they
only attract one another with a quarter that
amount. Supposing atoms attracted one
another according to the same law, then of
course they being so very close to one another,
the attraction would be considerably greater
than if they were visibly apart.

But besides this attraction between atoms
comprised in the molecules, there also seems
to be a repulsive action, into which I will not
enter more fully now, because that would be
almost beside my subject ; but I may say that
besides the atoms attracting one another (we
will take chloride of silver for instance) when
they get within a certain distance of one
another, they repel one another, and so there
is a continual oscillation between the atoms
composing those molecules.

I will try to show you on the screen how we
can picture the motion to ourselves. It is only
a mental picture, but still it will give us a sort
of idea of what happens. [An image was
thrown on the screen by means of reflection].

In this circular glass trough of water is
floating a little magnet, the magnet being
held at the surface of the water by a cork.
Passing round this coil, which is large enough
to surround the trough, is an electric current
from three Grove cells, and if I place it round
ell which contains the little magnet, and
not quite on a level with the water, you will
find that the single magnet goes into the
centre of the water. It is repelled from the
sides by the current that is floating round that
wire. Well, now, we have here one magnet.
Suppose I put another magnet in. The
ends attached to the cork have poles
of the same name. They repel one another
to a certain extent, and yet the force
from outside makes them go as near one
another as possible. By moving this coil
vertically we can make them separate and
oscillate, and we can picture to ourselves the
way in which two atoms in a molecule may
oscillate, and be attracted, and yet repelled
one from another. I put another little magnet
in, so now there are three ; and here perhaps we
haveapicture of chloride of silver, which I say is
composed of one atom of silver and two atoms
of chlorine. We can still make them vibrate
and oscillate. Here we have a mental picture
— at least it is a mental picture to me — of the
way in which the atoms of chloride of silver
may be made to oscillate. Again 1 take four,
and we repeat the same thing. Here we have
a picture of ammonia — three atoms of hydro-
gen and one of nitrogen oscillating. And so

I might go on. I might put five or six or a
dozen in, and we might get some idea of the
way in which they would all oscillate.

Here, then, we have endeavoured to draw
from visible phenomena a mental picture of
the way in which atoms of a molecule are
vibrating.

I must, however, call to your mind that
those magnets are vibrating only in one plane,
whereas of course the atoms of a molecule are
vibrating, not in one plane, but in space of
three dimensions ; but anyhow, I hope that
you have got into your mind at all events the
same kind of mental picture regarding the
oscillations or vibrations of the atoms which I
have in mine. I think that the case of the
magnets is a particularly happy one, because
from all the svidence which we have at present
we are led to the conclusion that all atoms of
matter are really charged with electricity, or
what answers to electricity of either one
name or the other; that is, either positive or
negative.

Now, we will throw a spectrum on the screen.
I will call to your recollection what it is. I
am now going to send the light of the lamp
through this bisulphide of carbon prism, and
I need scarcely say that the prism has to play
an important part in spectrum photography.
The wave length of the red is about one forty*
thousandth of an inch, and the wave length of
the violet, which is on the left of the screen,
is about one fifty-seven - thousandth of an
inch. Each ray of light is transmitted in
air at the rate of about 190,000 miles in
a second. Thus the number of vibrations
of the red rays is 500 million millions, and 700
million millions in the case of the violet rays,
and this rapid succession of blows batters
against anything upon which they fall. The
mean violet, I may say, is the photographic
light par excellence, and we shall recollect
that such rays might beat upon the sensitive
salt which we expose to it 700 million million
times in a second. Therefore, you see, if you
give an exposure of the 100th of a second
you still have seven million millions of vibra-
tions beating on the sensitive plate, so there is
ample vibration to effect any change on the
molecule of silver chloride, supposing -always
the amplitude (or distance of swing) is
sufficient. Instantaneous photography will
not be complete, I suppose, until you can reduce
by a million times.

We may take it that an atom vibrates some-
what in the same way that a pendulum
vibrates. Here I have a very rough contriv

ance to show what I mean. I set the pendulum
swinging. Now picture to yourself that the bob
is an atom, and picture to yourself, also, a
wave of light falling upon that pendulum ; if
the wave of light be synchronous with the pen-
dulum, it will increase the swing, or, in other
words, it will increase the amplitude of the
swing of the pendulum. For a rude illustra-
tion, suppose I take puffs of my breath as
illustrating the beating of the wave of
light, and suppose the atom to be at rest ; I
begin, and I blow ; every time I give a well-
timed puff to that pendulum, the pendulum
increases in amplitude, or swing. But if my
breath does not come in unison with that pen-
dulum [blowing irregularly], you see that very
soon I should bring that pendulum to rest ; in
other words, unless the wave of light beats in
unison with the atom, the amplitude cannot
be much increased. It is true that as long as
the breath strikes the bob as it is going away
from me the amplitude is increased, but if the
puffs are regular and slightly more rapid or
slower than the pendulum oscillation, the
amplitude must eventually be diminished.

Here we are met with a difficulty, and a very
great difficulty. I exposed the plate to the
spectrum, and you see the blackening not only
was where one wave of light synchronised
with that atomic motion, but that there
were a great many waves of light, extend-
ing from the ultra-violet as far as the
blue which affected it. How are we to get
over that ? That is a difficulty which has
puzzled a great many people. I would ask
you again to form a mental picture of how that
could possibly arise. I do not say that it is
the correct way, but all I say is that you can
form a picture in your own mind, can conceive
of how it could be done. Here, I have another
pendulum, but in this case the bob is attached
to an elastic band. The time of the vibration
of a pendulum depends upon the length of the
pendulum. Therefore, if during the time of
the oscillation of the pendulum I alter the
length, I also alter the rate at which the pen-
dulum vibrates during any instant. I pull
down the weight of the pendulum, and at
the same time set it swinging, and you will
see that during every part of this motion the
length of the pendulum is altered so that a
great many differently timed puffs of breath
might be synchronous with the pendulum. It
is not like this other rigid one, where it is of
a definite length, but here the length of the
pendulum keeps altering. I only ask you to
form a mental picture of the way in which such

a thing might happen. In this way you can
picture to yourself how a molecule might
vibrate, and still be synchronous with more
than one vibration of light.

Proceeding another step, I may say at once
that, to my mind, the theory of the photographic
image is well established. I know that there
are some people who differ, but in my own
mind the formation of the photographic
image is not a working hypothesis, but it is a
theory. The difference between a working
hypothesis and a theory is this — that you adopt
a certain idea and say, " I will work upon that
idea, and see whether every experiment fits with
the idea I have conceived. If it does not fit,
then that working hypothesis is no use. I
must give it up, I must take some other work-
ing hypothesis." As regards the idea of the
formation of the photographic image, I think
that it has passed from the stage of the work-
ing hypothesis into one of a really acceptable
theory. It does not follow that everybody will
accept it, but still it is an acceptable theory,
accepted by most people. I am not going to
enter into that very strongly to-night. At the
next meeting of the Photographic Society, I
propose to deal with it more fully ; but, at the
same time, I just wish to state publicly, to
perhaps a more extended audience than I shall
see at the Photographic Society in about three
weeks' time, that this photographic image
theory — that is to say, the theory as to the
action of light upon molecules of silver — is as
well established as, at all events, the wave
theory of light itself. Now, I am going to
show you an experiment which, perhaps, will
help to illustrate what I mean by the vibrations
of atoms. In this slide I have got a gelatine
plate, and I have a little flat iron which has
been made warm. It is rather too warm to be
borne comfortably. Here I have a phosphores-
cent plate, which I propose to illuminate with
magnesium wire, in order to give an even
source of light ; I press this flat iron against
the back of the plate which is in this slide for
a short time. I shall not let the plate cool, but
while it is warm I will expose it to the phos-
phorescent light for about fifteen seconds. The
plate is now allowed to become cold, and is
developed. If everything has gone right, we
ought to have something which shows us that
the oscillations of the atoms of bromide of
silver (which is the silver salt on this plate)
have been given extra amplitude by the action
of the heated iron to the back of the plate.
I am afraid that I cannot show you the
development in the light. [When the develop-

ment had been carried out the plate was
shown.]

You now see we have a picture of this flat
iron produced by the deeper blackening of the
heated part, though the whole plate was given
but a short exposure to the light from the
phosphorescent plate. I will impress this
further upon you. I have here a collodio-
bromide emulsion plate. But in this case,
instead of heating it by a flat iron, we
will heat it by immersion in hot water.
Of course a collodion plate is not so sensi-
tive as a gelatine plate. I put it into cold
water for a short time to moisten it, and then
dip half of it into some nearly boiling water;
on withdrawing it, I expose it to this candle,
and develop it when it gets cool, which we
effect by placing it a short time in cold water.
It will be seen that the part immersed in hot
water is much blacker than that which was
exposed cool. If I heat the plate and allow
it to cool and then expose, there will be no
effect. The plate will develop normally, for
the increased amplitude of vibration will have
ceased, and the light will have to perform the
same work on each part of the plate. Now,
in whatever manner increased amplitude is
given, when the cause of the increased ampli-
tude is withdrawn, the amplitude will cease in
the same manner. The case before us next
was the cause, and it will cease after a very
short period, in other words, when the plate
gets cold. One of the chief reasons against
what we may call the "vibration theory"
of the photographic image, namely, that the
molecule is unaltered by the action of light,
is this— that the increased amplitude would
cease with the same rapidity with which it
would cease when the hot iron was applied
to tin' back ; that is to say, after five or ten
minutes the amplitude of the vibrations would
come back to the normal extent, a condition
which is not fulfilled in the photographic
image.

I can illustrate this in a very visible manner.
I think you can all see this phosphorescent
plate. Now, what is the reason of that phos-
phorescence taking place? It is that the
atoms of the molecules which comprise this
phosphorescent material are swinging in a
certain rhythm, which gives us the sensation
of light. Now, if I apply a hot iron to the
back of this plate, I think at once you will see
that the image of the hot iron is present.
Here is the same kind of action taking place
in the one case as in the other.

Now we come to another point, which is a

slightly different one, and that is the energy
of radiation. I may say that the energy of
radiation is a subject on which I could dis-
course for a good many hours, but here I can
devote but two minutes to it. I must try to
make it as clear as I can. I hold in my hand
a little instrument which is called a thermo-
pile, which you see has a narrow slit which
could be narrowed to any degree of fineness ;
attached to it is a screw motion, which will
make that slit travel along the base of the
instrument ; beneath that slit are some thermo-
electric couples. It is not my business to
enter into how they are made, but still we
know that, when thermo-electric couples are
heated, an electric current is generated suffi-
ciently strong to cause the needle of a galvano-
meter to deviate ; and the amount of energy
of radiation which falls upon the face of the
pile can be measured by the deviation of the
galvanometer needle, from the energy heating
the lampblack at the junction of the couples.
In a great many experiments which were made,
this thermopile was caused to travel along
the spectrum by the screw motion, and at
every part of the spectrum at distances of,
say, a quarter of a turn or half a turn of the
screw, the amount of deviation which was
given to the galvanometer needle was read off.
By that means we are able to compare the
energy existent at different parts of the
spectrum. The spectrum used was that of
the electric light, the comparative energies
at different parts of its spectrum I have in the
diagram— at five turns of the screw we have
the end of the red, and at different turns we
have the yellow, the green, the blue, and the
violet ; whilst from five to twenty we have the
dark rays which lie below the red, and with
which we are not to deal to-night at all events.
The energy, I may say, being measured by
taking the amount of the deflection of the
galvanometer needle, you will see that the
dotted line divides the energy area into two
parts.

On measuring this area of the curve in
which lie all dark rays, and the area of the
curve for the visible rays, it will be found that,
roughly speaking, the energy of the latter
rays are about half that of the former. But
for photographic action we do not have any-
thing like that amount. The red rays for
ordinary photographic work are useless ; and
why that is we shall see by and by. We will say
that the photographic action stops at the blue,
and we find that the total energy of radiation
which is used for photographic purposes in

8

the electric light, is only about one-hundredth
part of the whole energy of radiation. The
remaining ninety-nine parts are wasted as far
as photography is concerned, except in so far as
they heat up the molecules in the same way as
the flat iron heated up the molecules on the
photographic plates. The other curves show
the energy of incandescent lamps. You will
see that they have very little of what is called
actinic power ; that is to say, they have very
little blue ray at all compared with the arc
light. In the lowest curve we have a lamp at
only a yellow heat, the middle curve being
that at a white heat, and you will notice the
enormous difference there is in the energy
between the two. The energy of the middle
curve, which measures the total energy of
radiation from the incandescent light, is about
twelve times that of the visible power. Yet,
when you have to measure the photographic
part of the spectrum, you will see that it is
only about eighty. That is to say, supposing
you have a filament of an incandescent lamp
which is one-hundredth of an inch wide and
half-an-inch long, then if you take an arc
electric light and cut off from the glowing
positive pole the same area, the photographic
value of the one, area for area, is about eighty
times that of the other. [A spectrum was
thrown on the screen.]

I will ask my assistant to put in front of the
slit something which I showed you at my last
Cantor Lectures, and which I dare say you have
forgotten all about. That something is a film of
the same silver salt with which I photographed
the spectrum at the commencement of the
lecture. You see that it cuts off all the violet,
and well down into the blue. I want to show
you that the colour of the photographic spec-
trum is perfectly different from that which the
human eye can see. I wish to show you a
little device by which, perhaps, I shall be able
to give you an idea of the integrated colour.
A tolerably bright spectrum is on the screen
of the camera ; I raise the screen so that the
spectrum falls on a lens placed a little beyond
it ; and if we had time, I dare say that we
should be able to get a screen placed in the
focus of the second lens, so that the recom-
bined colours would form a white patch, with-
out the slightest tinge of colour. We have
got a white circle, however, which is sufficient
for our purpose, though at one margin there
is a very narrow red fringe to it. [A white
patch about 6 inches in diameter was formed
on a transparent screen about 6 feet away
from the camera.] In the place where the

coloured spectrum is in focus, I place a hori-
zontal aperture, about £-inch wide, and by a
little arrangement I can, by strips of card, cut
off any colour I like from falling on the collect-
ing lens, so that it recombines only the re-
maining colours.

You remember that the photographic spec-
trum does not extend as far as the green,
ordinarily speaking, so now I cut off all
rays as far as where the photographic spec-
trum begins, and you can see the colour of the
light, which is really useful for photography.
It is a sort of sea-green colour. If I were to
take that light, and pass it through a slit and
a prism, you would soon find that the whole of
that spectrum would be photographically
active, because all the light which is not
photographically active has been cut off. I
shall have to revert to this in my next lecture.

I will show you one more method of re-
combining the photographically effective colour
disc ; that is by taking the ordinary disc, and
cutting out the red and orange. We have, then,
only the green, the blue, and the violet ; and
those, when they are combined together, ought
to give you pretty nearly the integration of the
colours which are ordinarily photographically
active. I will ask my assistant to spin it in
front of the lantern. [The instrument was
rotated.] I do not know whether you can all
see the colour-chart which I hold in my hand,
but those who can will see that the colours,
when placed in the blue-green light, appear
totally different from what they did in the
whiter light. The yellows are much deepened,
and the reds are much blacker.

I will ask now to have the spectrum thrown
upon the screen once more, and we will
again pass this colour-chart througli the
spectrum. The colours are very pure for
pigments. I think that it is the finest colour
chart of the spectrum which I have ever seen.
It is one prepared by Professor Piazzi Smyth,
and appears in his Madeira spectroscopic
observations. Notice that the blue appears
perfectly black when the chart is in the red,
the red at the left hand being brilliant.
Passing it into the yellow, the yellow is
vigorous ; the blue is black, and the red
undimmed. Upon my passing it still farther
on in the green, you will see that the red is
blacker, and the orange is blacker, whilst the
yellow still keeps its colour, and the blue
begins begins to get more bright. Passing
it still farther on into the violet, we see that
the yellow is now perfectly black, the red has
gone, and the blue begins to shine outi

9

Passing still further, you will see that the blue
still shines out, but is less intense, all the
other rays appearing black. Upon my passing
it again rapidly through, you will now be pre-
pared for the changes that take place. In
this lantern, which has been used to form
the spectrum, the light passes through a slit.
The slit, you see, is perfectly straight, with
parallel edges. Now comes the question, " Is
it necessary that light, in order to be decom-
posed into a spectrum, should be passed
through a slit of this description, or what
shapes may it be allowed to take ?"

I propose to try to answer this query in
an experimental manner a little. First of all,
we will see what the effect will be if we use
no slit at all. You see that the colours are not
pure. I replace the slit, and you will see at
once that we now have, not the various colours
light overlapping, but a tolerably pure
spectrum. Now let us take a slit of another
shape— a zig-zag slit; and here we have
another form of spectrum delineation of the
rays. Placing a metal in the arc, the bright
lines due to the vapour flash, and, it will
be seen, take the zig-zag form of the slit.
There is, then, no particular reason for
using a straight slit, except convenience.
Then, again, I may take a ring slit, and to
test its value we will put a little silver in the
arc to show you. I am not simply showing
this as a pretty experiment, but I want to show
you that such a slit is absolutely useful in
photography, the spectrum of silver now on
the screen shows rings of different coloured
rays. It is a very pretty spectrum. This form
of slit is extremely useful in one branch of
spectrum analysis.

You are perfectly aware that, during a total
eclipse, the body of the moon covers the sun ;
but that there are seen beyond the dark moon
certain red protuberances which belong to the
sun, and are known as "prominences." It
has been the work of astronomers to determine
the composition of those protuberances, and
also to form a definite idea of the corona of
light which surrounds the body of the sun,
and can only be properly seen during a
total eclipse. The picture on the screen is a
representation of the total eclipse of the sun
which took place in Egypt in 1882. It is a
negative picture, and of course the dark halo
which you see around was seen as a bright
halo, and the white disc is the black moon.
On the left-hand bottom corner you may
notice the comet which was discovered during
the eclipse, and which received the name of
Tcwfik, after the Khedive of Egypt. Round
the disc of the moon are little prominences.
Those prominences are vastly more bright
than the corona itself, which is the halo
extending some distance round the sun. Thus
we have a bright ring of light round the moon
surrounded by a feeble light. The former,
when viewed by means of a lens in front of
which a prism is placed, shows rings of colour
composing these prominences, and of course
these rings can be photographed.

I now show a transparency of a photograph
taken in Egypt by means of the slitless
camera, from which much valuable information
has been derived.

The ring slit was used by an Italian astro-
nomer about 1870 ; but the eclipse in Egypt
was the first time it was entirely successful for
photography.

10

LECTURE II.— Delivered April i-j, 1885.

In my last lecture, I left off with the use of
the slit in the spectroscope, and I showed you,
I think, that under certain circumstances the
slit which had the form of a ring was useful,
having previously demonstrated that it was
not necessaiy that the slit should be straight,
but that it was most convenient that it should
be so. I will next deal with the subject of the
prism. We know that prisms are employed to
separate the different coloured rays, as each
colour is differently refracted as it passes
through the prism, and it is this difference in
the index of refraction between the red ray and
the violet ray which gives the amount of dis-
persion in forming the visible spectrum. Of
course, if we go beyond the violet, there are
invisible rays, while again below the red there
are also dark rays, which also have their in-
dices of refraction, but I wish to show you the
influence that the material of the prism itself
has on the dispersion of the visible spectrum.

I have here a prism of 6o° built up of six or
seven different triangles of glass. It is appa-
rently homogeneous, but when we pass light
through it we shall find that it is anything but
homogeneous ; in other words, the different
portions are differentlyrefractive. The different
portions of the prism are all glass, as I have
said, but of different densities, and the denser
the glass, the more are rays refracted, and the
greater dispersion between the red and violet
there is. [A slice of light was passed through
this built-up prism, and the different spectra
thrown on the screen.] You will notice, by the
spectra on the screen, that the length of the top
spectrum between the red and violet is much
smaller than that of the bottom spectrum. The
glass which gives the dispersion to the latter
is much denser glass than that which gives it
to the former. Practically speaking, there-
fore, we may say the denser the glass the
greater refraction, and the greater dispersion
there is. For most purposes in spectroscopy,
it is as well to use as dense a glass as possible
in order to get the maximum amount of dis-
persion. I will now combine three prisms
together, two of a light glass, and one of a
denser, and we get a combination, in which,

although the main beam will pass straight
on to the screen, yet the presence of dis-
persion is also shown by the formation of
a spectrum. This is an example of what
is called a direct vision prism. The spectrum
is given by the differences of the refrac-
tive indices for each ray in the two kinds
of glass. For some purposes this kind of
compound prism is very useful, and parti-
cularly for lecture experiments, but, as a rule,
for photographic purposes I should not re-
commend it, on account of the internal re-
flections which take place between the different
surfaces of the glass, though they are cemented
together. You must recollect, wherever there
is a difference in density between two media,
in other words, a difference in the refractive
indices, there is always a certain amount of re-
flection, and those reflections, being white
light, are rather apt to fog the plate, and give
you false notions of what you get in the
photograph.

We come now to a much more important
point with regard to the spectrum, and that is,
what is the best material to use. In those
prisms which I have already shown you, the
material was glass. Now glass is, compara-
tively speaking, a mixture of materials, and has
no definite chemical formula ; but when we come
to a material which has some definite chemical
formula, we find that, as a rule, it has certain
properties which are invaluable in certain forms
of spectroscopy, more particularly when the
photographic plate has to be brought into
requisition. Quartz is an example of this ; it is
a definite compound of silicon and oxygen, and
we find that it has certain definite advantages
which are not to be found in glass prisms. The
dispersion is not quite so great as it is with
glass, but, on the other hand, it lets through
rays which are cut off completely by glass, as
I hope to show you on the screen. This quartz
prism has very well-worked faces, and we will
send a beam of light through it, and then
proceed to investigate its behaviour. [Spec-
trum thrown on screen.]

I may further say, in reference to this, that
the condenser in that lamp is quartz, the lens

n

is quartz, and the prism itself is quartz, so that
we are dealing with nothing but quartz.
Now, the question comes, is thereany advantage
to photographers in using such a material as
quartz. Let us first see the extent of the
spectrum. By placing a card which has been
washed over with quinine in the ultra-violet
part of the spectrum, you are able to see
these ultra-violet rays glowing with a pale blue
light, and you will notice to what a great
length these rays reach beyond the ordinary
visible point of the spectrum. Now, by placing
a piece of glass in front of the slit, you will
see thatilie ultra-violet spectrum is very much
shortened ; in other words, the glass has
absorbed these rays. I may repeat the experi-
ment with a card which has been brushed over
with paraffin oil, and the same result holds
good.

I have here a photograph of the electric
arc taken in another manner, to which I shall
have to direct your attention presently. The
ight in this case has to pass through no glass
whatever. The spectrum was taken by a
diffraction apparatus ; for the top part of the
spectrum a glass was interposed in front of
the slit, and we see the difference there is in
the spectra, owing to the use of glass in one
case, and not in the other. The glass ap-
parently cuts off many useful rays ; but I will
now draw your attention to the solar spectrum
taken in the same way, in which there has been
a glass placed in front of the slit for one spec-
trum, and not in the other. Both spectra,
practically, reach the same limits. We now
can answer as to whether it would be advisable
for photographers to use quartz lenses for
ordinary photographic purposes or not. Re-
collect that every ray of light you saw fluoresce
on the screen is useful for photographers when
they are using a light such as we have in the
electric light. You will see, then, from that,
if the electric arc light was usually employed,
all those rays which are cut off by the glass
could not be utilised by them, and, therefore,
there would be so much power wasted. Now
photographers, as a rule, do not work with the
electric light, but with sunlight ; we have
seen that in the solar spectrum taken under
similar conditions, the glass practically cuts
off none of the ultra-violet rays ; the atmo-
sphere of the earth, or of the sun, or both,
cuts off the extreme ultra-violet rays before
the light reaches us. We therefore come
to the conclusion that, so far as photo-
graphic work with sunlight is concerned,
there would be no advantage in using a quartz

lens over the ordinary photographic lens.
Some years ago, Mr. Claudet made an agate
lens, which he considered would give him
greater advantages over the ordinary photo-
graphic lens, simply because he could utilise
the ultra-violet rays, but I think you will see
from this there is no advantage in using such
a lens. Remember, however, if you are photo-
graphing the spectrum of the electric light, or
using it for illuminating a sitter, there is a
very great advantage in using quartz. We
may use another definite chemical compound in
the shape of Iceland spar. I have here a very
beautifully worked prism of Iceland spar,
which has a definite composition of calcium
and carbon, and I dare say that we shall
reach very nearly to the same ray limit as
we did in the quartz experiment. Iceland
spar holds an intermediate position between
quartz and glass. It was with such a prism
as that that Dr. Huggins took his famous
star spectra, and I thought it might interest
you to throw one or two of these on the
screen. They are very small, but the
definition is very beautiful. Many of the
black lines in these spectra indicate, probably,
hydrogen. It remains to be seen whether Dr.
Huggins has attained any advantage in using
Iceland spar instead of glass, for if the ultra-
violet stellar light is absorbed, as with sun-
light, no advantage would be gained. I may
mention that he gives the composition of the
stars by reference to the spectral lines of well-
known elements.

One more point is this : Would it be ad-
vantageous to use a mirror instead of a lens ?
There is a great deal to be said about this,
particularly in spectroscopy, where we have
to examine everything minutely. The material
we utilise most easily in the case of a mirror
is silver ; that is to say, we get a glass mirror,
and silver it on the front surface. Now, the
question is, does the silver reflect every ray in
the same way that quartz would transmit it ?
Here I have a photograph which should give
an answer to that question. The bottom half
of the spectrum was taken as reflected from a
quartz surface, the top half of the spectrum was
reflected from a silver surface, and you will
see that at one certain part of the latter the
rays are very nearly absent, though beyond
that again they are present. Where those
rays are wanting is just at the end of the solar
spectrum, and therefore, when using sunlight,
it is no great advantage to use a quartz re-
flector over a silver reflector. In spectro-
scopy it is necessary to know exactly the

li

qualities of all the substances with which you
are dealing.

One question in photography and in spectro-
scopy is, what width of slit you would use —
what slice of light would you allow to pass
through ? Here let me give you a demon-
stration. In the centre of this black disc there
is a fine line of light, and there is a micro-
meter screw by which we can tell how many
thousandths of an inch wide it is. As a rule,
about ji^ of an inch is the dimension used for
ordinary work.

I have been referring to the photographs
to two spectra on the same plate, and I
must show you how it is managed. For this
purpose, it is necessary to have an adjunct to
the slit, and that is a shutter, which is able to
cut off half the slit at one time, and after-
wards leave that part open and close the
other half already used. By this means we
can get one spectrum adjacent to another. In
comparing spectra of different metals with
each other, we are able to tell whether we
have any two lines coincident one with the
other.

Photographic spectroscopy is the easiest
thing in the world when you know how to do
it, but it requires a deal of patience to learn
every dodge. As a rule, a photographer is a
patient man ; indeed, there ought to be no
class of men who have more patience than
photographers ; hence spectroscopy should
not be difficult to them.

Here is another piece of apparatus which is
very useful in the spectroscope. It is an
apparatus by which you can take a great many
spectra on one plate. I need not enter into its
details, it is simply a dark slide, which by a
rack and pinion motion can be raised, so that
the plate gives a fresh surface at each ex-
posure. The only light accessible to the plate
comes through an opening of about three-
quarters of an inch wide cut longitudinally in
the shade. By this method we can get about
sixteen different spectra of different materials
on the same plate.

Here is another piece of apparatus which is
also useful in investigations with photo-
graphy. It is, a slide in which you can expose
plates in different gases or liquids, that is to
say in water, in alcohol, in nitrogen, and so
on. It is essentially a glass cell which slips
into a dark slide especially adapted for it ; on
the top there is an air-tight junction which is
screwed down, and there are two little tubes
through which you can fill the cell with gas or
water, or whatever other material you wish to

use. This is very useful in investigating the
behaviour of different sensitive salts under
different conditions of moisture, pressure, &c.
This cell has been used in a great many
hundred experiments, and I hope it will be used
in a great many more. Those who are going in
for spectroscopy should not be without such an
apparatus as this, for I do not believe much
real investigation can be done without some-
thing of the kind. The sensitive salt of silver
acts differently when isolated from its atmo-
spheric surroundings, and the only way to
ascertain how it does so is to expose it with
other surroundings, and to differentiate the
results one from another. There is no such
thing as a perfect vacuum ; you cannot say
you expose a plate in vacuo, and, for this
reason, I say you have to differentiate between
the different media in which you expose a plate,
in order to get at the true result which would
happen supposing you could expose the plate
in vacuo.

You saw last time how you could recombine
a spectrum by means of a lens, to form white
light.

Now, I want to show you that it is not im-
possible to develop a plate in white light. I
expose a plate behind a negative to the electric
light, and in the cell which is placed in the
patch of white light is some developing solu-
tion (which is quite colourless). The plate is
dipped into it. The image comes out into it,
although exposed to white light, without fog,
which was supposed to be an impossibility. I
have another plate placed behind the same
negative. I expose half of that plate for half
a minute to the white light on the screen, and
the other half to apparent darkness, but in the
same position on the screen, for a couple of
minutes. The plate on development shows that
the half which was exposed to what was pre-
sumably white light gives no image, while the
half exposed in the dark shows a perfect picture.
I dare say many of you have guessed my
trick, for it is merely a trick, but for those
who have not, I will show you how it is done.
It is perfectly easy, by mixing two elements of
light of different refrangibility, to produce a
colour which, at all events, to our eyes is a
white light. But you must not take it for
granted that wherever you can see white
light you can photograph with it, because
it is quite possible you may not. It is
only a trick, but some of these tricks bear
fruit in a very practical manner. I will
re-form white light again, and we will examine
it by means of the colour-chart I showed you

_

<3

last time. You will see that when the red is
placed in the white light there is blackness —
no colour whatever — the yellow looks bright,
as does the blue, all the other colours are
gone except some few which are of a non-
descript colour. The meaning of it is this,
we have simply a combination of yellow and
blue, which gives us the appearance of white
light. [The blue and yellow rays were shown
to be coming through two slits placed at the
focusing screen of the camera.] The blue has
no power of acting on the iodide or chloride of
silver, neither has the yellow, and, therefore, the
white light which is made by the combination
of those two colours is powerless to act on
films made of such materials as those. We
can also produce a white light, practically, by
a red and green, and if we examine this (which
is a very good imitation of white light) in the
same way, you will not see the whole series of
colours in the colour chart any better than you
did before. The red comes out perfectly,
but the blue is no longer visible ; the blue
becomes green, and the violet becomes red ;
the yellow is also not intense. This is because
we have only two colours present, viz., the red

and the green. The apparent darkness to
which we exposed the one-half of the plate was
in reality the dark ultra-violet light, and I
need say no more regarding that.

I told you last time that this was a very
interesting way of studying the spectrum. You
see how, by combining two lights together, you
may have a light which is perfectly safe for
certain salts of silver. On the screen is the
spectrum taken on the three ordinary salts of
silver — chloride, iodide, and bromide. The
iodide stops exactly at the violet. Below that
light we have no action whatever, and we,
therefore, may expose an iodide plate with
impunity to any rays below the violet.
A bromide plate, you see, is sensitive
down as far as the yellow, and, therefore,
it would be impossible to develop a bromide
plate in such a light as I showed you just
now, whereas iodide is perfectly capable of
being developed in such white light; the
chloride again stopped very nearly with the
limits of violet, so that it would be safe to
develop a chloride plate in such a light.

[The lecturer concluded with a brief explana-
tion of the diffraction spectrum.]

LONDON: PRI.V1KD EY W. TRCINCE, JO, GOUGJ1 SQUARE, EU ET STREET, E,C

SOCIETY FOR THE ENCOURAGEMENT

ARTS, MANUFACTURES, AND COMMERCE.

CANTOR LECTURES

ON

LIGHT AND COLOUR.

CAP1A1N W. DE W. ABNEY, C.B., RE., F.R.S.

Delivered before the Society of Arts November 26, and December 3, 10, and 17, 1888.

LONDON:
PRINTED BY W. TROUNCE, 10, GOUGII-SQUARE, FLEET-STREET, LONDON, E.C.

1889.

8 V L LABUS.

LECTURE I.

The production of colour and its dependence on the kind of illuminant. — The spectrum and its
recombination.— Simple colours. — The characteristics of colour.— Colours of pigments.

LECTURE II.

Interference colours. — Production of colour by absorption ; by fluorescence. — The measurement
of the luminosity of colours. — Colour contrast. — Colour-blindness.

LECTURE HI.
The effect of the dilution of colours. — Mixtures of colours. — Impure colours. — The measurement
of colours in terms of a standard. — The reproduction of the colours of a pigment.

LECTURE IV.
The action of light on pigments.— The cause of change.— The effect of sunlight, sky-light, and
artificial light. — Rays effective in causing change. — Moisture and oxygen necessary to cause change.—
Work done by the absorption of light.— Chemical effect, heating effect.

LIGHT AND COLOUR.

BY

CAFIAIN W. DE W. ABNEY, C.B., R.E., F.R.S.

LECTURE I.— Delivered November 26, 1888.

I hold in my hand a series of colours of various
hues and depths, some of them are fugitive
and others are fast colours, and it is the
object of the lectures I have been called upon
to deliver to show how we can measure and
mix colours, and what causes the fading of
some by light. In four lectures this subject
can by no means be treated exhaustively, and
I can only endeavour to explain, in as familiar
language as I can command, and by some
plain experiments, what I desire to enforce
upon your minds. A great deal has been
written in the last two years on the subject of
the fading of water-colours, and from what
I have gathered from the newspaper corre-
spondence, it Is not quite unnecessary that a
few familiar discourses on the subject should be
given, to prevent a repetition at all events of
some of the blunders that have been made in
physical phenomena. It may be known to
some who are present here to-night that Dr.
Russell and myself have carried on a series of
experiments during two years on the subject of
the fading of water-colours, and as our report
to the Science and Art Department, which was
presented to Parliament, pleases neither the
party who cry out that water-colours are stable,
nor yet the party who proclaim the contrary,
we may presume that our results are not
altogether wrong. To these experiments I
shall refer later in the course of lectures.

Now, to commence with the elements of
colour from the physicist's point of view.
I wish to show you that the colour of
an object depends on the composition of the
light falling on it, on the material on which
such light falls, and on the eye of the person.
The screen which I have here is what we call
white, when viewed by ordinary daylight or
artificial light, and such a screen not only will
reflect white light, but also all coloured lights
with the greatest brilliancy possible.

Let me throw a spectrum on the screen to
serve as a text. If a brilliant spectrum be
looked at, we see that it is really divided
into three colours, blue, green, and red, with
shades of other colours blending these colours
into one another. I am not going into the theory
of the matter, but I would ask you to remember
that the mean red light has a wave length of
about 38,000 to the inch, the waves being in the
luminiferous ether of whose existence we only
know by circumstantial evidence, the green of
about 50,000 to the inch, and the violet of
about 64,000 to the inch. The other colours
have intermediate wave lengths.

I would remind you of the old experiment
that red, green, and blue, when combined
together by means of rotation, give a grey
light which can be matched by a combina-
tion of black and white. Here we have such
a combination forming a grey in the electric
light. The reason assigned for this is, that
in the eye there are three sets of nerves, one
which responds to the red, one to the green,
and the other set to the blue. When the disc
is at rest, an image of these three coloured
sectors is formed on the retina, and the nerves
lying at the parts of the retina on which the
image falls respond to these colours, and we
see the sectors coloured. If there is astigmat-
ism, or defects in the optical apparatus of the
eye, the image is not sharp, then we have an
image of part of the two colours adjacent
blended into one another, or again if the
discs rotate rapidly, s» that the same part of
the retina receives the coloured images in
quick succession, all three sets of nerves are
brought into use, and we have an impression
of white, or rather grey, produced. But this
subject I shall allude to again in one of my
subsequent lectures.

We can recombine also the pure colours of
the spectrum by several plans, the simplest to

my prejudiced mind being that which I intro-
duced. I take away the lens of long focus,
and put one of shorter focus in its place
attached to a camera, for reasons which I will
shortly explain (Fig. i).

On a collimator, G, to which is attached the
usual slit, is thrown, by means of a condensing
lens, a beam of light, which emanates from the
intensely white-hot carbon positive pole of the
electric light. The collimating lens, L2 , is filled
by this beam, and the rays issue parallel to
one another and fall on the prisms, Pi and Pa,
which disperses them. The dispersed beam
falls on an ordinary camera lens, Lj, of slightly
larger diameter than the height of the prisms,
and a spectrum is formed on the focussing-

Fig. I.

///

////

//

//

screen, D, of a camera. When the focussing-
screen is withdrawn, the rays would form a
confused patch of parti-coloured light on a
white screen, F, placed some four feet off the
camera. The rays, however, can be collected
by a lens, L4, of about two feet focus, placed
near the position of the focussing-screen, and
slightly askew. This forms an image on the
screen of the near surface of the last prism,
P2 ; and if correctly adjusted, the patch of
light should be pure and without any fringes
of colour. The card, D , is a strip which fits into
the aperture left forthe focussing-screen in the
camera. In it will be seen a slit, s2) the utility
of which will be explained later on.

It often happens that a second patch of
white light, comparable to that formed, is re-
quired. Advantage is taken of the fact that
from the first surface of the first prism P, , a
certain amount of light is reflected. Placing
a lens. L5, in the path of this reflected beam,
and a mirror, G, another square patch of light
can be thrown on the same screen as that
on which the first is thrown, and this second
patch may be made of the same size as the
first patch if the lens, L5l be of suitable focus,
and it can be superposed over the first patch
if required.

We have now a square white patch upon
the screen, from the re-combination of the
spectrum. If I wish to diminish the bright-
ness of this patch, there are at least two
ways in which I can accomplish it. First, by
closing the slit of the collimator, and, second,
by the introduction of rotating sectors, M, which
can be opened and closed at pleasure during
rotation in the path of the beam.

The annexed figure (Fig. 2, p. 3) is a bird's-
eye view of the instrument. A A are two sectors,
one of which is capable of closing the open
aperture by means of a lever arrangement, C,
which moves a sleeve in which is fixed a pin
working in a screw groove ; D is an electro-
motor causing the sectors to rotate, and the
aperture in the sectors can be opened and
closed at pleasure during their revolution.
To show you its efficiency, if I place two strips
of paper, one black and the other white, on
the screen, and cast a shadow from a rod,
by the direct white light on the white strip,
and a shadow from the same rod by the
reflected light on the black strip of paper, and
interpose the rotating sectors in the path of
the reflected light, the aperture of the sectors
can be closed till the white paper appears
absolutely blacker than the black paper.
White thus becomes darker than lamp-black,
owing to want of illumination on the former.

We all talk about white light ; we say that
the electric light is white and that gas light
is white. I wish to show you that the white-
ness is a mere matter of judgment.

I throw the shadow, by the electric light, of
a thick rod on white paper, and another
shadow by gas light, on the same paper, and
we at once see that the shadow illuminated by
the electric light seems blue, whilst that illumi-
nated by the gas light appears orange, yet we
speak of both gas light and the electric light
as white lights. Evidently, if these two differ
so much in colour, pigments will take different
hues when illuminated by them. Putting

paper coloured with red, blue, and green pig-
ments in the shadows, the change in hue is at
once apparent. Placing in the shadow illumi-
nated by the electric light a strip of paper
coloured orange (Fig. 3), by orange chromeand
aureolin, we see that now the electric light re-
flected from it appears of very nearly the
same hue as the light from the gas reflected
from white paper. Gas light, we may say
then, is orange rather than white, if we
take the electric light as the standard.

We have seen that colours appear of dif-
ferent hue in the electric light to that which

they appear in gas light, and I wish to enforce
this more strongly upon you by an experiment
which I introduced a year ago. In front of the
condenser of the electric light lamp I place a
circular aperture some inch in diameter, and
by means of a lens throw an image of it on
a white screen. We may suppose this to
represent the sun, the colour of the light
being very much the same as that which it has
in England about midday in the middle of
May. In front of the aperture I place a
trough containing a solution of hyposulphite
of soda, and then drop into it dilute hydro-

Fig. 2.

chloric acid, and stirring up the two together
very fine particles of sulphur slowly separate,
and the white light, owing to the law of scatter-
ing by small particles, loses some of its com-
ponents, and we have a gradual reddening of
the sun — first yellow, then orange, and finally
a red— the series forming a very exact repre-

Fig. 3.

sentation of the colours of a setting sun. If
we place coloured pigments in this changing
light, we see how, towards sunset, the blues
become darker whilst the reds change but little
in hue. It may have been remarked that in
an evening the last colours in a picture to
disappear are the reds. The colour of sunset

light now imitated before you gives a clue to
the reason of this.

We may as well trace the cause of this
change in colour. Placing a cell contain-
ing hyposulphite of soda in front of the
slit of the spectroscope, and throwing the
spectrum on the screen, and then adding
the dilute hydrochloric acid, we find that
as the light from the reflected beam (which
we throw just above the spectrum) be-
comes yellow, orange, and then red, so the
spectrum loses the violet, then the blue,
then the green, till finally the red alone
remain.

Let me further exemplify that you cannot
know what effect the colour of the light has
upon a colour unless you know its composi-
tion.

The slit S, in the card D (Fig. 1) can be
passed through the spectrum, and as it cuts off
all the colours of the spectrum, except that
passing through the slit, we have different
coloured square patches of light thrown by—

what 1 will now call— our patch-forming appa-
ratus, the colour of the patch being that of
he colour issuing through the slit.

Now sodium, when ignited, gives a peculiar
yellow light, due to a line in the orange. If I
send the light from this sodium line through
the slit S4, we have a square patch of sodium
light on the screen. The rod casts a shadow
as before, but instead of casting a second
shadow by the reflected beam, I cast a shadow
from gas light, when it will be seen that the
two illuminated shadows have almost the same
colour.

I now shall perform a common Christmas
experiment, and ignite some spirits of wine
in which salt has been dissolved, and illuminate
with that light cards on which various blue, red,
green, and yellow pigments have been placed,
and we see that all the pigments partake of
various shades of orange, instead of the
colours seen by gas light. The reason of this
is apparent ; in gas light we have all rays
present, in the sodium light there is only
orange present. We shall see that as the colour
of a blue pigment depends principally on the
reflection of blue rays, that of a green of the
green rays, and so on, it is only to be ex-
pected that the colours of pigments, when
illuminated by pure orange light alone, will
only give different shades of orange.

This shows also that light or colour may to
the eye appear to be the same and yet be very
different in optical composition. I cast two
shadows of the rod in the patch-forming appa-
ratus, one by the recombined spectrum and the
other by the reflected beam, and pass the card,
D, with the slit, S3. in it along the spectrum.
One shadow will be illuminated by the white

light and the other by the light from the parts of
the spectrum coming through the slit S3. If I
place emerald green in the shadow illuminated
by white light, I find that there is one point in
the green of the spectrum which matches it in
hue, and I can make them of the same depth
of colour by the introduction of the rotating
sectors. Evidently, then, the coloured light of
this part of the spectrum and that of the
emerald green might be mistaken for one
another, and so with other colours. There are
some pigments, however, which cannot be
matched by the spectrum colours.

That emerald green is a combination of
colours I will at once show you. A strip of
card is placed in the spectrum, on one half of
which is this pigment. Half of the breadth
of the spectrum falls on the white card
and half on the pigmented card. It will be

Fig. 4.

Emerald Gr

-

seen that the emerald green reflects other
colours of the spectrum besides that which it
matched in the colour patch-forming appa-
ratus. The combination of all these other
colours in the proportions reflected from the
pigment, form the colour which, in the simple
colour of the spectrum, we should call emerald
green. So if we pass other pigments through
the spectrum we get similar results, though
not all pigments can be so matched.

LECTURE II.— Delivered December 3, 1888.

In the last lecture I finished the matching
of the colour of pigments with parts of the spec-
trum, and to-night I will endeavour to showyou
that colourless bodies can be made coloured,
under certain conditions, although the light
that falls upon them is colourless. I told you
last time that the waves of red light are such
that if you put 38,000 end to end they make
up an inch. If in the sea we have two sets of
waves, one set of which is exactly half a
wave behind the other, then the crest of the
one wave will exactly fill the trough of the
other, and instead of motion we shall have
rest. Suppose I have a colourless body, whose
thickness is comparable with a wave of red
light, and that a wave of red light when
reflected from the back surface is half a wave
length behind that reflected from the front
surface, we get darkness instead of light.
The easiest way to obtain a colourless body
answering to the above conditions is to use a
soap film stretched across a vertical aperture.
Its thickness is found to be comparable with
a wave of light, and as it gradually thins by
gravity, some part of the film becomes of the
thickness that the reflection from the back
surface is half a wave length behind that
reflected from the front surface, the red is
annihilated at such place. There will be
another thickness of film in which the green
light would be similarly absent, and yet
another in which the blue is absent, and so
on. The light reflected from the first locality
would be the components of white less the red,
in the second the same less the green, and in
the third the same less the blue.

I can show you the kind of colour that is
seen by the suppression of one small part of
the spectrum, by using our patch-forming
apparatus and passing a thin rod along the
spectrum, which cuts out the part required.
It will be seen that the patch is no longer
white, but coloured. These colours, remember,
are not simple colours, but white light, with
some colour abstracted.

Putting a soap film on a ring in the beam of
the electric light, at an angle of about 450 with
it, the light is reflected on the screen, and a lens
in the beam forms an image of the ring. At first

the film appars white, but after a short intervals
of time coloured bars appear horizontally
across it. Putting a piece of red glass in front
of the beam, we have a succession of red and
black bars, the red glass cutting off all the
remaining colours. A piece of green glass
placed in the beam shows green bars, and
so on.

The bars are brighter at the bottom of the
image, which is in reality the top of the film,
for the reason that the film is of a thickness of
'£> 2£> 3i> 4s> 5i> &c-> wave lengths of the
different coloured lights as we go from the
top to the bottom of it. The bars gradually
widen out and become very far apart, until we
see only 3. I now cause a gentle current of
gas to play on the film, and the coloured glass
being withdrawn, we get a magnificent series
of colours whirling one around the other.
Peacock green, golden yellow, azure blue, suc-
ceed one another, and give a most brilliant
effect. All these colours are due to white
light falling on a colourless body.

The next experiment is to throw a small
image of the film upon the slit of the spectro-
scope. We see the spectrum traversed by
black lines curving down from red to blue,
and rapidly shifting in position. These lines
show the colours which are absent in the
horizontal bars of coloured light reflected from
the film, a section of which passes through
the slit.

In this case we have a demonstration that
the colours reflected from the film are not
produced by any conversion of white light
into coloured light, but by the abstraction of
certain colours from the components of white
light.

In the opal we have an example of inter-
ference colours, caused by a thin layer of
material of different thicknesses, which abstract
a certain component of white light in exactly
the same manner as does the soap film. When
we have the light from the varying thicknesses
close together, as we have in the reflected
beam in the patch-forming apparatus, they
have very much the same appearance as has
the opal.

But one more example of interferences,

which is very beautiful, as time will not allow
me to go into the theory of the matter ; suffice
it to say that if parallel lines be ruled on a
surface very close together, and the beam of
light be thrown on them, the " interferences "
are such that pure colours are produced, and
we have a spectrum.

Next let me show you that the colour of
transparent bodies is also due to the abstrac-
tion of colour or colours from the white light.
In a cell I have a liquid which appears
green. A spectrum is formed on the screen
and in front of the slit of the spectroscope
the cell is placed. You will see that the blue
and most of the red is cut off, and that we only
have the green and a small band in the red left
of the spectrum. Recombining the remainder
of the spectrum to form a patch as before, we
have a square of green light, and side by side
with it is the patch formed by the reflected
beam, which is coloured by the light which
has not passed through the prisms, but only
through the cell and the collimator. They are
both absolutely of the same hue, showing that
the recombined spectrum gives the same
colour as the light after passing through the
cell. Repeat the same with a red liquid or a
blue liquid, and we obtain the same results.

A paper is coloured with the green dye
which I had in the cell, and I allow the
patch of white light to play on it, and you
see the light reflected from it is green. In
the path of the reflected beam I place a
cell containing the green liquid, and throw the
patch on white paper. The two patches, viz.,
the white light on the green paper, and the
green light on the white paper, are the same
colour. The white light which penetrates
colouring matter is the same in the two cases,
though when on the paper itself it traverses
the colouring matter twice. This leads to an
important axiom, viz., that the effect is the
same whether the colouring matter is in con-
tact with the paper or at a distance from it, so
long as the eye receives the light which has
traversed such colouring matter. I shall im-
mediately take advantage of this, for I wish to
show you that the depth of colour depends on
the thickness of colouring matter through
which the light passes. Of a double wedge-
shaped trough, half is filled with pure water,
and the other half with coloured water.
Different thicknesses of the blue colouring
matter are passed in front of the slit, and as
the thickness is increased so the spectrum
gets fainter in the blue than in the red.

The patch of white light is next formed, and

the wedge of coloured liquid is again passed
across the slit, and you will see that the
colour deepens as it passes through different
thicknesses. As this is true when the colour-
ing matter is in front of the light, so must it
be true when the colouring matter is in
contact with the papers.

There is another feature which I must not
pass over, i.e., what is known as fluorescence,
and though it does not enter into the effect of
pigments used in water colours, yet it has
much to say to the coloured materials of
every-day wear. One of the most beautiful
examples of this fluorescence is fluorescene.
In the beam of the electric light a jar
of water is placed, and in it is dropped
a concentrated solution of the fluorescene.
We have a fine example of fluorescence ;
the fine threads of liquid as they stretch
towards the bottom appear of a brilliant
green. I take another jar and repeat the
same with quinine sulphate, and we have a
gorgeous blue.

We will endeavour to trace this fluorescence
to its source. I take a piece of card and brush
it over with the solution of fluorescene, and
place it in our colour patch ; the different
colours of the spectrum illuminate one after
the other ; we now can readily see the light
which causes this fluorescence. It is the
green and the blue, but the light reflected
from the fluorescene is of a totally different hue
from the rest of the colour patch. So with the
quinine. We see that when the colour patch
is apparently dark, the paper covered with
quinine shines out with peculiar lustre. The
rays which excite fluorescence in this case are
the invisible rays in the ultra violet. Common
machine oil is fluorescent in the same part of
the spectrum, but shines with a greenish light,
and not blue.

We now come to the point when we must
ascertain the second constant of colour, viz., its
luminosity or brightness. Before showing how
this is done for pigments, it will be necessary
to show you how we can ascertain the lumin-
osity of the spectrum itself. The luminosity of
the spectrum varies greatly in different parts,
the maximum luminosity of the prismatic
spectrum derived from bright lights, such
as the electric light, being in the yellow, and
there is a degradation of brightness as we go
towards each end of the spectrum. Now suppose
we find that the reflected beam of white light,
when the rotating sectors are as widely open
as possible, is slightly brighter than a yellow
patch formed from the yellow of the spectrum

— it is manifest that other parts of the spec-
trum will be dimmer than that. If, now,
in the reflected beam, I rotate the sectors
at less than full aperture, less light will
reach the screen, and it is evident that there
are two parts of the spectrum, one on each side
of the yellow, which will match the brightness
of this degraded white.

In order to make this match, we place the
rod as before in front of the colour patch.
One shadow is thrown on the white screen
by the spectrum colour, and another shadow
is thrown alongside it from the reflected
beam. The white light and the coloured
light, each light up one of the shadows.
The slit in the card is moved across the
spectrum till we find (say) that when in
the blue the illuminated shadow is too dark,
and when the slit is in the green the
green illuminated shadow is too light. It
is evident that at some intermediate place
in the spectrum the coloured shadow is
neither too light nor too dark. This place in
the spectrum is found by moving the slit
rapidly, making the coloured shadow first
too light and then too dark, diminishing the
extent of the oscillations till equality of bright-
ness is seen to the eye. The same procedure
is carried on on the red side of the yellow.
The angular aperture of the sectors is again
altered, and a fresh determination made. Now
the card in which the slit is cut carries a
scale, and by means of a pointer the scale is
read off, which tells us the exact part of the
spectrum where the different equalities of
brightness are established. We then use the
apertures used as giving the relative lumin-
osities of the different parts of the spectrum
as measured, and make such a curve as we
have below.

Fig. 5.

The method, then, of ascertaining the lumin-
osity of a colour depends on the rapid oscil-

lation of either the white or coloured patch
between " too light " and "too dark."

This gives us a clue by which we can measure
the luminosity of a coloured surface in a direct
manner. The rotating sectors in Fig. 2 give
us the means of doing this in an easy manner.
Suppose the luminosity of a vermilion-coloured
surface had to be compared with a white
surface when both were illuminated, say,
by gas light, the following procedure is
adopted : — A square space of such a size is
cut out of black paper so that its side is rather
less than twice the breadth of the rod used to
cast a shadow. One half of the aperture is

filled with a white surface, and the other half
with the vermilion-coloured surface. The
light, L, illuminates the whole, and the rod,
R, is placed in such a position that it casts
a shadow on the white surface, the edge
of the shadow being placed accurately at
the junction of the vermilion and white sur-
face. A flat silvered mirror, M, is placed
at such a distance and at such an angle that
the light it reflects casts a second shadow on
the vermilion surface. Between R and L is
placed the rotating sectors, A. The white
strip is caused to be evidently too dark and
then too light by altering the aperture of

Fig. 7.

sf::^T-::;;i

•$

the sectors, and an oscillation of diminish-
ing extent is rapidly made till the two
shadows appeared equally luminous. A
white screen is next substituted for the ver-
milion, and again a comparison made. The
mean of the two sets of readings of angular
apertures give the relative value of the two
luminosities. It must be stated, however, that

8

if the screen remained unshaded, as repre-
sented, the values would not be correct, since
any diffused light which might be in the room
would relatively illuminate the white surface
more than the coloured one. To obviate this
the receiving screen is placed in a box, in
the front of which a narrow aperture is cut
just wide enough to allow the two beams to
reach the screen. An aperture is also cut
at the front angle of the box through which
the observer can see the screen. When
this apparatus is adopted, its efficiency is
seen from the fact that when the apertures of
the rotating sectors are closed the shadow
on the white surface appears quite black,
which it would not have done had there
been diffused light in any quantity present
within the box. The box, it may be stated,
is blackened inside, and is used in a darkened
chamber. The mirror arrangement is useful,
as any variation in the direct light also shows
itself in the reflected light. Instead of gas
light, reflected skylight, the electric light,
or sun light can be employed by very obvious
artifices, in some cases a gas light taking the
place of the reflected beam.

It will be in your recollection that I said
that the colour of an object depended on the
eye of the observer. Vision, I have told you,
depends on the fact that three colour sensa-
tions are necessary for the normal eye to
see white light. There are in fact, as I have
said, three sets of nerves, one responding to
the blue, one to the green, and one to the red.

Fig. 8.

White Light

White Light

Slue

Green YelloujOrange Red

If one of these sensations be absent, then the
eye does not see white light as we know it, but
as — what would to us be — coloured light. The
above diagram shows the three sensations

derived from Clerk Maxwell's measures. The
top line is supposed to be the spectrum as
the eye sees it, all colours being of equal
value. It will be noticed that at only three
places in the spectrum is the colour due to
single colour sensations, and all intermediate
colours are made up by mixtures of two sen-
sations, the height of the curves added together
giving the height of the straight line parallel
to the base of the curve.

Now, in order to test the eye for colour-
blindness, it is only necessary to get a person
so afflicted, to measure the luminosity of
the spectrum. For evidently, if deficient (say)
in red sensation, the spectrum would begin
where the green colour sensation commences,
and even then the luminosity would be much
smaller, owing to the absence of such red
sensation. Such a luminosity curve is seen in
Fig. 5 (p. 7), and in the same figure is shown the
colour deficiency. It is comparatively easy
to show the colour of the light which colour-
blind people see. If a certain proportion of
the light near the position which the blue
lithium line occupies in spectrum be mixed
with a certain proportion of the green light of
the spectrum near E, and the two be combined
in a patch, the colour of the patch will be that
seen by a red colour-blind person. [This
was shown on the screen, and the vermilion,
emerald green, ultramarine and gamboge
were placed in the mixed light, and the alter-
ation in colour of the pigments noted.] In
the same way the white light which, blue and
green colour-blind see, can be shown.

In measuring the luminosity of the spectrum
you cannot but have noticed that the shadow
illuminated by the white light never appeared
as white, but always coloured. Thus, when
placed in juxtaposition with the yellow, the
shadow illuminated by the white light appeared
bluish; when with the green, reddish; and
when with the blue, yellowish. The colour
given to the shadows illuminated by the white
light is merely the effect of contrast, and is
due to error of judgment by the eye. The
tendency of white in proximity to a colour is to
make it to appear of the hue of the comple-
mentary colour, to which I shall draw attention
in my next lecture.

LECTURE HI— Delivered December 10, 1888.

My first business to-night is to show you
the third constant of colour. You will re-
collect I told you that the hue is one constant.
the luminosity of colour the second, and that
the third is the purity of colour. The purity
of colour is that which is perhaps the most
difficult to measure, but not so difficult to
describe. No colour is pure unless it is un-
mixed with white light. I propose to show
you how you can get colour so impure that
eventually the colour will entirely disappear
and will leave to your eyes only the impression
of white. I think my first experiment will
very likely demonstate this.

The apparatus is exactly that which you
saw before, viz., the colour-patch apparatus.
I am only allowing a small beam of light to
come through the prisms, to get a small round
patch on the screen, instead of the big white
patch square to which you are accustomed.
Now, supposing I pass the slit in the card
through the spectrum, that patch becomes
coloured with any of the colours with which I
wish to experiment. The reflected beam gives
js a large square of white light, which I
superpose over the small coloured patch.
Let us see whether we can extinguish that
coloured light or not. I may take red, green,
or blue, and then if I place the rotating sectors
in front of the coloured beam you will see that
by making the coloured patch fainter it will
entirely disappear. This is the case whether
we have a blue, red, or a green patch. That
the colour is still present I can demonstrate
by cutting off the white light, when you see
the colour on the screen.

The lesson I wish to inculcate is this — that the
blue, green, and red which you saw disappear,
and which were mixed with more and more
white light, are essentially impure colours,
and most impure where the white light is
strongest. It was by this method that origin-
ally the luminosity of the spectrum was
measured. It was seen how much white light
it took to extinguish a colour on a screen, and
according to the white light it took, so the
luminosity was supposed to be proportional to
it. To my mind it is not a very satisfactory
way of testing luminosity, and I think the

way I showed you in the last lecture is far
preferable.

There is another deduction I want to point out
with reference to this, which is of importance
to artists. In water-colour painting it is well
known that in order to get what artists call a
certain amount of warmth in the picture, a
wash of yellow ochre is very often given to the
white paper before it is worked upon. Those
of you who are water-colour painters know
very well that, although you may appear
to have a wash of colour on the paper
when it is moist, yet when it is very dry
apparently there is nothing but white left
behind. The colour is so diluted with white
that it does not appear to the eye, but the
colour is there all the same, and if you
increase but slightly the amount of pigment
the colour may be visible. All the colours you
place on that apparently white paper mix
with the yellow ochre. Remember, then, that
if you have a wash of water-colour on a
sheet of white paper, and it does not appear
to the eye, yet subsequent washes of any
colour will bring out that colour, and in the
case of yellow ochre will give that warmth
which artists so often desire to have upon
their sketches.

Now, then, as to the question of diluting
one colour with another. We have, so far,
only diluted a pure colour with white light;
but in diluting one colour with another we
enter into a region which has been traversed
by a great many experimenters, amongst others
by Clerk Maxwell and Lord Rayleigh, and
there is an immense amount of interest in the
results which have been obtained. Some of
them 1 hope to show you in as simple a manner
as I possibly can. But I want you to recollect
that one can only touch on the fringe of the
subject, as it were, in an hour's lecture.

Let me pass some slits through the spectrum
of this patch-forming apparatus. First we have
a patch of white light, and by a simple means
I propose to show you what colours come
through the slits placed in the spectrum. If I
put another lens L6, Fig. i, in front of the big
lens, which condenses the spectrum to form
the white patch, you will find I can get the

lo

spectrum itself fairly defined upon the same
screen as that on which the patch was formed.
The second lens in reality produces an image
of the first spectrum which was formed in
the plane usually occupied by the focus-
sing screen. Now suppose I pass a series
of slits through the spectrum you will see
the kind of light I am going to use. I
have here two colours, and I will show you
what is the effect of blending those two— green
and red — together; I have only to remove this
lens, and we see an orange patch, I will allow
another colour to come through a third slit
(the card has several), and replacing the small
lens we see the three colours. If I blend those
three I get a green, and so I may go on blend-
ing the colours by passing more slits through
the spectrum. Here I have four, and I dare-
say we shall get a different result again— still
it is a green. Perhaps one of the most inter-
esting ways of showing colour mixtures is to
take away both lenses, and let different parts
of the spectrum pass through the slits, and paint
themselves upon the screen. We begin with
the red, and here we have a red patch. Then
I add yellow which forms orange, and then I
shall add a third patch, and pink is formed,
then green and blue by adding others until we
get nearly a white light in the centre ; so I
can keep passing these slits through the
spectrum, and get many varieties of colour.

Thus we see it is not necessary to have the
whole spectrum in order to get certain coloured
lights. All we have to do is to take certain
portions of the spectrum, and if properly
chosen their combination gives us what we
call a white light. For example, I wish to
show a crucial experiment. I believe every
artist will tell you that the combination of blue
and yellow gives a green. Now I want to
demonstrate that blue and yellow do not give
you a green in accordance with the artist's
notion, but something totally different. I form
my white patch on the screen, as before, and
by means of the small lens put a big spectrum
on the screen. Passing through the small
spectrum two slits, cutting off in the one case
all the spectrum except the yellow, and in the
other all except the blue, which you see on the
screen, and then removing the small lens, in
stead of getting green we get white. Thus it
requires only two parts of the spectrum to be
combined in order to get white. So we see
blue and yellow give white, not green. This is
a crucial experiment, because on this is based
a great deal of the theory of colour mixtures,
and I want you to bear that in mind.

I would once more ask you to remember that
the eye only appreciates three colour sensa-
tions, viz., red, green, blue, and that all the
other colours which are seen by the eye are com-
posed of two or more of these three colour sensa-
tions. I told you the luminosity of the spectrum
was greatest in the green. In the diagram

Fig. 9.

Yellow Orange

(Fig. 9) we have the luminosity curve on a
normal or wave length scale ; the maximum
luminosity is therefore a little bit more towards
the violet end of the spectrum than in the pris-
matic spectrum ; the red component, the green
component, and the blue component of the
luminosity of white light, are shown in the
diagram. These three luminosities together
make up the luminosity of the spectrum of
white light. The blue, you will notice,
has but little luminosity compared with
the green and the red. The luminosity in
the green is far greater than any of the
other two sensations. This I wish to get
firmly impressed in your minds, noting that
the blue is a much less important colour than
green or red ; in other words, it is far prefer-
able to be colour-blind to blue light than to
green or to red light. This, of course, is
founded on Clerk Maxwell's theory, though
the curves are derived from our own measure-
ments. I think the researches which General
Festing and I have made bear out in a very
great measure, although they differ in some
respects in detail, the results which Clerk
Maxwell himself got.

It may be said that we have been dealing
with spectrum colours, and not the colours
of every-day life. Is it possible that if you
are not dealing with spectrum colours that
yet you get the same result? The answer
to this question I will give by experiment in a
very simple manner, and we shall see that we
do get the same result whether we are using
the colours of pigments or the pure colours
of the spectrum. Recollect there were only
two rays combined to form white in the

II

experiment I showed, whereas in the colour of
a pigment you may have a great many colours
combined, although they give the sensation of
one colour to the eye.

The electric light illuminates a circular
aperture, behind which is ground glass, and
by a lens I can throw an image of this aperture
upon the screen. Instead of a simple lens I
have here a lens which is divided into two
halves. The centre of one half lens is raised
slightly above the other. Now every portion

Fig. 10.

of the lens will give an image by itself, and
therefore each half of the lens will give a
separate image, one overlapping the other.
Thus on the screen we now have two images
of the aperture which is in front of the lantern.
If I put a piece of yellow glass in front of one
half of this lens, I form a yellow disc, and
if I put a piece of blue glass in front of the
other I form a blue disc, and where the two
overlap you have the real colour which a
mixture of the blue and yellow lights will give.
You can see that yellow and blue do not make
green, but white.

But the artist, after all said and done,
is not wrong in one way, because he more
often than not mixes his pigments together
and not the colours reflected from them.
Supposing I put the yellow glass in front
of the aperture, I then get two yellow
discs ; if the blue glass be placed in front of
the yellow glass, however, I get two green
discs.

Now let us see why this is the case. I
must come back to my spectrum, to which
we have always to refer when we are dealing
with colour. I will put the two pieces of glass
successively in front of the light passing
into the slit, and ask you to notice what
happens. With the blue glass a great deal
of red is cut off, and a good deal of yellow ;
the blue is nearly as bright as it was before,
and the green is fairly bright. If I substi-
tute a piece of yellow glass for the blue, the
blue is cut off, and the green left almost as
bright as it was before, and the yellow
and red are also left. In the one case,
recollect, we had the blue and the green left,
and the red and yellow cut off. In the

other case we had the blue cut off, and
the green and the red left. If we take one
from the other we get the green left, so that
if I put these t;vo glasses together in front
we ought to get only the green left, which is
the case. Now if I take away the small lens
from the front of the big lens, and form
a patch, we have that patch of the same
greenwhich you saw in our previous experi-
ment. Here, then, we have the combination
of blue and yellow making up the green.
Now for one more experiment in relation
to this. If a blue sector and a yellow
sector be rotated together, and, if what I
have said be true, instead of forming green
they ought to form grey, i.e., degraded
white. Let us see whether it does so. The
two discs are now rotating, and we get what
is not, at all events, far from grey. Thus
we get a blue or a yellow forming a grey
or white, when the blue and the yellow are
each presented to the eye separately.

Now, I shall have to show you why it is that
when they are not presented to the eye sepa-
rately they form the green. This is a yellow
chromate solution in a cell. I place the
chromate solution in front of the lantern ; the
yellow light falls on the blue sector, which is
now at rest, and we have a green. The yellow
is almost unaffected, but there is no doubt
about the blue becoming green. Prussian
blue used in a similar manner leaves the blue
sector nearly unaltered, but the yellow has
now become greeu. If I take a still darker
blue, the green becomes more pronounced
than it was before. You recollect I proved
to you, or tried to do so, that it did
not matter whether a pigment was next to
the paper, or away from the paper, so long
as it was in front of the source of light.
Now in the case before you, when you
mix yellow and blue together, as an artist
mixes pigments, you have one particle of
yellow, say, in front of a particle of blue,
and, therefore, the light which passes through
the yellow is that which reaches the blue
particle, and that they both absorbed I showed
you in the spectrum. The yellow absorbed in
the blue alone, and the blue absorbed in the
yellow and red, green rays would, therefore,
only come through the two.

For the same reason, when I held the
yellow glass in front of the beam of light,
the blue became green, simply because the
yellow glass blocked out the blue, and the
blue particles on the paper only allowed the
green to pass through. This exemplifies

12

again what I told you, that it does not matter
where you have your colouring matter,
whether it is miles away from the paper or
absolutely in contact with it, so long as it is
between the source of light and the paper
itself. But artists, whether they do so know-
ingly or not, employ the method of mixing the
light reflected from the pigments, as well as
mixing the pigments themselves, of mixture
of colour. We know perfectly well that
gamboge and cyanine blue are a very favourite
mixture for greens ; but, on the other hand,
you will find that in some of the most beauti-
ful works of art broad washes, to obtain light
and shadow, are not adhered to, but, as in
the execution of portraits, stippling is resorted
to. Now stippling means that different
colours in fine dots are placed close to
one another, so close that the eye cannot
separate them, and the colours blend one into
the other. Thus, if you have, for instance,
a great many yellow (gamboge) dots distri-
buted amongst a great many blue (cyanine)
dots, the result is exactly the same as
you saw on the screen, viz., instead of
getting a green the general effect is a grey.
This is the whole principle on which stippling
depends, viz., the juxtaposition of very
different colours to give an effect which other-
wise cannot be obtained. Now, the explana-
tion may be new or it may be old, but from
having examined a large number of stippled
water-colour drawings, one can only come to
the conclusion that many of the tender greys
which are often seen in stippled works are
simply due to the fact that you have two or more
colours in dots and fine lines in juxtaposition
one to another, which colours, when combined
in a rotating apparatus such as you have seen,
give the effect of grey to the eye.

I must nowrepeatthe experiment with which I
began my series of lectures, viz., that the three
colours, vermilion, emerald green, and ultra-
marine blue, will give you white ; and I think
that this will be a proof— at all events, a minor
proof— that the three sensations which the eye
distinguishes are green, blue, and red, and not
yellow, blue, and red, as used to be held.
Here we have three colours rapidly rotating,
and those three brilliant colours give the sensa-
tion of white. What proof is there in this
that the three primary colours are red,
blue, and green ? Recollect that I showed
you just now that blue and yellow made
white, therefore red and green must make
yellow. Is that the case ? If that be the case,
I think the point is proved. Let us see whether

such is the case. We will go back to our
apparatus consisting of the half-lenses. There
is a reddish glass in front of one half-lens,
green in front of the other half, the part of the
discs which overlap is yellow ; hence red and
green make yellow. We have already seen
that blue and yellow make white, but it takes
red, green, and blue to make white ; therefore
yellow is equal to red filus green.

Let me further show this. I have a lens in
front of the lantern which forms a slightly
larger image of the aperture than before.
Cemented alongside one another I have three
coloured glasses — green, red, and blue. These,
when placed in front of the lens, and in close
contact with it, will, with a little manipulation,
show a disc of light, something approaching
white. The three colours combine to give this
result.

I am next going to show you how we can
get complementary colours. A patch of white
light is now upon the screen by means of our
much used apparatus. I have a card in which is
cut a wide slot to allow the whole spectrum to
pass through, and suspended from it is a little
prism, which will cut off a certain amount
of the spectrum. The part so cut off will be
reflected on to a mirror, and by means of a lens
will form a patch on the screen. The rest of
the spectrum will go through to the usual lens,
and form another patch of white minus the
colour reflected. The two patches when super-
posed give white, but a rod placed in the front
give two complementary colours side by side.
The complementary colour is that which with
the colour itself will give white. I will cut off
the different parts of the spectrum, and you
will see the real complementary colour. On
cutting out the different colours you will notice I
get almost every variety of hue, and the colours
complementary to them. This seems a very
simple way of getting complementary colours,
and I think it is instructive, as at the same
time it is seen that the background, where the
two overlap, is white.

The next point we come to is one that is
very germane to our subject, and that is how
are we to measure the intensity of pigments in
any satisfactory way ? As far as I know, a
paper which General Festing and I recently
read before the Royal Society explains the
only method which has been satisfactory, so
far, and I hope to show you how that is
done.

The desideratum is to compare the intensity
of any colour of the spectrum which is reflected
from any pigment with that which is reflected

•3

from a surface of white paper. When you get
that you know exactly the colour value of the
pigment.and by certain methods which I shall
show you bye-and-bye, you can at any time
make upon the screen by the spectrum alone the
exact colour of the pigment you have measured.
In order to take these measurements it is
necessary to have two similar spectra one
above the other, and this we get in the follow-
ing manner. Upon the screen a lens forms an
image of an aperture placed in front of the
lantern. Where the rays passing through the
lens cross, I put what is known as a double-
image prism, and by it we get two discs of
light, which will rotate round a centre as the
prism is turned round its axis. This double-
image prism is of Iceland spar, made by Mr.

Fig. ii

iff

i

i

13-

-Ev

Hilger with his usual ability. It gives us the
means of at once getting two spectra one above
the other having exactly the same quality of
light.

In contact with the lens of the collimator, as
it is called (which makes the rays which strike
the prism parallel), is placed the double image
prism ; we thus get two sets of parallel rays,
one set inclined at a slight angle to the other.
Two spectra by this artifice are formed by
the prisms, one above the other, and separated
by a breadth of about one-eighth of an inch.
Passing a slit through those two spectra, the
same colour is cut off from each when the
double image prism is properly in adjustment.
To the card, C, in which the slit is cut, two
right-angle prisms are attached, as shown, and
so adjusted that the beam, r, from the top
spectrum is reflected first by the prism A, and
then by the prism B, on to the screen. A
lens, F, of about two feet focus, in front of B,
makes a coloured patch on the screen, over-
lapping a patch of the same colour formed by
the lens D, which comes from the bottom spec-
trum. By this means we get a parallax of
lights of exactly the same colour, one from

the bottom spectrum, and the other from the
top spectrum. A rod placed in front of the
patch will cast two shadows, one illuminated
by one spectrum, and the other by the other.
The colour, orange, which I propose to
measure, is on one half of this card ; the other
half is left white, the coloured and the white
adjacent rectangles surrounded by a black
mask. In the left hand shadow is the white
card, and on the right hand is the colour
which we wish to measure. In front of the
beam which illuminates the shadow cast on
the white surface are placed the rotating
sectors, and by altering their aperture I can
make the two coloured shadows of exactly
the same intensity. Stopping the motor, the
angular aperture is read off. With another
part of the spectrum exactly the same thing is
done ; by that means we are able to compare
the amount of light which is reflected for the
pigment, and from the white card.

It is on this principle that these particular
colours were measured. To graphically show
their reflective power for different parts of
the spectrum, the following plan was adopted.
Suppose, for instance, that for one spec-
trum to match the other in intensity through-
out its length required an angular aperture
of ioo, and if for emerald green at a
wave length of (say) 5,500, it required an
angular aperture of 45, then in forming
this curve we set off the wave lengths as a
base line, and at 5,500 set up this angular
aperture, which gives us a point on the curve,
whilst the light reflected from the white sur-
face is represented by 100. Thus, at this
point, emerald green reflects only iVbths of
this particular light. By taking numerous
other parts of the spectrum you are able to
build up a curve, which is an absolute measure
of the light reflected from the pigment, as
compared with that reflected from the white
surface. I want you to notice how very
peculiar are the curves of the yellow pigments.
There seems to be very little difference in the
intensity of light reflected from them, but
to the eye they appear of decidedly different
hues. It is just these little differences in the
curves which make up the difference in the hues
which are so noticeable. Again, I want you to
notice cobalt. You see what a large proportion
of red there is in cobalt, and what a little red
there is in Prussian blue, Antwerp blue, indigo,
or French ultramarine. If we take a line tan-
gential to the bottom of these curves, and
parallel to the base line, the height of this
tangent shows the amount of white light which

»4

Fig.

I 2

ioo White

90
BO

70

00

2

70
6(1

y

^^

7

'

r^w

..

40
86
10

»'

40

'4-

\ i

/....

\ i

7

t—

20

ii-'i 7

/

in

1

/

0

B

f

\ '"HrT

1 T

()

4L P 5°°\ «°W1W J ° -""Vol -'S° fXS~7!° 7°oo '-

O £O0

Fig.

■3

fOOWMt*

90
80
70

eo

50

90
70

50
40
JO

/

^S

30

/

V

1~

j-P

/

V

20

y

"~~ '

JO

2

10

0
4C

a
i

f

£

0

V

r

0

»f^ ' S0°

c c

«3 ^ 3

«s

I

i. Indigo. 2. Antwerp Blue.

4. French Ultramarine.

Fig. 14.

3. Cobalt*

1. Emerald Green. 2. Chromous Oxide. 3. Terre Verte.

Fig. 15,

100

<

m

0 Wh

te

10 !

700 (V/i te

90

f

\1 .-"

90

no

90

-,

10

70

eo
so

—

J^

70

HO

70

'

70

—

~

v/

if

~

-

so

40

50
40

—

J

50

v:

/?

j-
4

l^***1

—

1

„.-.

le"

20
10
0

8
2

'1/
1~"

0

h

1=

— '-

D

1

B

I

f

f

£

0

C

S

I

0

10

ETJ"°

"i^koyi^1

| ^J" |)"« «u 7ooo *

kzj -•-?°t ?ji <

i

! J J |E I 111?

I II ^

■a- -o' J

6

i

03 ca £ v- L. ?^ Ot~ i5

1. Vermilion. 2. Carmine.

3. Mercuric Iodide.

1. Gamboge. 2. Indian Yellow. 3. Cadmium Yellow.

4.

[ndi

in I

led

<

• Y

L'llu

If 0

Jire

is reflected from the pigment, and is a measure
of its impurity. For instance, if you take the
curve of cobalt, you will see it has about 3
per cent, of white light mixed with it ; whilst
in the tint measured of Antwerp blue there is
about 23 per cent, of white light mixed with
the true colour of that pigment. You will
notice that, in all cases, a certain amount of
white light is reflected from the pigments, and
therefore not one is really a pure colour.

Now I want to show you another method, and
one which has never been exhibited before, by
which we can obtain the intensity of colours in
a very simple way. I use for convenience sake
a rather short focussed lens in the camera, as
I want to form rather bigger patches of mono-
chromatic light. Behind those black discs of the
motor is a disc of white card, and I am going to
measure the intensity of spectrum colour re-
flected from a coloured disc by a novel method.

I can put any coloured disc I like in front of the
sectors, and in contact with them. I rotate the
sectors in the coloured patch, and I can alter
the amount of white on the larger disc until I
get it to match the luminosity of the colour in
the centre. Knowing how much black has to
be mixed with white, in order to bring the tint
reflected from the colour in the centre to the
same value as that reflected from the rotating
black and white, I can readily determine the
intensity of the light reflected. (Several
colours were measured in succession, in the
manner described.)

Next on my programme is the method of
producing on the screen the exact colour of
any pigment. The researches of Dr. Russell
and myself on various pigments which have
faded in light would be of little value, unless
in, say, a thousand years' time those colours
could be reproduced with the same accuracy

'S

with which they were measured. We have a
means by which we can, without having the
pigment itself, absolutely reproduce that colour
from a card such as this. I will show you on
the screen how it is done.

Fig. 16.

40 1 4 «ls50

D Prussian Blue

If we mark off the scale of the spectrum
along the radius of a circle, and draw circles
at the various points of the scale from its
centre, and from the same centre draw lines
corresponding to the various angular apertures
of the sectors required at the various points of
the scale to measure the light reflected from a

pigment, the point where one of these lines cuts
the circle drawn through the particular point of
of the scale to which the aperture has reference,
gives us a point on a curved figure. Such a
figure, when rotated in front of the spectrum
in the proper position, will cut off exactly the
right amount of the spectrum at each part of
it to give the colour required. I will show
you one or two of these colours, and by that
means you will see that we have literally tem-
plates by which our successors in science will
be able to reproduce the colours which we
have measured in our experiments, and to see
whether any alteration has taken place in
those particular pigments we have used, and
which we propose to leave, either at the South
Kensington Museum or elsewhere, for the
benefit of those who come after us. (The
colours of various pigments of blue sky, gold,
and gaslight, were reproduced on the screen.)
By cutting out templates like these, and in
your laboratory carefully making the neces-
sary adjustments, you can always reproduce
on the screen any colour you may have
measured, and if you use the light in which
the colour has to be viewed, be it sunlight,
gaslight, starlight — whatever light it is— to
form the spectrum, you will get on the
screen the colour as it would be seen in that
light.

i6

LECTURE IV.— Delivered December 17, 1888.

We have, in the three preceding lectures,
seen how colour is produced, and how it
depends on three factors— the kind of light
falling on to the substance, the kind of sub-
stance itself, and also the eye of the observer.

To-night I have to endeavour to explain in
one hour what ought to take many more hours,
how light acts in altering the colour of pig-
ments through what I may call mechanical
means. A water-colour picture (I shall deal
only with such) is exposed in the ordinary
atmosphere of a room. Sometimes that room
is without a fire ; consequently the atmosphere
becomes more or less damp, and all absorbent
objects, such as paper, take up moisture. At
other times, when there is a certain amount
of warmth, the moisture which it would
take from the air is less ; so that a picture is
exposed to alternations of damp and dryness.
Dr. Russell and myself concluded that it
would be quite fair for testing the stability of
water-colour pigments if we exposed them to
the ordinary outside atmosphere, and then
traced the amount of fading which took place,
remembering this, that a picture inside a room
would certainly be more stable, supposing
moisture had anything to do with fading.

We prepared tubes, as in Fig. 17, perfectly

Fig. 17.

(n\

open at each end, but with a small cork in the
unbent end, the cork being pierced with a
large hole. A current of air could pass
throughout the tube when hung on a bar by
the bend and exposed to the sunlight. Inside

each tube were strips of paper, covered with a
pigment which had been tinted by hand in
graduated tints (such as you see here). There
were eight tints in all. One such strip was
placed in one end of the tube and another in
the other. The lower half of the tube was
covered with an opaque covering so as to pro-
tect it entirely from the light, and the other was
left free to the sunlight and the light from the
sky. By-and-by I shall show you why it was
we deliberately chose sunlight to which to
expose our water colours. From theoretical
considerations we arrived at the conclusion
that fading would take place in a shorter time
in sunlight than it would do if we exposed it to
the open sky alone.

In such a series of tubes, containing in all
somewhere about 100 colours — 39 being simple
colours, the others being mixed colours —
were exposed. The first reading of the amount
of fading was taken in August, 1886, or after
four months' exposure, and we found that in
many of the colours fading had taken place
to a certain extent, although perhaps not to so
large an extent as might have been antici-
pated. From time to time after that date the
tubes were examined, and the amount of fade
ing noted, our notes showing the deepest tint
which was visibly acted upon. Finally, we were
obliged to conclude our experiments, owing to
the impatience of certain gentlemen who were
anxious to get the results we had obtained,
apparently for their own advantage rather
than for that of the public. We thus stopped
our first series of experiments in March
of this year, or after these tubes had been
exposed about one year and nine months
outside my laboratory at South Kensington.

In these tubes, then, we had the ordinary
atmosphere, to which moisture and air had
free access. If the tube got the least bit
heated a current passed through it, much in
the same way as would be the case in a chim-
ney. The great point to settle was whether
the fading which we knew must take place,

17

and which we subsequently noted, was due to
the air itself, or to the air £lus moisture, or to
the moisture alone. In order to test that, we
passed air over various drying materials, dried
the papers and tubes very thoroughly. The
papers were then placed in straight tubes
sealed at one end, and when filled with dry air
the other end was sealed off, and they were
exposed to sunlight, one paper being shaded
from it as before. In the case of the open
tubes, we found out of 39 simple colours only
12 were not acted upon; and in Table I. you
have the 39 single colours in the order of their
fugitiveness.

TABLE I.

Carmine.
Crimson lake.
Purple madder.
Scarlet lake.
Payne's grey.
Naples yellow.
Olive green.
Indigo.

Brown madder.
Gamboge.
Vandyke brown.
Brown pink.
Indian yellow.
Cadmium yellow
Leitches blue.
Violet carmine.
Purple carmine.
Violet carmine.
Purple carmine.
Sepia.
Aureolin.
Rose madder

Permanent blue.
Antwerp blue.
Madder lake.
Vermilion.
Emerald green.
Burnt umber.

Yellow ochre.
Indian red.
Venetian red.
Burnt sienna.
Chrome yellow.
Lemon yellow.
Raw sienna.
Terre verte
Chromium oxide.
Prussian blue.
Cobalt.
French blue.
Ultramarine ash.

Vermilion is ordinarily supposed not to
change at all, but, as a matter of fact, it
does change, and in every sample there
has been a little blackening. Those last
on the list, yellow ochre, Indian red, and
so on, show no change whatever after being
exposed to as much sunlight as there was in
one year and nine months. They remained
perfectly unaltered, and, if you begin with rose
madder (all below which may be said to be
practically permanent) you have a very good
gamut on which an artist could work in water
colour.

In the closed tubes with dry air, out of
thirty-eight sample colours which were ex-
posed, twenty-two were not acted upon, so that
it is evident that moisture had something to
do with the fading of some.

TABLE n.

Name of Colour. Dry Air.

Carmine Faded to 7.

Crimson lake Gone to 5.

Scarlet lake Faded and darkened.

Vermilion Gone black.

Rose madder No change.

Madder lake No change.

Indian red No change.

Venetian red No change.

Brown madder Faded to 4.

Burnt sienna No change.

Gamboge Faded to 3.

Aureolin No change.

Chrome yellow No change.

Cadmium yellow No change.

Yellow ochre No change.

Naples yellow No change.

Indian yellow Faded to 4.

Raw sienna No change.

Emerald green No change.

Terre verte No change.

Chrom. oxide No change.

Olive green No change.

Antwerp blue Faded to 3.

Prussian blue Faded to 5.

Indigo blue Faded to ".

Cobalt blue No change.

French blue No change.

Ultramarine ash No change.

Leitches blue Faded to 5.

Permanent blue No change.

Payne's grey No change.

Violet carmine Faded and brown.

Purple carmine Faded.

Purple madder Faded to 4.

Sepia No change.

Vandyke brown V. si. faded.

Burnt umber No change.

Brown pink Faded to 4.

Note. — SI. means slightly ; V. si. means very slightly ;
No. 1 is the faintest tint.

The next series was interesting. The same
kind of tube was taken and filled with
hydrogen, and also with as much moisture as
the hydrogen and paper would take up. The
tubes were then sealed and exposed to light
approximately for the same length of time as
the other tubes. As a matter of fact, out of
thirty-six colours twenty-two remained un-
changed, the same as before. Hydrogen, I
may say, is practically an inert gas for this
purpose, as we proved subsequently.

Then we come to the most interesting series
of all, when we excluded air and moisture from
the water colours. We took exactly similar
tubes, dried the papers very carefully indeed,
dried the tube, inserted the papers, put a

i8

Sprengel pump to work, and made a vacuum,
and then when the vacuum was very complete,
sealed off the top and exposed them.

TABLE III.
Name of Colour. Vacuum.

Carmine No change.

Crimsonlake No change.

Scarletlake No change.

Vermilion Gone black.

Rosemadder No change.

Madder lake No change.

Indian red No change.

Venetianred No change.

Brownmadder No change.

Burntsienna No change.

Gamboge No change.

Aureolin No change.

Chromeyellow No change.

Cadmium yellow No change.

Yellow ochre No change-

Lemon yellow No chanSe-

Naples yellow No chanSe-

Indianyellow No change"

Raw sienna SI. darkened.

Emerald green No change.

Terre verte No chanSe-

Chrom. oxide No change.

Olivegreen No chanSe-

Antwerp blue No change.

Prussianblue V. si. faded.

Indigo blue No chanSe-

Cobalt blue No change-

French blue.... No change.

Ultramarine ash No change.

Leitchesblue No change.

Permanent blue No change.

Payne's grey No change.

Violet carmine SI. darkened.

Purple carmine SI. darkened.

Purple madder V. si. gone.

Sepia SI. faded to 6-

Vandyke brown No change.

Burnt umber No change.

Brown pink No change.

Indian yellow and rose madder No change.

Rose madder and raw sienna No change.

Raw sienna and Venetian red No change.

Vermilion and chrome yellow More yellow.

Burnt sienna and Naples yellow .... V. si. faded.
Indigo, Indian yellow, raw and burnt

sienna No change.

Indigo and gamboge Gone blue.

Prussian blue and gamboge Gone green.

Burnt sienna and Antwerp blue Gone red.

Raw sienna and Antwerp blue Gone brown.

Prussian blue, raw and burnt sienna,

and Indian yellow Gone brown.

Piussian blue and burnt sieuna .... Gone brown.
Indigo and Vandyke brown Faded.

Name of Colour.
Prussian blue and burnt sienna ....

Prussian blue and raw sienna

Indigo and raw sienna

Indigo and burnt sienna

Indigo, raw and burnt sienna

Prussian blue and Vandyke brown . .

Indigo and Venetian red

Prussian blue and Indian red

Indigo and Indian red

Prussian blue and crimson lake ....
Antwerp blue and crimson lake ....
Indigo, Venetian red, yellow ochre . .
Prussian blue, yellow ochre, Venetian
red

Note.— SI. means slightly;
No. i is the faintest tint.

Vacuum.
Gone brown.
Gone red.
No change.
No change.
No change.
Gone brown.
No change.
Gone red.
No change.
Gone pink.
Gone pink.
No change.

Gone red.
means very slightly ;

We here arrived at the very interesting fact
that out of thirty-nine simple colours which were
exposed, only five were acted upon in the very
least, and the amount of change was so slight
that you might almost say every colour re-
mained perfectly unchanged in vacuo. The
five that were changed were vermilion (which
went black to a very slight extent), raw sienna,
Prussian blue, purple madder, and sepia. We
are apt to look on sepia as one of the most
permanent pigments ; as a matter of fact it is
fugitive in ordinary air, and those who have
examined sepia drawings made in the early
part of the century will see there has been
certainly a distinct fading of those drawings.
By the process of exhaustion, we arrived at
the fact that it requires both moisture and air
to cause the fading of these pigments.

Now the question arose— Would heat with-
out light cause the fading of pigments? Where
they were exposed to sunlight it might be
surmised, perhaps fairly, that in the sunlight,
which we know has a heating effect, the fading
might be due to this cause in the open tubes.

This could not be the case in the closed
tubes, as in them the colours did not fade.
To test the action of heat alone, we took
tubes in which the papers were sealed up
with moist air, and exposed them for three
or four weeks, at the temperature of boil-
ing water, in the dark. There was a certain
amount of fading in these colours, but^I
need scarcely say that the fading was small,
and also that the temperature to which they
were exposed was something far beyond that
to which colours in our open tubes were sub-
jected. If you put a thermometer up one of
the open tubes when it is in full sunshine, the
difference between the temperature of the air
inside it and the air outside only varied between

'9

three and four degrees. That was simply due
to the fact that there was a draft created up
the tube, as already pointed out.

But another point, and a very fair point for
the critics to take hold of, is this. It is all
very well to say light alone causes fading, but
how about light and heat together, would
not the heat aid the light ? This possible
criticism was combated, I hope, in a suc-
cessful way. A certain series of pigments,
washed on paper, were taken and exposed
on a vessel containing boiling water ; similar
papers were exposed to the sunlight free,
that is to say, without the presence of
the boiling water. In some few cases the
fading was rather more rapid, in others less,
and you will very readily see why, in some
rases it was rather less rapid. You require
moisture plus air in order to cause fading,
and if you heat the paper of course you take
away part of the moisture — one of the agencies
which are conducive to fading. But the differ-
ence between those exposed on boiling water,
and those exposed without, was so small that
you might take the action of light plus heat
as equivalent to the action of light alone.

There was another experiment we had to
try, and that was as to the rays which
caused the fading. I have shown you in my
previous lectures that beautiful band of colour
we call the spectrum. I daresay you noticed
that the beam of light which passes through
the slit to form the spectrum is uncommonly
narrow ; for accurate experiments we should
not use it more than i-ioooth inch wide, and
that has to be spread out into that band of
colours, so that really the light which strikes
upon the screen is very feeble indeed. If we
had attempted to expose some of these pig-
ments in the spectrum, we should have had to
expose them for some thousands of years, and
as life is shorter than this, we thought it was
better to take some other means of arriving at
the conclusion as to what coloured rays were
the active agents ; so we adopted a method
which, perhaps, may be called crude, but I do
not think it is crude when you know how you
are going to work. We exposed slips of paper
beneath coloured glass — red, blue, and green,
and also white. Here are some of the pig-
ments which were actually exposed. We
got the results as shown in Table IV. (p.
20).

We exposed 39 or 40 simple colours besides
compound colours, and I want you to notice
how very few faded in the green, in the red
less than the green, but a very great many more

under the blue glass than under either of the
other two, You will see that the blue and the
white were almost equally effective. Had a
certain proportion of the blue rays in the white
light been cut off by the glass, practically
those two columns, white and blue, would
have been identical. Under the red and
green glasses the fading of the few pigments
which succumbed was so small that it re-
quired a practised eye to distinguish it.

Now I will read you some conclusions we
came to with regard to the fading of water
colours : — " Mineral colours are far more stable
than vegetable colours, and amongst those
colours which have remained unaltered, or
have very slightly changed after an exposure
to light of extreme severity, a good gamut is
available to the water - colour artist. The
presence of moisture and oxygen are in most
cases essential for a change to be effected,
even in the vegetable colours. The exclusion
of moisture and oxygen, particularly when
the latter is in its active condition, as experi-
ments to be described in our next report show,
would give a much longer life even to these
than they enjoy when freely exposed to the
atmosphere of a room. It may be said that
every pigment is permanent when exposed to
light in vacuo, and this indicates the direc-
tion in which experiments should be made for
the preservation of water - colour drawings.
The effect of light on a mixture of colours
which have no direct chemical action on one
another is that the unstable colour disappears,
and leaves the stable colour unaltered appreci-
ably. Our experiments also show that the
rays which produce by far the greatest change
in a pigment are the blue and violet com-
ponents of white light, and that these, for
equal illumination, predominate in light from
the sky, whilst they are less in sunlight and
in diffused cloud light, and are present in
comparatively small proportion in the artificial
lights usually employed in lighting a room or
gallery."

Now, it has been said that moisture
and oxygen are essential for the fading
of water-colour pigments. Is it possible
that they can fade without light ? I have
here a stream of oxygen passing through this
tube in which are some papers coated with
pigments ; half of each paper has been
damped and the other half is dry. In con-
nection with this tube is an ozone generator,
and a Ruhmkorff coil produces ozone, or
the active state of oxygen, which is said to
be particularly present near the sea. In this

20

TABLE IV.

Purple Madder. . . .

Antwerp Blue ....

Leitches Blue

Violet Carmine. . . .

Payne's Grey ....

Indigo

Prussian Blue ....

Rose Madder

(2 experiments.)

Brown Pink

Crimson Lake ....

Vandyke Brown . .

Vermilion

Carmine

Gamboge

Indian Yellow

Sepia

Burnt Sienna

Faded to 2 ...
No experiment

SI. faded

Faded to 1 ...
Faded to 1 ...
No experiment
No experiment
SI. bleached .

Antwerp Blue . .
Prussian Blue . .
Purple Madder..
Burnt Sienna

Gamboge

Indian Yellow . .
Vandyke Brown
Brown Pink
Crimson Lake . .

Carmine

Vermilion

Rose Madder .
Violet Carmine .

Payne's Grey .
Sepia

No experiment
No experiment
No experiment
Darkened . . .
No experiment
No experiment
No experiment
Become lighter
No change . . .

Colours Mixed

No experiment
No experiment

Bleached

No change . . .
No experiment
No experiment
No experiment
No experiment
No experiment
No experiment
Blackened . . .

Blue.

Faded to 1

Faded

SI. faded . .
Faded to 1

Bluer

Faded to 1
SI. faded . .
SI. faded . .

Faded to 3

Faded

Faded to 1

V. si. darkened . .

Faded to 3

Faded to 1

No change

Become lighter . .
No change

Green.

Darkened

Blue

SI. faded
SI. faded

SI. faded

Red.

Darkened

SI. faded.
V. si. faded.

SI. faded.

SI. bleached

Bleached to No
and darkened
2 and 3

Bleached to 1 . . .

Lighter

'I

with Chinese White.

Bleached ... .

Bleached

Bleached

No change . . .
SI. bleached .
SI. bleached .

Bleached

Bleached to 3 .
Bleached to 3 .
Bleached to 3 .

Blackened under 1
and 2

V. si. bleached

Same as under white
glass

Become bluer ,
Lighter

Become bluer

SI. faded.

Note.— SI. means slightly ; V. si. means very slightly ; No. i is the faintest tint.

21

frame [shown] you have a series of colours
which have been exposed to moist ozone.
A gTeat many are bleached entirely, thus
proving, if you have ozone and moisture
together, you get a bleaching without the
presence of light at all. Here are some
papers which were exposed, I think for about
ten minutes, to moist ozone before the
lecture, and you will be able to see the
amount of fading that has taken place. In
the example of indigo the bottom part was
damped and the top part left dry ; the first
half has faded, the other has not. In carmine,
too, you will see that where it has been
damped the colour has entirely gone ; the dry
part is much less changed. We come then to
the conclusion that oxygen and moisture are
sufficient for the fading of water-colour pig-
ments, and that it is not absolutely necessary
that there should be light present in order tha
this fading may take place. Now, as before
said, you are supposed to have more ozone at
the seaside than inland. It is therefore a
matter for consideration whether it may not
be the fact that water-colour drawings fade
more rapidly near the sea, where there is more
ozone present, than they would do inland.
That is a question I am not going to touch
upon now, but when we make a subsequent
report no doubt that will be brought forward
prominently.

We have seen the results of light, and I
wish to show you how it is that light acts upon
matter. Matter is formed by molecules, or
very minute particles, far beyond the vision of
the best microscope that was ever made ; you
can only reason and argue about them from
the circumstantial evidence which nature from
time to time puts before us. The molecules
themselves are composed of atoms. Thus,
in the molecule of water it is supposed
there are two atoms of hydrogen and one
of oxygen. Each molecule is presumably
of identical shape, and size, and composition.
There has been a certain amount of evidence
brought forward that perhaps some molecules
of the same kind of matter are rather bigger
than others, but to my mind such evidence is
incomplete, and I cannot accept it. At any
rate, as a rule, we may take it that the size of
the molecule is the same for the same species
of matter ; that, for instance, all water mole-
cules are the same size and composition, as
are those which go to form the molecules of
these pigments we are considering.

I want to give you a homely notion of what
a molecule is like, and how we may suppose

the atoms vibrate. I have here a little cell of
water, through which a vertical beam of light
can be thrown, and again be deflected to the
screen. A lens forms an image of the sur-
face of the water on the screen. Around
this cell of water I can cause a current
of electricity to pass through a coil of
wire. When you have a current passing
there is a certain amount of magnetism
produced which repels magnetism of the
like kind. I have here some little needles
which are magnetised, and inserted in small
bits of cork by one end, the same poles being
in the corks. The corks will float on the
surface of the water, thus supporting the
needles. Now, if we float some of these little
magnets in the water, they will repel each
other and tend to go farther apart, the reason
being that magnetism of the same kind repels.
Now if I turn on the current in the wire
passing round the cell you will see that they
are found to approach one another, and as I
move the wire up and down, they alternately
approach to and recede from one another.

You must recollect that at the same time that
these atoms are vibrating one towards the
other, the molecules themselves are vibrating
to and fro from one another, so that we have
vibrations of the molecule and vibrations
of the atom. Now I have told you that
the waves of light vary in length ; the red
waves are the longest, and the blue waves
are the shortest, and as they all travel at
the same speed, the time of oscillation of
the red wave is longer than the time of oscil-
lation of the blue wave. We may take it that
the oscillation of a molecule is slower than
that of an atom, and it is much more likely to
be isochronous with a wave of red light than
it would be with one of blue light. Similarly,
the waves of blue light are much more likely
to be isochronous with the time of oscillation
of the atoms than the molecules, and, as a
matter of fact, such we find to be the case.

Now let me give you another homely example
of what we mean by oscillation on the part of
an atom or a molecule. You can quite under-
stand, I think, that if you have a body oscil-
lating to and fro from another body, both of
which attract one another, if you increase the
oscillation, a time comes when the attraction
between the two is so small that there is a great
tendency for them to fall apart. If there is
another body at hand which is willing to take
up one of those atoms — which has a great
affection for such atom— it will take hold of it,
and bring it to itself. The bob of this pen-

22

dulum, which is of iron, is supposed to be an
atom swinging to and from another atom, and
some three inches behind it is fixed a magnet.
By puffing with my breath at the same rate as
the pendulum vibrates I can increase the am-
plitude of that oscillation to such an extent
that, eventually, the attraction of the mag-
net for the bob of the pendulum is greater
than the force of gravity, and it reaches the
magnet and is held by it. This very simple
experiment teaches us a lesson. Here we have
an atom swinging away, we will suppose, from
another atom of something. My breath timing
itself with the swing may be taken as the oscil-
lation of the waves of a ray of light. The waves
of light perpetually beating on the atom will
increase the amplitude of swing of that atom
so greatly that if there is another body near it
which will take up the atom, it leaves the
original atom for it. When such a re- arrange-
ment of atoms takes place, we say that a
chemical action has taken place, that is,
that light is able to decompose a molecule by
robbing it of some of its atoms, and giving
them to another body. We get, then, by the
decomposition new molecules formed, and con-
sequently new matter, and such a new body
may be in the shape of a faded pigment.

Throwing a spectrum on the screen, I put a
layer of pigment in front of the slit, the light
passes through it, and we get, as you saw by
previous experiments, some colour taken away
from the white light, and other colours left
behind. In the case before us the red and the
green and the blue are left, but most of the
green is cut off. I will put another substance
(permanganate of potash) in front, which gives
a beautiful absorption spectrum, and there are
a number of dark bands in it. If I take the
iron salt which I used in the experiment in
measuring the quantity of light which came to
galleries of South Kensington, you see that it
cuts off the blue almost entirely. You can see,
then, that these various solutions cut off a
certain amount of colour from the spectrum.
Now the question is this, what becomes of the
rays that are cut off ? The whole principle of
the chemical action, and the heating effect
of light upon pigments, is answered by the
answer to that question. It is this. Where
you have an absorption of light, there you have
work done upon the body on which it falls.
In that permanganate of potash, for instance,
which you saw gave a fine spectrum — the
rays missing, which gave the black spaces,
were doing work on it. They were heating up
the permanganate of potash, or chemically

changing it into something else. You cannot
have work done on any body unless there is
absorption by that body. You understand
what I mean by absorption — the cutting off
tne light by the body. When there is chemical
action taking place, the work done is the
swinging the atoms away from each other,
when heating effect takes place the molecules
are swung further apart from one another.
I hope I have made clear to you that my view
is that when you have chemical action
taking place, the absorption takes place
in the atoms ; when it is a heating effect
which takes place, it is the molecules which
are acted upon, and made to jostle each
other more vigorously. As far as chemical
action is concerned we have a very familiar
example in photography. I am going to
develop a spectrum for you. This has been
done before in this room by myself, but as
there are many here who have not seen the
experiment, I think it might be as well to
repeat it. [The photograph of the spectrum
was developed.] The paper was covered with
bromide of silver, and if I place a slab of
bromide of silver in front of the slit, you will
see that the absorption exactly agrees with
the locality where chemical action has taken
place.

Now I have another experiment to show you,
and that is the heating effect of radiation. I
have here a little instrument called a thermo-
pile which consists of strips of two metals
soldered together at one end. If the junction
be heated, a current of electricity will pass
through wires attached to the other ends when
joined ; and if a galvanometer is in the circuit
the galvanometer needle will be deflected.
By means of a mirror attached to the needle,
which will reflect the light from a lamp on to a
scale behind, I can show you the deflection.
I now form a very small spectrum, and cause
different parts of the spectrum to fall on the
junction of the metals. The needle deflects very
slightly with the blue, showing that the heating
effect is small ; as it gets towards the green and
travels into the yellow the deflection is greater,
and when we get into the red portion itis again
more. At the very limit of the red the deflec-
tion is greater still, and outside this colour and
in apparent darkness we see that the light on the
scale travels further still, showing an increased
heating effect. Thus an invisible part of the
spectrum which lies beyond the red heats this
junction of the two metals more than any part
of the visible spectrum. We have here a proof
that not only the rays which cause the sensation

why we made experiments with light ///wheat.
I told you we exposed the pigments on paper
against a vessel of boiling water to see
whether the decomposition was accelerated.
It was possible that these dark rays might

Fig. i 8.

have heated up the paper to such an extent
that the heating action aided chemical decom-
position by the blue rays which we found most
effective.

Now I want to call your attention for a minute
to this diagram (Fig. 18), which represents the
heating effect of different sources of light.
The height of the curve is a measure of the
heating effect. The curves on the right hand
of the dotted line show the energy of the dark
rays, whilst on the left the heating effect
of the visible spectrum is shown. The heating
effect (which is a measure of the energy) of
the dark rays is very much greater than the
heating effect of the rays which lie in the
visible part of the spectrum. I want to call
attention to the solar curve ; you see what a
peculiar jagged curve it has. The jagged in-

in the visible spectrum, and a very large effect
in the dark part of the spectrum. The same
applies to gas-light and candle-light.

In estimating the chemical action of radia-
tion on a body, there are two factors to be con-
sidered, the intensity of the radiation acting,
and the time during which it acts. This is
very important. Thus if a certain coloured sur-
face be exposed to a radiation whose intensity
we may call ioo, which bleaches it in one hour,
then if a similar coloured surface be exposed
to intensity I, it will require ioo hours' ex-
posure to effect the same amount of bleaching.
There is an idea abroad that if the light be
very feeble, no matter what length of exposure
be given, it will not affect a bleaching ; this,
however, is not the case. The same propor-
tion of the total energy absorbed by the body
which, with intense radiation, effects chemical
decomposition, on exposure to feeble radia-
tion is doing the same kind of work. We
may say, briefly, that the deductions from
scientific experiments lead us to believe that if
strong light causes fading, a feeble light will
do the same, if the exposure to it be prolonged.
The pendulum experiment, I think, fully illus-
trates what I mean. I will give you a rather
fuller illustration, however. The amount of
increased swing that light can give to the
atom means an increase in the amplitude of a
wave, and the amplitude of a wave in the sea
is the height from the crest to the trough. Sup-
pose we have a heavy church bell hung without
friction on its supports, and without any resist-
ance to its motion, and suppose it to make a
complete swing once a second. Suppose also
that at the end of the bell-rope there was a small
horizontal plate, and at intervals of a second
a thousand grains of water fell from a fixed
height on the plate. The bell would gradually
oscillate ; the bell would be like this pendulum,
and finally it would oscillate so greatly that
the bell would ring. Now, if instead of 1,000
grains falling from the same height, we had

*J** i *- i > j

24

but one grain falling every second, it would
take 1,000 times longer before the bell rung;
or if the weight were i-ioooth of a grain, it
would take one million times as long before it
rung. The work done by the dropping water
may be looked upon as the work done by the
amplitude of the wave of light on the atom,
as it, too, moves without friction and without
resistance.

As to the light which pigments in water-
colour drawings are ordinarily exposed, a few
remarks may be made. There is no doubt
that pictures as a rule are carefully protected
from direct sunlight, but it is nevertheless true
that the greater portion of the light they
receive is reflected sunlight. On a bright day
the clouds reflect sunlight, and on a dull day

Fig. 19.

Sun Ugkt

urn

80
80
70

9

\"

50
40
30

S

J32

J.

10

1

L-

H

j_

19 1

< 1

7 16 15 14 13 12 11 10 9 8 7 8 5 4 3 2 .1

the diffused light is also sunlight, which is
reflected according to the laws of geometrical
optics, and a large per-centage reaches the
earth from the clouds. There is also a fair
proportion of light from the sky ; this is
bluer than that reflected or diffused from the
weakened sunlight. In cases where the
windows of a gallery are in vertical walls,
which is the most ordinary case, and have an
interrupted view of the horizon, the blue light
reflected is comparatively small, the light
near the horizon being distinctly more like
sunlight than that nearer to the zenith. In
galleries lighted like those at South Kensing-
ton the light comes from above. The artificial
lights to which water-colours are exposed are
gas-light, electric arc and incandescent lights.
The first and last are very deficient in blue
rays (see Fig. 19). You see, for instance, how
deficient gas-light is in blue rays compared
with sunlight or blue sky. Blue sky, you will
notice, possesses hardly any red light what-
ever.

Now I think you will see why we were
justified in exposing our pigments to sunlight

instead of skylight. If you know the amount
of blue rays that are in any particular light,
and the amount of work such rays are capable
of performing, it is quite fair to translate the
action which one source of light has upon a
pigment into the amount of effect from a
different source of light. That is to say, if I
know what action sunlight will have upon a
pigment, then from diagrams such as the
above, we can calculate the amount of action
which skylight will have, and also the gas-
light, whether the intensities of the total light
are the same or different.

It is now necessary to explain to you how it was
that we came to use three kinds of glasses for our
experiments to see which part of the spectrum
was most effective. As a preliminary, I should
like to show you that a pigment may be very
rapidly acted upon, although apparently per-
fectly inappreciably to the eye. I have here
two transparent films which were treated with
two dyes. Those two films were exposed behind
a transparent cross to the electric light for ten
seconds, and were then floated over with a silver
salt and a developer. From previous experi-
ments we knew that where these particular dyes
had been acted upon by light there silver would
be deposited on them, and I think you will
see that these two show that such is the
case. The first film was dyed blue origin-
ally, and you will see where the light has
acted the silver has deposited upon it. Here
is another film, originally red, on which the
same thing occurs. I want you to lay this
thing to heart. Do not think that because an
object does not visibly fade in a year that,
therefore, it has not begun to fade at all. A
year to one pigment may be the same as 30
seconds to another pigment, and if you expose
pigments for a year, which will only fade as
much as that particular pigment faded in 30
seconds, then, applying this silver salt, you
will probably get exactly the same action after
a year's exposure as you did with that shorter
exposure on the more fugitive colour.

One more experiment. Here I have a piece
of paper which has been impregnated with a
silver salt, and has also been dyed with a
colour. I want to show you that the smallest
action of light on this particular colour will
cause the reduction of silver salt. I am going
to expose the paper to the spectrum for 10
seconds. [The paper here was developed.]
You see in this case that we have a black band
corresponding to the absorption spectrum of
the dye with which it was dyed. This band is
absent where the silver alone without the dye

*s

is acted upon. The dye has been acted upon,
and thus caused a reduction of silver to take
place where it has been altered, although such
alteration is perfectly invisible to the eye.

Now I can show you why we chose red,
blue, and green glasses for our experiments.
I want you to notice the different parts of the
spectrum that these particular glasses absorb.
Passing the glasses through the spectrum, the
red glass allows the red, and a little bit of
yellow and green, to pass (see Fig. 20). With
the green glass a great deal of the red is cut off,
and all of the violet. With the blue glass you
will see that a great deal of the red is cut off.
Thus, in the case of the blue glass, we have
the blue principally left, in the case of the
green the green principally left, and in the case
of the red glass we have the red principally left.
Now, suppose I put the red glass and the blue
glass together, what would happen ? We only
ought to have a bit of the red of the spectrum
left, and if I put the green glass with these

Fig. 20.

1

OU

1

40

1

1 1

! llpkt

-;q

ru

'J"

rt

'1.

>'fv

itnjin

1J

&

30

'&

"v.

Y

20

.C;

•*/

's

\

,-r

'>

^y_'

s_

i*S

f.

<fe. 1

as

r4

z

1

a i

7 1

, i

i, 1

* 1

3 ;

e '

1 1

j 1

'

! <

' '

1

1

'

I

o <B E = S V

s « 2 J

we ought to have nothing left, which is the
case. In other words, the rays transmitted by
those three glasses make up the whole spec-
trum, so that when using those we are utilising
the rays of every part of the visible spec-
trum. It was for this reason we choose those
particular glasses through which to expose
our pigments. Fig. 21 shows the heating
effect of the light after passing through the
different glasses. Notice the dark rays.
They are nearly entirely transmitted through
the red glass, very slightly through the blue
and green glasses. Had the fading of the
colours we have examined been due to the
dark rays, it ought to have been shown beneath
the red glass far more than under the green or
the blue glass. This was not the case, as a
reference to Table IV. will show. We may,
therefore say that the blue, violet, and ultra
violet rays are those which are by far the most

active in producing a change in the pigments
with which we have experimented.

I may say a word or two about the exposures
we gave, and the results deduced. We ex-
posed between May, 1886, and the middle of
August, and we found that during that time
these pigments had 705 hours of bright sun-
shine. That bright sunshine we reduced to so
much sky light, and the total amount of
effective sky light received in that time was
1,700 hours. Allowing for overcast skies, and
for blue sky light and sun light, we find that
these pigments had an average of 2,225 hours
of average of blue sky — or, roughly speaking,
2,500. We may now go a step further, and
calculate the amount of illumination which a

Fig.

21.

7t

?!

W

./

.53

-57 0

*7 >?

i0/

*7#

/ a

iM,

y .

M

a

'

/

-A

>-

-■ •

to

~ 1

V

G

jf_

t

':■

'"/

%'J

V

,

picture shown in a gallery such as those at
South Kensington would have during the
same period. There is no direct sunlight,
and making calculations from photometric
observations, and seeing how much light
came into the gallery, compared with that
outside, we came to the conclusion that to
have the effect on these pigments in the
galleries which took place in the sunlight, 32
years would have to elapse, supposing the
light was always equally bright to that between
May and August. But we know it is not
equally bright, and we came to the conclusion
that it would take 100 years to get the very
little fading such as we got outside the labora-
tory in four months.

Now let us see what would happen to a pig-
ment supposing it were exposed to gas light.
Calculating the amount of blue light in such
light, and also the total illumination in the

2 6

gallery in question, we found it would require
at least 2,000 years of continuous exposure for
the same amount of effect to take place as oc-
curred in the four months of sunlight exposure.
After an exposure of one year and nine months,
we have the astonishing result that to obtain
fading of the same amount in the colours ex-
posed, it would have taken 485 years of average
daylight in the galleries to have got that
amount of bleaching. If we had exposed it
continuously to gas light, the time required is
almost incredible to believe, viz., 9,600 years.
With these facts before us I think you will say
it is not at all surprising that we chose to
use sunlight instead of any other source of
light for our experiments. I am afraid that
neither Dr. Russell nor myself are good

for 480 years, and therefore we preferred to
use the shorter time of one year and nine
months in order to arrive at the conclusions
we did.

The methods of measurement that I have
brought before you are for the most part
new, but I believe they can escape any very
serious criticism. The details of many of the
experiments, from which our calculations have
been derived, have been published in various
papers laid before the Royal Society and
the Physical Society. I may say we have
the greatest reliance on the accuracy of
them.

I have now finished my course of lectures,
and I have only to thank you for the great
attention which you have paid to me.

PRINTED BY W. TROUNCE, 10, GOUGH-SQUARE, FLEET-STREET, LONDON, E.C

SOCIETY FOR THE ENCOURAGEMENT OF ARTS, MANUFACTURES,

AND COMMERCE,

CANTOR LECTURES

PHOTOMETRY.

CAPTAIN W. DE W. A B N E Y,

C. B., F. R. S.

Delivered before the Society, Afiril 2nd, qth, and \btk, 1894-

LONDON :
PRINTED BY WM. TROUNCE, 10, GOUGH-SQUARE, FLEET-STREET, LONDON, E.C.

PHOTOMETRY.

Lecture I. — Delivered April 3, 1894.

The lectures on photometry are not given
with the idea that they will be of practical
value for the measurement of gas light. There
is excellent literature on the subject, part of
which I shall have to refer to during my
course. What I have undertaken in these
lectures is to endeavour to give an idea of
the general principles of photometry, almost
restricting myself to the scientific aspect of
the question. Photometry, in its broadest
sense, is the measurement of light, at least,
so we must think, from its derivation. Now,
the light measured may be light coming from
an object, or from a self-luminous body, such
as a candle or the sun, or it may be the light
transmitted through objects. In the second
case, if an appropriate screen be used to
receive the light, we are in reality measur-
ing the illuminating power of the source of
light, rather than of the light itself. Hence,
almost as much depends upon the screen on
which the light is received as on the light
itself. A screen is usually what is called
white, and by white is meant a screen wrhich
reflects every colour equally well ; but, I would
remark that in London the white may become
imperceptibly brown, and such colour may
interfere materially with accurate results.
But the photometry that I am alluding to
not only includes the measurement of the
illuminating power of light, but the measure-
ment of the light transmitted through bodies
of various kinds, when they are transparent,
like plain glass, or translucent, like ground
glass or paper. The requirements of the
candle-power of gas I shall not enter into,
as it is a subject which others than myself are
much better fitted to deal with.

We may take it, I think, that the first
matter we have to consider is the light we
have to use as a standard. Parliament, in its
wisdom, in i860, pronounced its standard of
light to be the light of a candle 6 candles to
the pound, each burning 120 grains of sperm

per hour, and this is at present the only legal
standard known in England, though why, in
the name of common sense, such a definition
has been continued our legislators alone can
guess, when it has been proved to be so
faulty. The standard of light for France is
the Carcel lamp, which is equal to about
0/5 candles. Now, a light from a candle
is a very pretty thing theoretically, but
practically it is anything but practical, as
it has the unhappy knack of burning inaccu-
rately, particularly when one is anxious to
shield it from draughts. Heat affects the
rapidity of combustion, and if it be confined,
and no proper access of air be given it,
its light may be most irregular. We have to
remember that part of the energy of combustion
is taken up by melting the sperm, or wax, or
whatever it may be, and if the surrounding air
be heated the wax is at a temperature nearer
its melting point than it should be when at a
normal temperature. When the melting point
is attained the liquid is decomposed and the
flame results, and there is more liquid to be
vapourised and vapour to be improperly con-
sumed than in the normal state.

I show you a trace made by photography of
the light from a candle burning under normal
conditions. The light was admitted through
a slit to sensitive paper, and a fresh portion
of paper was continually being exposed. You
will now see the irregularity of the burning.
Of course, by taking several candles the varia-
tion is not so great, but even then you have
to be sure that the proximity of the candles
to one another does not alter the rate of
burning.

An Argand burner, however small, will not,
during a long series of experiments, differ
1 per cent, in light value. Here we have a
proof of this. This small paraffin lamp was
allowed to burn for three hours, and you will
see that the band it makes is perfectly uniform
in appearance, and when the measurement is

made of the blackness produced by it on the
photographic paper, it proves my statement is
correct.

The apparatus by which these diagrams
were made is a very simple one. It consists
of a clockwork arrangement drawing a pulley,
which pulley is in connection with a drum,
which can rotate on its axis. Round this
drum is placed sensitive paper, and a box,
with a long' slit in it, covers the drum. The
light is placed opposite the slit, which is
covered by a moveable lathe, in which is an
aperture of a convenient width. As the drum
moves, this aperture moves across the slit,
and so we have a corkscrew band of exposure
produced. With some clockwork the motion
is regular in its irregularity, and every tooth
of the train can be counted on it, by noting
the bands of varying exposure, and for this
reason the clock was at one time abandoned,
and the smooth motion of the sinking of the
height in subsiding water was substituted.
This gave very good results, but for my pur-
pose the clockwork was sufficient.

The sources of light 1 have mentioned are
what may be called feeble sources of light,
and cannot be used when a body is fairly
absorptive, if the transmitted light is to be
measured. We want in such a case a
stronger source of light, and one which is
practically constant. Such a source of light
we have in the electric arc light. If we
project upon the screen an image of the points
where the positive pole is slightly behind the
negative pole, with a fairly long arc, we
become aware that there is a central part,
which is higher than any other [shown]. It
comes from a depression in the positive pole,
and for the last eight years I have been in the
habit of using this as a source of light of uni-
form intensity, and many hundreds of measures
have proved it to be so. This, as several years
ago I pointed out, was due to the fact that the
temperature of this spot was that of the
volatilisation of carbon. It is an intense
light, and may be taken as 50,000 A.L. per
inch of surface, and very useful for a great
many purposes, as we shall see as we proceed.
Now we call all these lights which I have men-
tioned white, but it is quite evident that there
is white and white if all these be white. I
believe myself that Mr. Lovibond's definition
of white is a good one, which is the light which
is seen in a white fog about midday, and if we
compare this light with any other we shall, I
think, comeback to it as being a very practical
white light. Now the electric light is not far

from this quality of light, and as such is very
useful in comparing the transparency of objects
by what is approximately daylight. We can
measure the light of each part of the crater
passing through a small hole.

We can at once see the difference between
all the ordinary lights by a simple experiment.
This box is divided into partitions with tissue
in front, and in each partition we have a
different source of light — a partial gas jet,
an Argand gas - burner, a candle, and a
paraffin lamp. It will be noticed that the
light enclosed in a chimney is much whiter
than those burned in free air, but you will
also see that all these lights have various
depths of yellowness when compared with
the electric light. It is quite evident that
even supposing they gave the same illumina-
tion, that they are not all fit for standard
lights. I take it that a standard light in
photometry must always have the same quality
of light as well as the same quantity of light.
Now we can, by appropriate means, make the
electric glow-lamp light of the same visual
intervals as a gas jet. The one before us is so,
but it is evidently not of the same quality. One
of the very best tests that we can make of
ascertaining whether any difference in quality
exists is to see if, when they are equally strong'
visually, they give the same photographic
results. [An experiment was made with an
electric light and an amyl acetate lamp, in
which both were made of the same visual
intensity, but photographically they differed
materially.] You see that the amyl acetate
lamp is decidedly the worse photographically.

Perhaps I can show you why this is. I take
an incandescent lamp, and cause it to glow :
it goes red, to begin with ; then I increase the
current, it gets yellow, then whiter, and so
on, till it is nearly white. 1 cannot make it as
white as the arc light, for the reason that, as
the temperature increases, the fusing point of
carbon is reached, and that, as I pointed out,
is the temperature of the crater of the arc
light. These temperatures, however, are sub-
ject to different amounts of energy expended
upon them ; and here I have a diagram, show-
ing how, with an increased energy expenditure
on the same filament — that is, with an in-
creased temperature — the different rays of the
spectrum are altered in proportion. These
diagrams are taken from measures made
with a linear thermopile, moved through the
spectrum. You will see that the higher the
temperature, much more rapidly do the rays
of high refrangibility increase.

60

Iff - Vfk

TKHnIt

=Jm'l)>lv.

125 ..

.. 5,900

13 ••

.. 14.650

5

.. 7.250

17 ..

.. 20,750

9

.. 9,900

21

.. 27,500

These numbers apply to both diagrams, and
in Fig. 2 the numbers attached to the different
curves, are those which are attached to the
abscissa? in Fig. 1 .

Let me show an experiment. I will balance
an electric light against the amyl acetate
lamp, and expose a piece of paper to its
action. I will increase the temperature and
balance again, and expose another portion of
the same paper to its influence for the same
time. Notice, please, the difference in the
two. You will find that the highest tempera-
ture filament is much more "photographic."
By this means all lights, which are due to the
incandescence of solid particles of carbon, can
be tested as to quality. Make them visually
equal, and then see if they are photographi-
cally equal. For my own part, I believe that
a knowledge of the photographic value of light
is essential in the near future ; for I cannot
help thinking that there will have to be a
registration of photometric values for record,
independent of the eye, and this must be by
photography.

For this purpose the photographic value,
and the visual value of every light used, will
have to be known and carefully recorded. We
shall see soon how these records can be

s

>

j/A

A*/

&

y

Wjl

s\k

W

,(

V> .3

\Q,

I2J

17,

'

,

'V

£*&

6^-

s£

?n

a

.

-*

if" 1 —

O I 2 3 4 5 6 7 8 9 10 1,1 I? 13 14 15 16 17 18 19 20 21 22 23 24

turns of screw

Fig. i.

The value of the abscissa; in Fig. i (in wave
lengths) is as follows : —

10 20 30 40 50 SO 70 80 00 100 110 120 130 140 150

THE NUMBERS

Fig. 2.

utilised, and become of permanent value in
themselves, being capable of being measured
at any date after being made, and re-
measured if required. I throw on the screen
the photographic values of a candle, an
amyl acetate lamp, a gas jet, a paraffin lamp,
and an arc light — all made of the same value
as a candle visually [shown]. You will see
that they vary enormously, and the scale of
opacity below, which was made by exposing
different parts of a plate to a steady light for
different times, gives us a means of comparing
one with the other.

I have said that all lights which are due to
solid particles of incandescent carbon can be
tested by means of photography, and I have
shown you the deposits which certain lights
cause on a photographic plate. There can
now be but little doubt that a luminous candle
flame is as much due to solid incandescent
particles as the glow-lamp we have been using.
The final proof has been long in abeyance,
but 1 think no doubt now can exist regarding
it. First of all, if we examine the spectrum of
the luminous part of the flame, we find that it
is continuous, though occasionally a bright
line of sodium in the orange puts in an appear-
ance, but it is of no account. Now any light
which emits a continuous spectrum must be
due to a solid or liquid body in a state of in-
candescence, or to a gas in similar state, but
nndir great pressure. The flame is certainly
not liquid, nor is it gaseous under pressure.
It seems, therefore, the light must be due
to solids, and those solids must be so small

that even a microscope of low power will
fail to distinguish them. This fact (if it be a
fact) enables us to put the matter to a good
test. If we project a beam of light against a
cloud of small particles, the rays which are
most refracted (the violet and the blue) are
violently scattered in all directions, as Lord
Rayleigh has shown should be the case
theoretically, and the greater the number the
more yellow is the light coming through them.
There is one peculiarity, however, about these
scattered rays, viz., that those which are scat-
tered at right angles to the beam are what are
termed polarised in one direction — that is, that
if they pass through a Nicol's prism turned in
one direction, they become quenched, whilst
they will pass through readily if the Nicol be
turned in the direction at right angles. You
will see what I mean by the scattering by an
experiment which I now make.

Fig. 3.

If to this clear solution of hyposulphite I
add a few drops of hydrochloric acid, it
becomes cloudy, owing to precipitation of
fine particles of sulphur. I allow a beam
of light to pass through the solution before
I make the addition to the screen, and then
add the HC1. The light becomes yellowish
and then reddish, as the number of fine
particles increase ; that is, the more par-
ticles the redder it becomes, and the more
light is scattered, as a look at the cell
testifies.

By precipitating mastic in water we get the
same results. Here is some which has stood
two years or more, and while it is turbid the
beam of light passes freely through it, but
scatters light on each side. Now, if I pass
that broad beam of light first through a Nicol's
prism, turned in one direction, and then through

the solution, the path of the beam is clearly
visible, but if I turned it in a direction at
right angles it is at once quenched. Its ex-
istence, in the first case, and its absence, in
the second, shows that the light, coming at
right angles to the beam, is polarised. This
you can see for yourselves, at least most of
you who sit in the proper direction ; but for
the sake of those who do not I take two photo-
graphs, one with the Nicol turned, so that
the polarised light passed, and the other when
it was turned, so as to present the beam. You
see the result.

Now let us apply this to the small carbon
particles. If a beam of intense light, such as
that coming from a small image of the sun, be
thrown on the flame of a candle, a white beam
of sunlight should be seen on the flame, and a
beam of white "light passing through the
flame. Unfortunately, I have not the sun at

Fjg. 4.

my command here to-night, so I cannot show
it, but you may take my word for it that such
is so. Sir G. Stokes examined this white beam
in a position at right angles to its direction,
and found, by means of a Nicol's prism, that
it was completely polarised ; that is, that
when the Nicol was turned in one direction,
the streak of white light in the flame dis-
appeared altogether. This establishes the
fact that the luminous part of the flame is due
to small particles, independently of any other
proof. It appears to me, therefore, that one
is correct in stating that the bright flames are
due to measurement carbon. Into the theory
of flames I will not further enter at the present
time ; this is enough for my purpose.

In case there be any doubt amongst you, I
will show you some photographs of the pheno-
mena I have taken.

Fig. 5 is a photograph of an Argand gas-
fiame, on which the rays of the sun, collected
by a lens of about 8-inch focus, were concen-
trated so as to pass along part of the circum-
ference of the cylinder. The Xicol prism was
turned in such a direction that the scattered
rays would be unaffected in the left-hand

Fig. 5.

photograph, whilst it was turned at right
angles to the first direction for the right-
hand photograph. In the left-hand figure the
track of the beam is readily seen, whereas any
trace of it is absent in the right-hand figure.
Fig. 6 is the same, but the electric arc light
was used in place of the sun. The results are
the same.

Fig. 6.

Fig. 7 shows the results when the beam from
the electric light is passed through a candle
flame. In the one figure a broadish white
band is seen, whilst in the other it is absent.

We are now in a position to see why it is
some flames are whiter than others. When a
chimney is used with gas, for instance, we
find that the illumination is whiter — bluer, if

you like the word better. The function of a
chimney is to supply air to the flame, ample
room being found through interstices to allow
as much air as is needed to be drawn up into
the chimney. In the case of hollow flames,
such as an Argand burner, not only is the
air admitted to the outside shell of the flame,
but also to the inside. The consequence is
that the small particles of carbon are heated
to a higher temperature, as they are in the
blacksmith's forge by the bellows, and they
then emit a whiter heat before they are con-
verted into carbonic acid. When one has a
smoky lamp, there is one of two things hap-
pening— either the supply of air is insufficient
to the chimney, or else the flame is too high
and the sudden access of cold air chills down
the incandescent carbon particles till they
become black, and smoke results. One of the

Fig. 7.

most instructive experiments as to the need of
air and warmed air to a flame is shown by
lighting a paraffin lamp. It is an orange
smoky flame, but directly you place the
chimney on it the light whitens and the smoke
ceases.

I should here like to correct a very common
notion which exists regarding the blackening
of ci ilings by gas flames. As a matter of
fact, the carbon in a gas flame ordinarily is
totally converted into carbonic acid. It is the
ascending current of heated air that catches up
the floating motes in the room and dashes
them against the ceiling, to which some cling
tenaciously, and gradually the blackening is
encountered. A friend of mine lately put up
the electric light in his house, and placed the
glow-lamps close to his ceiling. He was
astonished to find that the ceiling above them
blackened to an extent which reminded him of
gas. It was the current of warm air which

caused the blackening. Similarly, hot-water
pipes will do exactly the same thing. Heated
air will ascend, and when it ascends it carries
the motes and particles with them. In South

Kensington Museum, ceilings which adjoin hot-
water pipes blacken quicker than where there
is gas, the reason being that the volume of
heated air is so large.

Lecture II. — Delivered April 9, zSgj.

I omitted, from want of space, to say in my
last lecture that the fact that a flame viewed
end on is from 10 to 35 per cent, less luminous
than when viewed sideways. Fig. 8 gives a
measurement if taken with a flame at different
angles to the screen according to Mr. Dibdin.
The variations in the light of a burning candle

Edge to Bar.

■jog 0} oOpj
Fig. 8.

has been shown you, and I think that for
scientific working it must be dismissed as un-
worthy of serious consideration. There are
only three what I may call feeble light
standards which I shall refer to, viz., the
amylacetate lamp, due to Hefner Alteneck ;
the pentane illuminant, and the ether illu-

minant. I put the amylacetate lamp first
not because of its superiority, but because it
requires such little manipulation. This is a
lamp which is a great favourite of mine
because it is so accordant in its results. It
consists of a tube of German silver, 8 mm. in
diameter, and 25 mm. high. The flame is 40
millimetres hisfh, and when it has been burnt

Fig. 9.

for five minutes the flame remains of a con-
stant height. It burns amylacetate, but it is
not necessary to use this compound, as any
similar one will consume as well. Hefner
Alteneck gives a Table of the results of the
different compounds and their comparative
luminosities : —

Constitution.

Per cent,
of

Boiling

carbon.

point.

69-7

c.

1

646

62-1

[22=

1 :i

Il6

980

Intensity
of light.

Time for the
combustion of 1

gramme of
the substance

Carbon

consumed
in 100

. 1 ■

Valerate of Amy] C,0 H20 0a

Acetate of Am) 1 ' C, H,, (>.

Foimiate of Amyl C„ H1£ O.,

Acetate of Isobutyl C0 H, „ 02

Formiate of Isobutyl C ., H10 0„

I -03

TOO

roi
099
097

430

372
373
355

0-I()2

o-i 66
0-163
0-163
0-166

The drawback to this lamp, as originally
constructed, is that the metal takes a green
deposit, which is tiresome ; if it be plated
with silver, this disappears.

Dibdin's pentane Argand, which burns pen-
tane, is the next one to refer to, and is
the lamp which appears to me most per-
fectly to utilise the pentane, employed as an
illuminant, in a simple method. Pentane is
a hydrocarbon of the paraffin series, but is not
perfectly pure at all times. The illuminant is
air passed over a carburetter containing the
pentane. The height of the flame is 3 inches,
-fVths of which are cut off by a screen at the
top. By these means a standard flame is
obtained, which is equal to 10 candles. The
great point in this is that the height of the
flame does not affect the result, at least it
does not to the eye. Temperature has no effect
on the result, as Mr. Dibdin has thoroughly
tried.

The next standard is a very simple one,
introduced by Mr. Dibdin more especially
for photographic purposes ; ether, instead of
pentane, is burnt in a pentane lamp, and gives
a very fine light. Photographs taken with
hese two lights at different heights of flame,
but of the same visual intensity, do not give
quite the same photographic effect, so that
there is a deviation from the definition of per-
fect standard.

We have seen what kind of a light we must
use for photometry as to quality and quantity.
Now we come to photometers. The photo-
metry we will first consider is the comparison
of two lights together. How are we to com-
pare two lights ? There is one evident way,
and that is to place side by side two white
surfaces which are illuminated by the two
lights. This is the principle of Rumford's
photometer and nothing else. We are usually
told that it is the method of shadows — the
comparison of shadows one with the other.
Now it is nothing of the kind, it is really the
illumination of a surface by two distinct lights,

the one illumination being not interfered with
by the other, and this is secured by making
one light cast a shadow of a rod on the screen,
which is illuminated by the other, and this
last light to cast a shadow of the same rod at
a different place, which is illuminated by the
first light. These two illuminated surfaces
can be made to touch by moving the rod
or the angle of the light, and by various
plans these can be equalised in brightness.
No less a distinguished authority on photo-
metry than Mr. Dibdin, in an excellent book
he has written, says, although this method has
certain advantages, " the method is one
which few practical photometrists of the
present day would venture to adopt." Well,
I am a tolerably practical photometrist myself,
and T must confess I prefer it to any other kind
of photometry, as it is simple, and very few
errors can creep in if one is ordinarily careful,
which is more than can be said of some others,
as we shall see. One error that may be
met with is that if the lights make a great
angle with each other, and if the screen is not
placed at right angles to the line bisecting
the angle, an error may creep in.

Let me show you this experimentally, and
this experiment really demonstrates another
mode of photometry.

This white cube is placed between two lights,
one of the right angles of the cube being
towards you. I place a square aperture in
front, so that it is bisected by the edge.
The cube is rotated round that edge as a
centre, till the two sides appear equally illumi-
nated. The reason of the equality of illumi-
nation is quite plain. It is because the side
nearest the light is skewed at a greater angle
than the other to it. If we have a diagram, we
shall see why this is. In Fig. 10 (p. 8) A n and
B A are the two sides of the cube illuminated by
rays K and R. It is evident that the side a B
will not receive so many rays as BC, in fact,
the amounts are measured by / q and mn.
If the lights are unequal, of course when the

intensity of the one multiplied by ^ t/ is equal
to the intensity of the other multiplied by m?i
the two will equal. The intensities, where
a balance is struck, is found by taking the
cosines of the angles through which the cube
is turned.

IE

Fig. io.

We now see that light illuminating a sur-
face varies as the cosine of the angle through
which it is turned. If it be turned 50 more
towards one light than the other it is evident
that we shall get a variation. The amount
would be as 1 to '996, or an error of -nfW>
or t4t7- If it were 10° it would be 1 to "984,
or -jiJ-go- or -gL part, which would be appreciable.

There is still one more error which might be
felt, and that is that the eye receives more
light when the angle which the screen makes
with the eye and the source of light is greater
than a right angle (see r and r1 in Fig. 10).
This must always be the case, but what may
be called the difference in the specular reflec-
tion is so small for ordinary angles, that it is
of the same order as that given for the wrong
placing of the screen, and becomes practically
negligible.

For great accuracy the illuminated shadows
should touch, and if the lights be not too
broad, there is no difficulty in causing this to
be done ; sometimes, however, a white line or
a black line will separate the two owing to the
penumbra of the shadows, and then making
the illuminations of the two strips equal be-
comes more difficult. As the black line has
greater contrast to the two illumined surfaces
than the white line has; the former is the
worst kind of line to put up with.

The next method that is adopted is what is
known as the Bunsen method. It consists of
equalising the brightness of a greased spot in
the centre of a paper disc, or its total dis-
appearance. The principle on which this is
based is the translucence of the spot. If as
much light goes through the spot (if perfectly

made) from one light as goes through from the
other, the spot is equally illuminated through-
out its thickness, and appears the same white-
ness as the paper. If it be greater on one
side it will appear dark on one side, and lighter
on the other. It is evident that with such a
method every suspicion of stray light must
be rigidly excluded, unless it be exactly the
same on both sides of the disc, and only that
coming directly from the sources of light
utilised. Light reflected from the sides or bars
will give fatal results as far as accuracy is
concerned. I have met with some instruments
in which reflections seem to have been encou-
raged rather than allayed. To my mind the
method should not be accepted except in the
hands of those who are thoroughly practical
and scientific. I show the design of a Letheby
photometer, kindly lent me by Mr. Sugg.
[The instrument itself was in the lecture-
room, through Mr. Sugg's goodwill.] The
grease spot is viewed on both sides by inclined
mirrors, and when the grease spot disappears
on both sides, or at all events appears to
equally dim on each side, the light illumi-
nating the spot may be said to be equal.

There is one thing to be noted, and that is
that very much depends upon the kind and
amount of grease, and the kind of paper,
employed. I have made a good many grease
spots in my day, and I have found the sensi-
tiveness of the method vary considerably
according to the attention paid to these
details, but I have abandoned the method
in my laboratory, except under special circum-
stances, in favour of the old Rumford method.

Mr. Dibdin, in his work, says : —

"When first setting up a disc for use, special
experimental readings should be taken ; and if any
material difference is found between the indications
when one side or the other is turned towards the
standard flame, it should unhesitatingly be rejected, as
no amount of after allowance can compensate for the
trouble and doubt arising from contradictory results.
The disc should be clean and perfectly free from
scratches or other markings of any kind ; it is but sorry
economy to work with a defective instrument. The Gas
Referees went so far, a short time back, as to run a
new disc, to be used every week. As, however, a
good disc, when taken care of, will last much longer
than that period, the point has not been insisted
upon ; but that is no excuse for the continued use of
a defective one, which should be instantly destroyed
as soon as detected."

We see from this that a disc photometer is
open to a very grave objection, and it is for
this, if for no other reason, that I prefer the

Rumford system, where there is no liability
to err on this matter. A modification of the
Rumford method of shadows is that employed
by Prof. V. Harcourt. He casts his shadows
on ordinary printing- paper, rendered partially
translucent by a wash of spermacetti dissolved
in petroleum. Instead of a rod, and about three-
quarters of an inch from the paper, he places a
brass screen, having two rectangular apertures
cut in it exactly their own breadth apart. The
two lights are placed at equal angles on each
side of the line perpendicular to the screen,
and the illuminated shadows are caused to
just touch one another. It will be noticed

slightly altering the values that should be
obtained.

Before quitting the subject of the Bunsen
method, I ought to mention that in photometry,
for the grease spot is sometimes substituted a
star of thin paper, sandwiched between thicker
paper; that is known as a Leeson disc, and
has been much improved by Mr. bibdin.

Methven proposed to use a slit placed in
front of an Argand gas-jet as a regulator, if I
may call it so, of the quantity of light issuing

Fig. i i.

that really there are four shadows illuminated,
one from one light being touched by the other
two, and the fourth falling on an opaque or
black space. A great advantage of this plan
is that they are looked at from the back of the
screen, no rod being between the eye and the
screen. If two lights of approximately the
same colour are looked at, the fact that the
light has to traverse the paper is of no moment,
though, when coloured lights have to be com-
pressed, there is a danger of absorption

Fig. 12.

on to the grease spot. This appears at first
sight an admirable arrangement, and it would
answer well if the grease spot were always kept
at the same distance from the source of light,
but when it is moved, an error, though it may
be very small, must be introduced. An Argand
flame is practically a hollow cylinder of liglit,
of acertain thickness (Fig. 1 1). Asyouapproach
the light the section of the cylinder varies,

twMf . otM^fel dflllf

and consequently the quantity of light falling
on the spot must vary beyond what it should
do. It may be remarked that putting aside

Fig. 13.

this error the measurements are made from the

slit and not from the source of light, which is
a decided advantage. Messrs. Hurter and

10

Driffield have to a large extent got rid of
this light and employ a flat flame, of large size,
as the source of light, and use a small square
aperture in front of the flat side. As the
section of such a flame appears to be uniform,
the inaccuracy of measurement introduced is
done away with. In reference to this, it may
be interesting to show that in an ordinary
flame the light varies in intensity at different
points. This can be done well by means of
photography, reducing the exposure each
time. Fig. 12 (p. 9) is an Argand burner flame,
Fig. 13 (p. 9) a batswing, and Fig. 14 an ordi-
nary candle.

It will be seen that in the candle flame we
have an almost expected result. The Argand
gas is more surprising. The batswing gas is
perhaps the best, as it shows that in the wing
used the intensity remains almost constant.
I think these photographs will demonstrate to

Fig. 14.

you that if the quantity of light to be admitted
to a screen is to be determined by an aperture,
the burner should be of the batswing type.

Before quitting the subject of photometers,
I must introduce to your notice the radial
photometer of Dibdin (Fig. 15). The diagram
almost explains itself. The object of the
photometer is to measure the illumination of
a flame in all directions. It will be seen that
the arm which carries the light to be tried
remains always at the same distance from the
screen. The screen itself is so arranged that
its surface bisects the angles between the lines
joining the two lights and itself — a most
necessary thing, when Fig. 10 is taken into
consideration.

We have now to turn to the method of
judging the equality of light ; that is, how the
eye can best appreciate the light.

We are told very frequently that the eye can

appreciate about the ,'0th part in the intensity
of light, or, say, 2 per cent. There is a story
told of a celebrated witness who, when asked
whether such and such a thing was the case,
said : — " Yes and No." Now if I were asked
the question as to whether the above limit was
true, I could safely answer in the same terms.
First of all let me show you an experiment,
which will prove that this limit is both under-
stated and also overstated. I have on this
screen a variety of greys between black and
white. We can now see them all, and the
difference between them. If I turn down the
light, a great many of these appear the same
tint. If I turn on to them a very strong electric
light, those not neaily white, when looked from
where I am, appear white, and it is only when
there is considerable black in the shade of
tint that they appear grey.

Fig. 15.

As a matter of fact, there is an intensity of
light, in which much smaller differences than
the jVth can be perceived. I believe, for my
own part, that, when the light is suitable, a
difference of nearly ^J-j is recognisable. But
it is not necessary that the eye should be so
sensitive as the above, so long as proper pre-
cautions are taken in balancing the light. If
we balance from "too light" and then from
" too dark," the mean will be fairly exact, and
probably not be far off the truth by a good deal
less than 1 per cent. But there is another
plan, which is better still, and that is by rapid
oscillations in intensity on each side of the
true point. This is difficult with many photo-
meters, but not with all. When this plan is
adopted, supposing we are using the shadovv
method, the two shadows appear to wink, and,
when exactly balanced, this winking stops.
It is curious how, without this artifice, readings,

which can be proved to be palpably wrong,
are made. For instance, when one shadow
is intensely darker than another, the eye of the
observer will fail to see it, when the alteration
is made slowly. If the eye, however, has a
rest, by looking away at some black object,
the inequality of the shadows will at once be
seen. This cannot happen when the method
of rapid oscillation is adopted.

What the cause of this may be is not abso-
lutely proved. When the eyes look at two
objects (spots or shadows) the images of the
two are projected on different parts of the
eye, these portions get fatigued, and the
longer they are looked at the greater the
fatigue. The brightness of the two gets lowered
and they gradually approach one another.
When the system of oscillation is adopted,
though both images are lowered in tone, yet
there is a constant brightening and dimning in
both, not sufficient rapid to make each of them
practically uniform in a tone midway between
the two, but scintillation is produced. We
can see how the eyes can be fatigued by a very
simple experiment. I will throw a bright patch
from the electric light upon the screen, which
is also partially illuminated by gas-light. If
the audience look at it for a few seconds, and
keep their eyes fixed on the screen when I cut
off the electric light, they will see a dark
spot where the bright patch was, and it will
appear to travel about as the eye wanders
over the screen. This shows that the part of
the retina on which the white patch was re-
ceived is fatigued, and is less sensitive to the
feeble gas-light illumination with which the
screen is illuminated.

Some very instructive measures of the sensi-
tiveness of the eye to different shades of light
can be made by a sector arrangement. Black
dots of any size required (in the case in point
they have one-eighth of an inch in diameter)
can be placed on a white disc, as shown.
This disc is cut radially from the centre, and
a black disc is marked out in the figure. The
proportion of black and white can be altered
at pleasure, and a further slight alteration in
the grey produced is made by the dots ; the

Fig. 16.

smallest alteration, of course, being when the
dot subtends the smallest angle. By this plan
the sensitiveness of the eye to any small
change in light can at once be found. The
sector may be varied between all white to
nearly all black. Similarly white dots may be
placed on a black disc, a white disc over-
lapping, and unique measures made. It must
be remembered that in all cases the black
itself reflects a certain amount (in this case
about 4 per cent.) of white light.

Lecture HI.— Delivered April 16, 1894.

The sensitiveness of the eye to changes in
intensity, I have shown you, varies according
to the intensity of light from which the varia-
tion takes place. As my time is short, I must
omit some other theoretical considerations
which it was my intention to show you. I will
first of all commence by showing how it

can be ascertained whether a light is up to the
standard temperature, such standard tempera-
ture being required for visual and photo-
graphic comparisons.

It is well known that by mixing two properly
chosen spectrum colours white light can be
formed, and when I say white light, I mean the

12

colour of the light under trial. Now, for
lecture purposes, it is useless for me to
try and use the light of a candle to form a
spectrum. It would be invisible to you all ;
but I can use the electric light just as well for
the object I have in view, viz., the demon-
stration of the principles involved. Now the
whiter the light, the more blue and violet there
is in its spectrum. There is, therefore, a large
quantity of blue and violet in the electric light.
I will form a spectrum, and place a slit in the
orange and another slit in the blue, so that I
can have a slice of each coloured light. By
means of a lens I can cause these two slices of
spectrum colours toexpandandoverlapand form
an image of the face of one of the prisms used,
and I can then cause a beam of the original
light to illuminate a white surface alongside of
it. The two slits can now be opened till they
form by their mixture exactly the same colour
as the original beam. Let us see if we use
another source of light whether we shall get
exactly the same result, keeping the slits as
they are. I tone down the electric light by a
very pale yellow glass : the light imitates very
closely gas-light. If we place it in front of the
slit of the spectroscope, so that the spectrum
is the spectrum of the yellower light, and the
incident beam is the yellow light, you will see
at once that the mixture of the two colours no
longer gives the same colour as the yellow
light. Making the light the same as the
amylacetate lamp light, you will see again that
the balance is upset, the two patches of light on
the two white surfaces are no longer the same.

Here, then, we have an indication of the
method to pursue in ascertaining if lights are of
the same quality. By having two adjustable slits
in the spectrum, which will with a standard light
exactly match the colour of such a standard,
we can at once see if any other light is of the
same value ; if it is not, the two illuminated
surfaces will be of a different hue. Another
plan is to use proper coloured glasses in front
of a lens, and allow light to pass through
them in such proportions that they cast an
image of a beam of exactly the same colour as
that of the standard light itself. When
another light is used, equality of colour no
longer exists.

There is one method of altering the in-
tensity of a light, if it be a glow-lamp
which may interest some. In the first lec-
ture I showed how the visible rays increase
in intensity in a parabolic curve. This was
further investigated by General Festing and
myself. If each ray goes up parabolically, it is

probable that the sum of them does the same.
In a paper read at the Royal Society on
December 8, 1887, we showed that our surmise
was correct, and that if a constant was de-
ducted from the current multiplied by the volts
the result was the square root of the light
multiplied by a constant — (w - m = n /y~\
which is a parabolic. By altering the resist-
ance in the lamp, and reading an amperemeter,
and a voltmeter, the result is obtained, though it
is sufficient if the amperes alone be read, for
then c2 - s = t /y~ very nearly when
c is current and y the light, s and / being
constants.

As to the use of the sectors, it has been
brought to my notice that Mr. Ferry
has called in question the accuracy of the
sectors when comparing lights of different
colours with one another, such as lime- light
and a glow-lamp. He states that for light of
the same colour, and for monochromatic light;
no error can be found in its use. I may lefer,
however, in opposition to this, to some experi-
ments which were carried out by General
Festing and myself, in which the luminosity of
the spectrum was measured without the inter-
vention of the sectors, comparison having been
made with a glow-lamp. It was found, as
published, that the two methods gave identical
results. There are many other experiments
which show that no error in the results obtained
with the sector have been found by us. That
this is the case, we may take to be the fact by
direct and by indirect measures.

There is, in my opinion, no method so good
in photometry as that of using properly moving
rotating sectors, whose open apertures can be
altered at will. It allows both lights to re-
main stationary, as, also, the screen. This
method of diminishing the intensity of the
light was, I believe, first introduced by Fox
Talbot more than fifty years ago, though he
had not the advantage of using moveable
apertures. This principle of altering the
aperture during rotation I first saw exhibited
by Mr. Kempe, Q.C., at a soiree of the Royal
Society. It was applied to a colour top.
Without entering into the history of the
matter, however, let me show you the ex-
actitude with which such sectors can be
employed.

In doing this, I wish to introduce to your
notice a photometric method which I brought
out, and is, I believe, very fairly successful.
I am not saying it is the very best for
comparing ordinary lights, but it fills a
gap for measuring light transmitted through

>3

bodies, which is very convenient. The principle
of the screen, you will at once see, is different
from almost any other. It consists of a square
aperture cut in a thin disc, and over this is
stretched a white piece of paper of such a
nature that the light from an illuminant is
only scattered, and no direct image can be
seen under any circumstances. On the other
side is cut a mark in black paper or black
retint, which is exactly double the size of the
cut-out square, and this is filled up by the
white paper stretched over the aperture, so

Fig. 17.

Bach View.

that we have a rectangle of paper half of
which is translucent and the other half opaque.
If now we place a light behind the aperture,
the half is illuminated by transmitted light,
and if a light is placed on the other side,
the whole rectangle is illuminated. By placing
a rod in the path of this last beam, we may
cast a shadow which prevents the last illu-
minating the half through which the trans-
mitted beam comes, and then we have half
the paper rectangle illuminated by transmitted
light, and the other half by incident light.
If the paper be of good quality, the light will

Front View.

Fig. 18.

appear of the same colour. By placing the
rotating sectors in the path of the front beam,
and altering the apertures, we may cause the
two to appear of equal brightness.

Now suppose I want to examine the amount
of light transmitted through this piece of
ground glass, I can readily do it. If I place it
near the candle, and use an ordinary Bunsen
or Rumford photometer, I shall find that it
varies according as I place it close to the
source of light, or half way, or close to the
screen. It is quite evident that the closer I

place it to the screen, the truer will be the
measure of the total amount of light trans-
mitted. With this photometer I can get the
ground glass close to the screen, and we then
get a measure of the transmission of light
through it. An objection has been made that
light has been reflected back from the surface
of the white paper to the glass, and back from
that surface again. This may be true to a
very limited extent. If I take a piece of ordi-
nary glass, and hold it close to the lamp, lean
balance the two lights, bringing it closer and
closer, till it in fact almost touches the aper-
ture, you will see that the balance is undis-
turbed. A variety of experiments has shown
that any error caused by this is negligible.
We can take a piece of a photographic nega-
tive, and test it in the same way, and balance
it, and move it at different distances towards
the screen ; we find that if we strike a balance
when it is near the light it becomes apparently
darker as it approaches the light, then gets
lighter and lighter, till it appears lightest of
all as it approaches the screen. Another
point is this, that it need not be used in a
totally dark room, where provision is made
that any light there is must pass through the
body under measurement ; a small amount of
diffused light is of no very great moment any-
where, since it illuminates the front of the
rectangle, and has no effect on the measures
of the light transmitted. We can also use it for
coloured objects, such as coloured glass. For
ordinary purposes it suffices if the glass be
placed against the aperture, or in the path of
the beam somewhere, so long as the aperture
is only illuminated by the light transmitted
through the glass. This makes one half
coloured ; but it is easy to balance the illu-
minations by the oscillations of the sector
[This was experimentally demonstrated.] The
light passed through is then very easily found.
Again I may use coloured paper and do the
same. To myself it is more easy to balance a
coloured light against a white one than a
white one against the white. I need scarcely
say that, first of all, the illuminations of the
white surfaces are balanced, and the sector
opening read before the light coming through
any coloured or other body is measured. If
the white surface require a sector opening of
8o°, and only 40J when a body is against the
aperture in the screen, half the light is
transmitted.

We may often want to know the amount of
light reflected from a body, and the next
photometer I shall show you is used by me

H

for that purpose. It is very similar in principle
to the last. The aperture is cut as before, but
instead of being covered up, it is left open
to allow the coloured object to be placed in it,
alongside a white square. Instead of two
lights one light may be used for this photo-
meter, a reflection being used instead of the
second light. This avoids any alteration in
the relative intensities of the two lights used,
for they both are from the same source. A
rod casts shadows, one on the aperture and
the other on the white square. The aperture
is fitted with (say) a grey square, and the
sectors in the direct beam altered till the
two appear of the same colour ; or I may
introduce a coloured object and repeat the
process. In this case, of course, first of all
the aperture should be fitted with a white
surface and a measure taken, and the aperture
of the two measures of the sectors gives the
relative brightness of the two objects.

There are often cases where we may wish to
measure bodies which only allow but very little

light to pass, though they are transparent.
In such a case we have to use a very powerful
light, and it may be that the body varies
greatly in absorption at different parts. For
this reason I use the electric arc light as the
source, and concentrate it so as to give a
brilliant beam. There are, however, variations
in the electric (arc) light from time to time,
and unless the comparison light, with which
the relative intensities passing through
different parts of the wedge are measured,
varies at exactly the same time and in the
same proportion, the measurements will often
be very much out. If we merely wish to
measure the white light transmitted, the
apparatus to employ is not very extensive,
and Fig. 19 will show what it is. E Lis the
electric light, placed in a lantern or box of
some kind, to prevent the room, which should
be slightly darkened, from being flooded with
light. L, is a condenser which throws an
image of the crater of the positive pole upon
the slit, S, of the collimator, C. The rays issue

M,

/•::-•

Sectors

e i_-=:::::.

L,

R C I— n LMI

--:--- --:--::-p.-:::a

Fig. 19.

W

Screen

parallel, and are caught by a lens Lin, which
forms an image of the slit upon the surface of
the wedge, w, placed in a proper position and
in its mountings. The light, after passing
through the wedge, forms a circle of light on
the screen. It will be noticed that the image
of the slit may be as narrow as one wishes by
opening or closing S, and that we have a line
of light passing through the wedge, such as is
required to effect the graduation. Calculation
will show that, with a fairly narrow slit, the
measured intensity passing through it may
be taken as that passing through the mean
thickness of that part on which the image
falls.

Placed in the path of the beam, and between
the wedge and Lnl, is a plain mirror, M
(for which I often substitute a prism of i£c,
and so obtain a single reflection), which re-
flects the light at right angles, or any con-
venient angle to its path. It is again
reflected from M„ a silver on glass mirror.
An image of the slit is formed in the path,

and a second disc is formed on the screen.
The centre of this disc is made to coincide
with the centre of the disc formed by the
light passing through the wedge. A rod,
R, is placed in the path of the two beams,
which casts two shadows, one illuminated
by one beam and one by the other. The
usual black mask is used on the screen,
to confine the attention to a small part of
the shadows.

It will be seen that, when any variation
takes place in the light, it equally affects both
the illuminated shadows ; hence the measures
may be made without fear of error creeping in.
Sectors with apertures, moving at will whilst
they are rotating, are introduced, as shown in
the figure, and sometimes a second set of
fixed sectors are introduced between M and W
should the light passing through w be too
bright. The screen is placed perpendicular to
the line bisecting the angle made by the two
beams. It should be noted that this plan
almost necessitates movable sectors, but

•s

sectors which are fixed at known apertures
can be used at a pinch, and the balance made
by moving the wedge in its settings.

It should be remarked that though the
wedge may not be pure black the readings
can be very readily made by the method of
oscillating between "too light" and "too
dark" for the shadow whose brightness is
controlled by the sector. In making a valua-
tion of the wedge, the first thing to do is to
compare the lights without the intervention of
the wedge, and then to take readings.

For certain purposes it is necessary to know
how much of each colour of the spectrum is
transmitted through a wedge, and Fig. 20
shows how this is accomplished.

The electric light and the collimator are
placed as before, but the parallel emergent

rays fall upon a pair of prisms, and the
spectrum is brought to a focus by L,, on to
a screen in which there is a slit against which
the wedge in its setting is placed. The slit
can be placed in any spectrum ray, and the
wedge surface is always kept perpendicular to
that ray. A lens, Lln brings the rays to a
focus, so that a monochromatic image of the
surface of the last prism is formed on the
screen. From the surface of the first prism
parallel rays are reflected : these are caught
by a mirror and fall on a pair of precisely
similar prisms, and the remainder of the
apparatus is exactly the same as that de-
scribed above, a second patch of coloured
light being formed over the first patch. The
slit, S, , ,, is so adjusted in the spectrum that
the two patches are of the same colour. The

L,

el-:::x;

S, C c=??

Fig. 20.

sectors are placed as shown in the figure, the
rod, R, forming two shadows, as before. The
method of procedure is to place the slit, S,, in
some colour in the spectrum, and S,,, in the
same. The wedge is then graduated for this
beam throughout its length, another position
is taken up, and the same process gone
through. By this means we get the logarithmic
factor of transparency for each part of the
wedge for the whole of the spectrum colours.

The last point that I shall have to refer to is
an apparent failure of the law of inverse squares
as regards photometry.

I have upon the screen two patches of
spectrum light — a red and a green — of equal
intensity, if anything the red is rather the
brighter. I place the rotating sectors in front
of them and gradually close them. Notice

that the red begins to fade away much more
rapidly than the green. When very nearly
closed the red has disappeared and the green
remains not of its light green colour but as a
green grey.

Let us argue from this what should result.
If when we illuminate a screen with red light
we can remove it to such a distance that the
screen becomes invisible, though if we have
green light, which appeared of equal bright-
ness when close to it, we should be able to
remove it much further before the same screen
became invisible. The point at which the
screen disappeared from view would evidently
be the zero point from which the illumination
wi mid have to be reckoned for the colour which
was used. So with white light, there is a point
at which the screen would become invisible.

It

Evidently then the law of inverse squares for
illumination appears to fail for low intenseness
of light, and this is owing to the insensitive-
ness of the eye. Theoretically, of course, the
screen may be moved to an infinite distance
and still be visible. The law is obeyed prac-
tically of course. It may be thought that this
limit of vision is of no practical account. But
I must say that it is. For instance, in the
photographic room we use red light, and we

find that the corners of a fairly-sized room are
invisible. If we use canary medium the corners
will be well illuminated. This is owing to
what I may call the superior space penetrat-
ing power of illumination of the yellow-greenish
light over the red.

I have not been able to show all I wished
owing to want of time, but I trust that what I
have shown may not be lacking in instructive-
ness.

LONDON :
PRINTED BY WILLIAM TROUNCE, 10, GOUGH SQUARE, FLEET STREET, E.C.

SYLLABUS

LECTURE I.
White light — Sources of light — Standard light — Quality of lights from different sources.

LECTURE II.

Principles of measurement — Different methods of Photometry — Oscillation and scintillation in
light measurement — Colour no bar to measurement.

LECTURE III.

Applications of Photometry to various scientific purposes.

SOCIETY FOR THE ENCOURAGEMENT

ARTS, MANUFACTURES, AND COMMERCE.

C A N TO 11 LECTURES

ON SOME CONSIDERATIONS CONCERNING

COLOUR AND COLOURING.

DELIVERED BEFORE THE SOCIETY OF ART.-!, MARCH, 1S90,

PROF. A. II. CHUliCH, M.A , F.R.S.

Reprinted from the "Journal of the Society of Arls," July A, 11, £ 18, 1890.

LONDON:
PRINTED BY W. TROUNCE, 10, QOUQH SQUARE, FLEET STREET. E.C.

1800.

SOMK CON8IDERTIONS CONCERNING COLOUR

AND COLOURING.

By PROFESSOR A. H. CHURCH, M.A., F.R.S.

LECTURE h— Delivered March 17, 1890.

SYLLABUS. — Definition of terras — Constants of Colour — Primary Colour Sensations — Classification and Nomenclature of
Colours — Luminosity of Pigments — Translation of Colours into White, Greys, and Black — Effects of Varying Illumination —
Dichroism— The Chromatic Circle— Chromatic Equivalents.

It was after considerable hesitation that I
accepted the honour of addressing the Society
of Arts on the subject which is to engage our
attention to-night, and on the two remaining
Monday evenings of the present month. I
was doubtful as to the freshness of my
material and as to my ability in presenting my
opinions to your notice. Then, too, I remem-
bered that Captain Abney had handled some
parts of this subject so ably and so recently,
in the course of lectures which he delivered in
this room in 1888. I felt, however, that if I
could but succeed in developing your interest
in the connection between the science of colour
and the art of colouring, I might perhaps be
justified in acceding to the request of your
Council.

The syllabus of the course indicates the
limited range of these lectures, and shows that
they make no pretension to an exhaustive treat-
ment of the subject of colour. To my " Manual
of Colour," published in 1887 by Messrs.
Cassell and Co., I must refer my auditors for
a systematic exposition ; in this place I pur-
pose developing some only of the topics dis-
cussed in my book, especially in relation to
our everyday surroundings. But I cannot
refrain from expressing the great debt I owe
to the "Modern Chromatics" of Professor
Ogden Rood, published just eleven years ago.
You will notice how often I have to cite his
authority and to quote his results.

Definition of Terms. — The sense given to
certain terms, which will be freely used in these
lectures, ought to be explained at the outset.

The essential difference between tones lies
in their brightness or luminosity, that is the
amount of optical sensation caused by a given
area. Tints contain differing amounts of
white, shades contain differing amounts of
black, broken tints contain differing amounts
of grey— all these are tones ; the particular
descriptions I have given being so framed as

to apply to coloured and uncoloured sub-
stances rather than to lights. When a series
of tones is said to constitute a scale, it is
formed by the addition of equal increments of
the modifying element. Each hue admits of
three scales — the reduced scale, made up of
tints ; the darkened scale, made up of shades ;
the dulled scale, made up of broken tints.
Theoretically, each possible scale starts from
the pure normal colour ; practically, when we
deal with pigments and other coloured bodies
and not with lights, such a perfect starting-
point is unattainable. It should be added
that the tints of black and the shades of white
are the greys, and that passage-tones are
distinguished from the tones which constitute
the above-named scales in one respect only ;
they pass by imperceptible gradations from the
beginning to the end, and not by definite steps
of ascertainable value.

Constants of Colour. — Of these there are
three, namely, hue, purity, brightness. The
hue of a colour appeals at once to the eye ;
we endeavour to identify it, perhaps to name
it, calling it red, orange, yellow, green, &c,
and so on. The purity of a colour refers, in
the accepted language of chromatics, to its
freedom from white ; in this sense no pig-
mentary colour is pure. In the spectrum, under
certain conditions, we have pure colours. The
brightness or luminosity of a colour is measured
by the total amount of light sent to and per-
ceived by the eye, and is independent of hue
and of purity. Colours which do not sensibly
differ from white, are therefore far removed
from purity, are bright because the white
light freely reflected from them produces more
optical sensation than any positive colour. On
the other hand, a pure colour is not necessarily
a bright colour, for not only is the absence of
white light the exclusion of the cause of the
highest luminosity, but a pure colour may be
so low in tone as to be barely recognisable.

When a colour is at once as bright and as
pure as possible, it is called saturated ; such a
colour may be called "full" in speaking of
pigments and other coloured materials. Its
less pure tones are, as I have mentioned before,
called tints, and are more or less "pale." Its
less luminous tones or shades are more or less
" dark." Such of its tones as are at once less
pure than its shades and less luminous than its
tints, are more or less " dull."

Of all the above terms there are two only
which are applied in a manner somewhat
different from that belonging to ordinary
parlance. These two terms are shades of
colour and purity of colour.

Primary Colour- Sensations. — Colour is,
we know, an internal sensation, and has no
external and objective existence. It originates,
in all the cases which we purpose discussing
in the present course of lectures, in the impact
on the optic nerve of that energy or mode of
motion which we call light. Certain waves or
vibrations which affect the nerve-fibrils of the
eye are translated by the brain into colour. I
do not intend to place before you the various
theories which have been propounded as to
the method and mechanism of this transla-
tion ; they involve numerous intricate and
obscure physical, chemical, physiological,
and psychological problems. But I ask
you to accept, as the best of all working
hypotheses, the view of the subject taken by
Dr. Thomas Young in 1802, and since de-
veloped by Helmholtz, Maxwell, and Rood.
The idea that there are three primary colours
must be abandoned, but that of three primary
colour-sensations may be very conveniently
accepted, at least provisionally. What selec-
tion shall be made has been debated on
various grounds. Several triads have been
chosen, but a certain red, a certain green, and
a certain blue have been accepted by some of
the most distinguished investigators of the
subject. I venture to add my own testimony
as to the superiority of the triad of red, green,
blue, over that of red, green, violet. With
lights which provoke these three colour-sensa-
tions we can obtain all the colours of the
spectrum as well as the purples which do not
exist therein. Equivalents— I say, equivalents,
not equal quantities — of the standard or normal
red and green produce the normal yellow ; of
the standard or normal green and blue pro-
duce the normal sea-green ; of the standard or
normal blue and red, the normal purple. Inter-
mediate colours are produced when the quanti-
ties of the two lights, which are together

received by the eye, are not equivalent. Thus
with red and green an excess of red produces
orange ; with green, and blue, an excess of
green produces a greenish sea-green ; with
blue and red, an excess of blue produces a
bluish purple or violet. No tertiary colour is
possible, the colours called tertiary by artists
being broken tints of the three primaries and
of the three normal secondaries. In fact, if
you mix the three normal coloured lights in
their proper equivalents together, you get
nothing but white, or grey if the brightness of
the lights be small ; and if you mix these
three normal lights together in any other than
equivalent proportions, you get tints of the
primary or secondary colours — that is, these
colours mixed with white. By using the initial
letters of the various normal colours to repre-
sent their equivalents, we may place the above
results in the form of equations, which we will
give after the manner adopted by chemists : —

R -\- G = Y, or yellow.
2 R -f- G z= O, or orange.

G -|- B = S, or sea green.
2 G -f- B = greenish sea green.

B + R = P, or purple.
2 B + R = V, or violet.

R + G + B = W, or white.
2R4-G-f-B = W + R, or pale red.
2R + 2G + B = W + Y, or pale yeUow.

R + 2G-J-2B = W-f-S, or pale sea green.
2 R + G + 2 B = W + P, or pale purple.

An essential difference between the colours
produced by mixing lights and those produced
by mixing pigments consists in this, that the
addition or commixture of lights produces in-
crease of brightness, while the commixture of
pigments diminishes brightness. This differ-
ence of result is caused by the fact that the
mixture of two lights possesses the added
brightness of each of its elements, while the
colour resulting from the mixture of two pig-
ments possesses merely the residual bright-
ness left after a twofold or manifold absorption
of light. For when blue and red lights are
mingled on the retina, the eye receives the
combined brightness of both ; but when the
blue and red pigments are mingled, the eye
receives only that portion of the light which
has escaped the absorptive action of both the
blue and the red pigments. In this and
similar cases the dulness arising from absorp-
tion is very marked, and an artist who wishes
to secure the greatest possible brilliancy in
his work avoids, wherever he can, the com-
mixture of pigments, rather placing touches of

colour side by side, so that the colours they
severally reflect shall mingle on the retina ;
the works of Samuel Palmer, William Hunt,
and J. F. Lewis may be named as illustrating
this method.

But there is a second difference between the
colours produced by mingling lights and those
produced by mingling pigments. This differ-
ence is not one of brightness or of purity ; it
is a difference of hue. You know the effect of
mingling yellow and blue pigments together
on the palette is the production of a green,
more or less decided and bright, according to
the chromatic constitution of the two pigments
employed. But the commixture of yellow and
blue lights in equivalent proportions results in
the production of white more or less bright—
that is, either white or a shade of grey —
according to the brightness of its constituents.
If there be an excess of yellow light the colour
produced is pale yellow ; if an excess of blue,
pale blue. You may prove this point by the
use of Maxwell's rotating sectors. Take the
same two pigments, chrome yellow and ultra-
marine blue, which mingled on the palette
produce green— rather a dull green — and
mingle the lights they reflect, by the method
of rotation, on the retina, and you get a white
of low luminosity, that is, a grey. But, as in
this case, there is not the double absorption
produced by the commixture of pigments, the
grey is brighter than the green made by such
commixture, and you will have to add much
black to it (by means of a third sector of black)
to obtain the same low degree of luminosity by
rotation. I have purposely selected for citation
this case of the mingling of blue and yellow,
because it is the crucial instance by means
of which is most strikingly shown the differ-
ence in hue above-named, and it also serves
to demonstrate the fallacy of the still prevalent
notion that blue and yellow lights produce
green. When green is formed by the mingling
of two pigments, one yellow, the other blue,
this result is attained by the suppression of the
yellow and blue and the survival of the green
which was present in both. So when white
light, transmitted through a yellow and a blue
glass, appears green, it is the residual green
which has escaped absorption by both glasses
which colours the emergent light. We know
that the simple yellow of the spectrum, and
the compound yellow produced by the mixture
of red and green lights, affect the red and
green nerve-fibrils of the retina in exactly the
same way, and it is for this reason that we
affirm that yellow is a compound colour-

sensation, and not one of the three primaries
I do not now dwell further on this point, as I
I have dealt with it at length in chapter vii. of
my " Elementary Manual of Colour," to which
I beg to refer you for a fuller treatment of the
differences between mixtures of lights and
mixtures of pigments. I ought, perhaps, to
mention here that the advocates of the red-
green-blue theory of three primary colour-
sensations, while denying the truth of the
explanations afforded by the red-yellow-blue
theory of Brewster, do not contest the occur-
rence of the chromatic phenomena presented
by the admixture of pigments, or by the trans-
mission of light through coloured media, which
Brewster's theory incorrectly interprets.

Classification and Nomenclature of
Colours. — The simplest arrangement of
colours must include white, a balanced or
neutralised compound of two or more hues ;
black, the negative correlative of light and
colour ; and the positive hues, with their tints,
shades, and broken tints. In bare outline it
may be given thus : —

Darkness Black.

I r, , ( White.

Colourless.. . ,

( Greys, or shades of white.

Light., i (Hues.

| Coloured . . ■ Tints, shades, and broken

\ tints of hues.

But directly we endeavour to enter into further
details, to give quantitative values and appro-
priate names to our many materials, we are
confronted by serious difficulties. Whites and
greys must have their brightness determined ;
all other tones require the determination of
the three constants of colour, namely, hue,
purity, and brightness. The fixed lines of the
pure normal solar spectrum enable us, it is
true, to locate all pure and saturated hues
save the purples. Prismatic analysis reveals
the chromatic elements of any colour ; the
methods of shadow photometry enable us to
ascertain relative luminosities, but the in-
telligible expression of our results in reference
to definite standards has not yet been attained.
The problem is, moreover, greatly complicated
by reason of certain modifications of tone or
of hue caused by contrast, dilution with white,
&c, &c, as well as by the immense number
of colours, with their shades and tints, which
the human eye is competent to distinguish.

Of the modes of classifying colours, the cone
of W. von Bczold and the cube of W. Benson
have many merits, but are not free from limita-
tions and even defects. The colour-cube may

be described in a couple of sentences, so far
at least as its principle of construction is con-
cerned. At one solid angle of the cube black,
or the absence of light is placed, at the
opposite solid angle white. At the three solid
angles nearest to black, the full red, green,
and blue are respectively placed, while at the
three corresponding and opposite solid angles
nearest to white the three complementary
secondaries occur, namely, sea-green, purple,
and yellow. The beautiful diagrams in Mr.
\V. Benson's " Principles of the Science of
Colour" will fully explain the position of tints
and shades in the cube.

The chromatic arrangement of Chevreul is
vitiated by the adoption therein of the erroneous
theory of the red-yellow-triad, nor are the hues
represented referred to a definite standard.
Radde's colour-chart, though nominally based
on certain spectrum colours, lacks precision in
plan and execution. The scales of Lovibond's
tintometer are represented by numbered and
named glasses, referred to a certain arbitrarily-
fixed pale tint of which the value of one unit is
given, and then the true chromatic elements
of these glasses have not been determined.

The nomenclature of colours is intimately
connected with their classification. It has
difficulties all its own owing to (i) the shifting
meaning of well-known colour names ; (2) the
difficulty of coining new names. Purple some-
times means dark blue, sometimes a colour
half-way between violet and crimson. The old
German verb bleucii, to strike or beat, has
originated the modern German blauen, and
there is no longer a distinction between the
procedure of the schoolmaster and the laun-
dress, though blauen is given in some diction-
aries for to dye blue, blauen to beat blue, but
bliiue is the colour blue. The system, if such
it can be called, of the mineralogist Werner
(1774) included 92 terms arranged in nine
groups. It lacks accuracy and orderly
sequence, but is of interest in showing the
use of terms for designating colours derived
from certain typical, animal, vegetable and
mineral substances, the hues of which vary but
little. (See "Colour," page 63.)

The long list of colour-names given to silks
by the Lyons manufacturers will furnish some
expressive terms, but many of the names are
quite arbitrary and fanciful. R. Ridgway's
"Nomenclature of Colours for Naturalists"
(Boston, U.S.A., 1887), is useful in giving at
once many colour-names, and actual water-
colour washes of pigments representing them.
Unfortunately, many of these names are very

vague (dahlia, aster, phlox), and they are
placed in an order based on no recognisable
chromatic system. The 51 degrees of the
cyanometcr, for measuring the blue of the
sky, serve to define its tone only, not its hue
nor its purity.

We want an international colour conference,
in which artists, manufacturers, and scientists
shall be represented. We want an agreement
upon the names to be assigned to a number of
different hues. We want representations of
these standard hues reproduced in enamel,
preserved like our standards of weight and
measure, and distributed to every educational
institution in the United Kingdom.

For my own part, I should be content to
employ a combination of the systems of name
and of number in the nomenclature and
classifying of colours. The importance of
having a definite nomenclature, of quite
intelligible character, at our disposal, when
we are talking or writing about the decorative
employment of colour, is so important that I
venture to make a few suggestions which may
tend towards the attainment of this object.

Let, then, equivalents for the full red, the
full green, and the full blue (as already
described), be represented by the symbols R,
G, B, the seagreen, the purple, and the
yellow, which are the several complementaries
of the above hues, being similarly expressed
by the symbols GB, RB, and RG. The
capital letter represents the hue ; where no
number is subscript, 1 equivalent is meant;
for lower or higher luminosities decimals are
used. If the sea-green, to select one in-
stance, inclines towards green, that is, has
more than one equivalent of green in it,
we may easily express this modification
of blue by a fractional addition to the symbol
for green, and a simultaneous fractional sub-
traction from the symbol for blue. Such a
series would result as this : —

GB

GVI B

G,-2 B

Gri'3 B

G,-4 B

Gi-S B

Gr6 B

Gr, B

Gr8 B

Gi-9 B
G

= Seagreen.

5 := Greenish Seagreen.

4
3

= Green,

The corresponding modifications in the five
other principal series of colours would be ex-

pressed in a similar manner, the symbols, ecc,
being used exactly in the same way as in
chemical notation. In order to obtain a scale
in a concrete form, I would recommend the use
of Maxwell's rotation method, by which each
step in the gradation could be matched. I
would use for the series between sea green
and green two discs painted with as near an
approach as possible to the full normal green
and the full normal blue, and the former
should be increased and the latter diminished
(in the form of sectors) in accordance with the
system I have suggested. The rotation colour
produced at each step should be copied, or

rather matched, and would then take its place
in the series with a definite symbol attached
to it. By the introduction of a white sector
similarly treated, the tints of any colour could
be made to match ; with a black sector the
shades, with a black and white sector the
broken tints. The symbol for black may be
conveniently called Z, that for white should be
R G B or \V, as the case may require. But it
is also possible to express the shades of a
colour by reducing its luminosity, and the
broken tints by at once reducing its luminosity
and adding white. Here are the symbolic ex-
pressions or formula; for such changed hues ; —

S = GB = Seagreen.

S-q = Gq B 9 = A shade of, or darkened, Seagreen.

S\V , =1 GBW-i = A tint of, or a pale, Seagreen.

S-9 W-t = G-9 B-9 W-, = A broken tint of, or a dull, Seagreen.

In the case of pigments or other coloured
materials the proposed symbols, though quite
exact enough for any purpose for which they
are employed, will not admit of accurate
analysis. This is due to the fact that no pig-
ment reflects a pure coloured light, there is
always some white mingled with it, rarely less
than 20 per cent, of the whole. And even with
coloured lights there are also variable sources
of error, for the addition of white light to
coloured light, and also the reduction of its
luminosity, alters the original hue. To this
change, which occurs also with pigments, and
often in a very conspicuous manner, I shall
invite your attention shortly.

Luminosity of Pigments. — One often hears
the remark with reference to a work in black
and white, an etching or line-engraving, for
example, representing a coloured picture, that
it has got good colour in it. Such an expres-
sion may be interpreted to mean that the
relative tones of the original are preserved in
the copy, and suggest the force of the original
colours. Now it requires much experience
and singular skill thus to value and interpret
the relative luminosity of differently coloured
portions of a painting. An aid in the work of
translation is afforded by making determina-
tions of luminosity or brightness by means of
shadow-photometry, or by comparing the
b'ightness of various white and coloured
materials with the brightness of greys of
known composition. The latter method,
though not very easy or very exact, involving
the use of black and white sectors in the
rotation apparatus, gives results which are, at
all events, approximately true. These results

represent the comparative luminosities of the
particular specimens of pigments, &c, em-
ployed, and vary somewhat with the method
of applying the pigment, its thickness, and
the medium (if any) used. The following
figures were obtained in this way, some by
Professor Rood and others by myself: —

Chinese white, dry, in thick layer

on pastel paper - 100

Plain white paper 90-9

Whatman's paper (not hot-pressed) 88-2
Chrome yellow, pale water-colour

wash 73'°

Emerald green, pale, in thick paste 44-2

Cobalt blue, water-colour wash 32-2

Vermilion, in thick paste 23-4

Natural Ultramarine 8-3

Artificial Ultramarine 6-9

Black paper \",

Lamp black, dry, in thick layer on

pastel paper "8

I cannot help thinking that a more extended
series of comparative luminosities would prove
useful to the translators into black and white
of works executed in colour.

Effects of Varying Illumination. — The
very considerable alteration of appearance ex-
perienced by coloured materials when the hue
of the light by which they are rendered visible
is changed, is rendered familiar to us by the
case of pictures and dresses, seen first by
candlelight and then by ordinary daylight.
We discover that what we thought was purple
is really violet, crimson is purple, green is
blue, and so forth. The deficiency of the
more refrangible rays of the blue end of the

spectrum, and the superabundance of the less
refrangible rays of the red region in the light
of burning oil and gas, and even in that
emitted by incandescence electric lamps,
furnish the explanation of these changes. It
is of course more conspicuously evident when
we view richly and variously coloured objects
in a light which is virtually monochromatic.
The pure orange yellow of sodium cannot
furnish the vibrations to which the majority of
hues are alone capable of responding ; only
what is yellow can be seen, and even in this
case not the yellow compounded of red and
green vibrations. But I must not enlarge upon
this point, as it has been often treated, and at
length, elsewhere ; and I want to direct your
attention to those changes of hue in coloured
objects which are caused by alterations, not
in the hue of the incident light, but in its
brightness. The observant student of Nature
will have learnt that the hues of many-
coloured objects are most characteristically
brought out by a comparatively moderate
illumination. The chief cause of this pheno-
menon is often traceable to the large excess of
unaltered white light, which such objects
reflect or scatter when the illumination is
intense, while when the illumination is mode-
rate it just suffices to develop properly the
particular hue of the material. I recollect a
case in point. Wishing one day to show to a
connoisseur in precious stones a fine aurora-
coloured zircon, which had lately been cut for
me, I took the specimen into the sunshine.
The colour was so altered and so impoverished
that my pride in the recent acquisition was
humbled ; the stone was distinctly inferior to
a similar specimen in the British Museum.

This pallor was not the result of the glittering
reflections from the polished facets of the
gem, but the coloured light reflected to the
eye from the lower facets was greatly reduced
in purity by much admixture of white light.
Out of the strong light the specimen resumed
its rich and beautiful appearance, and showed,
moreover, not only a much deeper but a much
redder and less yellow hue than when sunlight
fell upon it. Painters have long been familiar
with such changes of hue. If you examine
the works of the great colourists of the Italian
schools you will see many striking examples
of their appreciation of this natural phe-
nomenon ; indeed, I may say that these
changes of hue with changes in the brightness
of the illumination have been not infrequently
much accentuated, even greatly exaggerated.
At all events, artists have long known that the
high lights of a self-coloured drapery cannot
be properly represented by merely adding
white— that is, by forming tints of the same
hne— although it does indeed sometimes
happen that the addition of white paint to
one having positive and strong colour brings
about changes of hue analogous to those we
have been considering. Be it remembered
that I am not now speaking of those purely
accidental modifications of hue which are
caused by reflections from coloured objects,
and which often cause a white drapery to
assume the beautiful orient tints of the pearl.

If increase of light above what we may call
a normal standard (one capable of showing in
perfection a colour) effects a change in hue, so
also does decrease of light, although in an
opposite direction. These changes of hue are
well brought out in the following Table : —

Decreased Light. Standard Light. Increased Light.

Purplish Red Scarlet.

Red Scarlet Reddish Orange.

Brown Orange Yellow.

Olive Green Yellow Paler Yellow.

Greener Yellow-Green Yellower.

Greener Blue-Green Bluer.

Violet-Blue Blue Turquoise.

Violet Violet-Blue Blue.

Violet-Purple Violet Bluer.

Purplish- Violet Purple Red- Purple.

Similar changes of hue occur when coloured
lights instead of coloured objects are reduced
or increased in brightness. Thus by lowering
the luminosity of a pure solar spectrum, there
will not only be a selective reduction or even
extinction of some of the constituent hues, but
also a shifting of the position of the hues. The

red will invade the orange region, so that the
line I), for example, will be bordered by a kind
of red-lead colour ; the green will extend to-
wards the sea-green, and even include the line
F, and the pure blue will contract. We should
then describe the darkened spectrum as prac-
tically consisting of red, green, and violet. By

a further reduction of brightness the red will
become brownish, and the green and violet
dull. In the next stage the violet disappears,
the green is more dim, while the brownish-red
acquires a sort of chocolate hue. The last
colour to be recognised is a very faint green
hue. I ought to mention in this place that
when white light is added to coloured light
the changes of hue which occur are generally
to be explained by the statement that they are
such as would be produced by a slight ad-
mixture of blue light with the white light.

The bearing of these observations upon
pictorial and decorative work in colour are
obviously of considerable moment. If we want
to be sure of the chromatic effect of a decora-
tive scheme we must arrange and modify its
elements in such a way as to adapt it to the
usual illumination of the apartments in which
it is to be carried out ; and here one curious
result may be noted. In a very dimly lighted
room, saturated and comparatively pure
colours may yield delicate and refined har-
monies, while the same colours in a blaze of
strong daylight, or direct sunshine, may also
lose their garishness because of the large
amount of white light which then becomes
mingled with the proper hues they reflect.
But with a medium illumination these satu-
rated colours, if freely used, become intoler-
able, just because they are perceived in their
original strength. We thus learn how it is
that a scheme of colouring, which seems
exquisitely choice when occurring in the dimly
lighted apartments of a Cairo or Damascus
house, may become to the last degree crude
when transferred to the full illumination of an
English home, where the walls are pierced
with large windows of plate-glass admitting
floods of light. The employment, under such
altered conditions, of colours saddened with
black and dulled with grey, has been long
recognised and practised as one of the easiest
ways of modifying the crude effect of such
chromatic arrangements. If saturated colours
are still used we reduce their area, or employ
them in such small and involved touches that
the mingling of their hues produces what Owen
Jones called a neutralised bloom.

Before I leave this important subject (of the
effects of varying degrees of illumination upon
the hues of objects), I wish to direct your
attention to a very simple and decisive class
of experiments, by means of which some
of these changes of hue of which I have
spoken may be recognised. You need
only take some pieces of richly coloured

material, paper, cloth, or silk, of uniform hue,
and partially crush and fold them, so as to
get various portions illuminated in different
degrees. An orange coloured silk will look
yellow in its high lights, and brown in its
hollows. A blue paper will appear violet in
its shaded parts, and of a purer blue, or
perhaps of a somewhat greenish blue, in its
projecting portions. A crimson cloth will
seem scarlet where it catches most light, and
crimson where it reflects the least. These
changes of hue are in part due to errors of
judgment, but they are none the less real.

Our next subject for study is intimately
allied with that we have just been considering.
I regret that the narrow limits of my available
time will permit me to give little more than the
barest reference to it. I refer to the alteration
of hue caused by the addition of white, of
grey, and of black to a pigment. White added
to a red pigment does not usually produce just
a paler tint of red, but the new tint varies in
hue, tending in some cases towards crimson,
in others towards scarlet. The dilution of a
transparent red pigment with a colourless
medium— that is, attenuation of its thickness —
produces analogous alterations of hue. Similar
changes are caused in pigments ot other
colours by dilution and by commixture with
white substances. Some of the changes pro-
duced by adding black are perhaps more
striking. Rood found that, on mixing carmine
with lamp-black, the mixture on the palette
was more purplish in hue than the colours
obtained by mingling the pigments optically
by the method of rotation. This approach
to purple was of course due to addition, with
the black, of more blue to the carmine. An
analogous observation, also due to Rood,
relates to the admixture of black with white.
Black pigments as free as possible from
any tinge of positive colour are generally ob-
served to yield a bluish-grey when mixed with
white pigments. This result has usually been
attributed to the fineness of the particles pro-
ducing a blue by the same action on the light
as an opalescent medium exerts. But Rood
found that when white and black are mingled
optically on a rotating disc, the grey they yield
is matched in hue (not in brightness) by a
white disc into which much blue (in one case
17 per cent.) has been introduced by means of
a sector covered with a strong wash of indigo.
Dickroism is the next topic for our study.
This term has been applied to similar phe-
nonema produced in similar ways. A liquid or
solid is said to be dichroic when the light

8

which is transmitted through it differs in hue
as well as in tone when the thickness traversed
differs. Many liquids, such as solutions of
double potassium and chromium oxalate, and
potassium permanganate, well illustrate the
phenomenon, so do many solids, such as copper
sulphate, and several kinds of coloured glass
and gelatine films. But there is another kind
of dichroism occurring in solids which are not
optically homogeneous. Many crystals are
dichroic, or even pleiochroic, for when white
light is sent through them in different direc-
tions it suffers thereby different selective
absorptions, and so the emergent beams are
differently coloured. The ruby, the emerald,
and the tourmaline amongst precious stones
are characteristic examples of this group of
dichroic phenomena. Even a third class or
group may be named, where the light trans-
mitted differs in colour from that reflected and
scattered, solutions of chlorophyll and of
uranin, and the solids uranium glass and fluor
spar belong here.

The first class of dichroic phenomena
(depending upon the thickness of the coloured
medium) is largely concerned with the pro-
duction of beautiful quality in colour, as we
shall see in the next lecture. It depends upon
the increasing absorption of certain of the con-
stituents of white light with increasing thick-
ness of the medium. Thus, a solution of blue

vitriol or, better still, a crystal, allows blue,
green, and yellow light to pass through it if it
be thin, but when thick only blue. Analogous
phenomena are shown by purple films of
gelatine, and by a solution of chromium
potassium oxalate. These are cases of in-
creasing selective absorption where waves of
certain coloured lights can pass in some
measure through weak solutions or thin layers,
but are at last completely stopped or absorbed
by layers of increased thickness.

The Chromatic Circle.— \ can dwell no
longer on this very convenient mode of arrang-
ing an orderly sequence of colours than will
just suffice to explain the uses which I shall
have to make of it in subsequent lectures. The
disc of which our chromatic circle is the
boundary may be regarded as the base of a
colour cone, differing from that of von Bezold
in the approximately correct angular positions
accorded to the several colours which find
place in the circle.

I have followed in the main the construction
given by Rood to the chromatic circle, but I
have not adopted his nomenclature, except
in a few cases, while I have used capital
letters to indicate the hues corresponding to the
three primary colour-sensations, and have em-
phasized these and their complementary secon-
daries by joining them by diameter lines of
extra thickness. (See diagram.)*

Crimson
Heel-Purple,
Put-pie

RED

Purple-Violet

Violet,
Violet -blue.

BLUE

Jurqut.

Orange -Yellow

GreenishYellow
'Green-Yellow

'Yellowish Green

Greenish-blue

'EEN
Emerald Green
Bluish Green

Green-Blue

Professor Rood determined the position of
his normal red ("spectral red"), in a pure
solar normal spectrum, and found it to be at
285, when the space between the fixed lines A
and H was divided into 1,000 equal parts.
Similarly his normal green is situated at 600,
and his natural ultramarine at 785. His normal
blue was taken at a point nearer the blue-green

side of lapis-lazuli than that which I am able
to accept. The three primaries I adopt are
approached very closely in hue by the follow-
ing pigments : —

Crimson Vermilion, for the red.

* This diagram is reproduced from "Colour," p. 91
(Cassells,)

Emerald Green with a trace of chrome, for
the green.

Pure Natural Ultramarine, for the blue.

Putting aside the questions of purity and of
brightness, each one of these three primaries
is neutralised, so far as hue is concerned, by
one of the three secondaries, thus — Red, by
green -f- b'ue or seagreen ; green, by blue
-f- red or purple ; blue, by red + green or
yellow. Thus the pairs, red and seagreen,
green and purple, and blue and yellow, are
what we call complementary colours, and
united produce in each instance white. But
of course there are an infinite number of com-
plementaries, the pair of colours at each
extremity of every diameter of the circle
having the right to be so designated. This,
however, is a subject to which we must return
in the next lecture, and so I wish now to con-
fine our attention to two other matters — the
angular intervals which separate the selected
colours to which I have assigned names, and
the doctrine of chromatic equivalents.

In the earlier attempts at arranging colours
in a circle, an inaccurate and quite arbitrary
plan was adopted of placing the three primaries
at three points equidistant from one another.
Supposing, however, that we obtain a normal
diffraction spectrum, and merely glance at the
position which our three selected fundamental
hues occupy, it will be immediately obvious
that the green is separated from the blue
by a smaller interval than that which divides
it from the red. And if we take the nume-
rical differences obtained by comparing the
wave-lengths of the three fundamental hues,
we shall arrive at the same conclusion.
It is by fixing the spectral position of
the pigments which most nearly approach
these fundamental hues, and consequently
learning their corresponding wave-lengths,
that we are enabled to assign approximately
correct angular positions on our chromatic
circle to the pigments in question. For the
practical purposes we now have in view, the
allocation of proper positions to our most
characteristic pigments is the problem which
our chromatic circle is intended to aid in
solving.

Chromatic Equivalents. — Still dealing with
pigments, we may ask— What areas of com-
plementary pigments neutralise one another's
hues ? and — What are the equivalents of the
three fundamentals as represented by the
corresponding pigments ? We can answer
these questions with a sufficient degree of
accuracy by means of rotation-experiments

with sectors suitably painted, using such areas
of the different pigments as shall produce a
neutral grey. Here is one set of results thus
obtained with a circle divided into ioo equal
degrees : —

Red (crimson vermilion) 3^"i

Green (emerald green and chrome) . . .^o

Blue (natural ultramarine) 26°7

On rotation these sectors produced a grey
equal to that derived from 25 ° white (paper),
with 750 black (paper). This large quantity
of black shows how great a stride towards
darkness is made when pigments (owing their
colours to selective absorption) are used.
From the above chromatic equivalents, re-
duced to their simplest expressions, we get
the following equations : —

Red = 12. Green = n. Blue = 9
Red -J- Green = Yellow = 23
Green -(- Blue = Seagreen = 20
Blue + Red = Purple = 21

It will be remembered that Field, by the
erroneous method of combined absorption
through coloured glasses, found the following

equivalents : —

Yellow = 3

Red = 5

Blue = 8

And on these equivalents, widely as they
differ from the truth, have been based
those calculations generally current amongst
colourists as to the proportionate areas of
coloured surfaces, by which a chromatic
balance and the so-called "neutralised
bloom " are to be secured.

I may add here that Rood, employing a
red paper of moderate brightness along with
emerald green and artificial ultramarine, ob-
tained for these pigments the following values
in degrees : —

Red = 50. Green = 31. Blue = 19.
The neutral grey these sectors produced on
rotation was matched by the rotation mixture
of 13 of white (paper) and 87 of black (paper).
These values differ, as one expected they
would, from those I previously gave, but they
lend no support to the figures of Field. On
the contrary, they correspond to the ratio —

Yellow, 81 : Red, 50 : Blue, 19;
while Field's figures, multiplied by 10 for the
sake of easier comparison, arc —

Yellow, 30 : Red, 50 : Blue, 80.

Next Monday 1 purpose discussing further
the subject of complementaries and of con-
trasts, inviting your attention also to various
arrangements of colours in association.

LECTURE II.— Delivered March 24, 1890

Svj.LAUUS.

■ Complementary Colours— Contrast of Tone and Colour — Harmonies of Colour — Combinations of Colour,
Dyads, Triads, Tetrads, &c. — Separation of Related Hues — Throbbing or Pulsing Colour.

Complementary Colours. — The chromatic
circle, to which I drew your attention in last
Monday's lecture, is also, you will have seen,
a contrast circle. On its circumference may
be arranged a host of colours, limited in number
only by the discriminative power of the human
eye ; these colours will be united in pairs by
the diameters connecting th^m. Each colour
has its complementary in that colour which is
1800 off, that is, most remote in position from
its own place on the circle ; the pair united in
equivalents from white ; in other than equiva-
lents one hue preponderates, but a tint is thus
formed, for it is merely the uncombined excess
which is perceived mingled with the white
produced by the equivalents. Conventionally,
we speak of white as present in the centre of
the circle ; in truth, it is present at different
points on each of all possible diameters, in
accordance with the value of each chromatic
equivalent. And when we concern ourselves
with pigments, we shall soon discover that
their positions are variously disturbed by their
divergence from normal standards of hue,
purity, and brightness.

Besides the numerous complementary pairs
which we can place and recognise on the
circle, there are many corresponding pairs
■within its area situated on its diameters ;
these are complementary tints. And outside
the circle, on the prolonged diameters, will be
found complementary shades — that is, colours
of reduced brightness.

There are several ways of learning the com-
plementary of any hue. In practice we gene-
rally desire to ascertain the complementary
hues of coloured materials ; to this question
we confine ourselves now. As such materials
produce their hue by selective absorption, we
cannot expect to obtain white from the union
of the rays they send to the eye, but must be

content with a neutral grey, often with a very
dull or dark grey. One method of ascertain-
ing the complementary of, say, a pigment,
consists in erecting a sheet of glass in such a
way that we can see directly through it the
pigment of which we want to ascertain the
complementary hue. We then cause an image
of each of a series of trial complementaries to
be seen by reflection on the glass, so that the
direct and the reflected images coincide.
When the patch of light thus produced has no
positive hue in it, we know that our trial com-
plementary colour is the true one. But this
method is troublesome, and so we generally
use another, mixing the lights reflected from
the pair of colours by rotating coloured sectors.
The case of brown paper is generally taken in
order to illustrate this method. We prepare a
disc of such paper, and associate it with such
angular amounts of other discs as shall pro-
duce on rotation a neutral grey. In this parti-
cular instance Rood found that it required 45
parts of artificial ultramarine and 14 parts of
emerald green to produce a greenish blue
which should neutralise 41 parts of brown
paste-board colour. The grey these comple-
mentaries yielded was matched by rotating 24
parts of white and 76 of black paper. Just
look a little more closely into the chromatic
constituents of this pair of complementaries.
The three pigments — vermilion, emerald green,
and artificial ultramarine — required in order
to yield a grey need to be rotated in the pro-
portions 36 : 34 : 30. But the complementary
of our paste-board contained the emerald
green and ultramarine in the ratio 14 : 45 or
9'3 : 30 ; consequently the paste-board must
have furnished the equivalent of 247 parts of
emerald green as well as that of 36 parts of
vermilion. But these proportions of emerald
green and vermilion, when mixed by rotation,

II

produce an orange - yellow of rather low
luminosity— a broken tint of orange-yellow—
and it is by this term that we should define the
colour of the particular brown paste-board.
[It should be noted here that by the successive
impressions of colours on the eye, produced in
the rotation experiments, we do not get the
added luminosities of the pigments, but only
their mean luminosity ; and that our bright
yellow and orange pigments possess a degree
of brightness quite out of proportion to that
of most of our pigments of other hues.]

If we put aside the questions of brightness
and purity, taking into consideration hue only,
we may describe all complementary pairs as
made up of—

[a.) One Primary and one Secondary , or of

{6.) Two Secondaries.
Examples of a are furnished by the pairs —

Red : Seagreen.
Green : Purple.
Blue : Yellow.

Among the countless instances of b, I name —

Orange : Greenish-blue.

Orange-yellow : Turquoise.

Greenish-yellow : Violet-blue.

Yellowish- green : Purple-violet.

Bluish-green : Crimson.

There is one characteristic of all true com-
plementary pairs of colours— they do not affect
each other's hues. They are already as unlike
in hue as possible, and cannot therefore
mutually modify this colour-constant. They
may, and often do, modify their apparent
luminosity and purity. So in the pair " blue :
yellow," which affords, next to "black:
white," the strongest possible difference of
tone, the blue makes the yellow at once
brighter and paler, while the yellow makes
the blue at once purer and less luminous. The
characteristic of the non-alteration of hue
when the complementaries are placed in con-
tact affords a criterion of the truly comple-
mentary nature of any pair of colours. For
instance, in this diagram of true and false
complementaries, in the pair " red : green,"
the red makes the green look bluer than it is,
while the green makes the red appear to
incline towards purple. So in the pair
"yellow : purple," the yellow looks some-
what greenish, the purple somewhat violet ; and
in the pair " blue : orange," the blue inclines
towards green, and the orange towards yellow.
These are examples of the peculiar optical
effect known as simultaneous contrast, in
which two contiguous surfaces, which differ in

one or more of the constants of colour, have
such differences mutually enhanced or modi-
fied. Of such contrasts of hue, brightness and
purity, I shall have something further to say
directly, but I want to ask you first which of
my columns of supposed complementary pairs
commends itself to your judgment and taste as
preferable ? I do not say that the strongest
possible contrasts of hue, as afforded by our
approximations to the complementaries, are
necessarily agreeable, but I do feel that the
similar series constructed on the false basis of
the red-yellow-blue theory is decidedly un-
pleasant : the angular intervals between the
pairs should have been greater or less. Red with
seagreen seems to be more agreeable to the eye
than red with green ; green with purple looks
better than with yellow; blue with yellow
better than blue with orange. That these
effects are not wholly due to differences in
brightness may be argued from this case of
blue with orange, where I have tried to lower
the luminosity of the orange so as to make it no
brighter than the blue with which it is asso-
ciated. But the inherent falsity of the red-
yellow-blue theory precludes one from giving
to these pairs of complementaries their proper
share of brightness — assuming for the moment
that we could really accomplish such a result
with pigments.

Contrasts of Tone and Hue. — Contrasts
are of three kinds, namely, contrasts of hue,
contrasts of brightness, contrasts of purity. I
have already directed your attention to con-
trasts of hue ; the other two kinds of contrast
now demand notice. The best illustration I
can give you of contrast of tone will be one in
which we shall produce simultaneously a series
of greys. By means of this black figure
mounted on a white disc we shall produce a
series of optical mixtures of black and white,
regularly increasing the white by five steps.
The outermost ring of grey will be the palest
tint of black, the innermost the deepest shade
of white. But you will observe another and
unexpected appearance produced in this ex-
periment. Each ring seems to be itself tinted
and shaded, although it must really be of one
uniform grey. Simultaneous tone-contrast is
produced, not by optical fatigue but by an
error of judgment. The effect is enhanced by
the difference of tone on each side of these
contiguous rings, for the middle ring is
bounded by a darker shade of grey on one-
side, and is there made lighter by contrast,
while on the other side it is in contact with a
paler tint uf grey, and is there made darker by

contrast. The immense importance of this
phenomenon in pictorial and decorative art is
well known ; it, of course, takes place with
positive colours as well as with negative. If
in an engraving or Indian ink drawing a
certain dark space seems monotonous and
heavy a single touch of a darker shade will at
once lighten and vary it, while just in the same
way a lighter stroke on a tint which is too pale
will enrich and deepen it. In both cases the
result of simultaneous contrast will relieve the
flatness of the work.

When positive colours come into the
arrangement of differing tones the results
may be indicated by means of a few diagram-
matic figures. Suppose we begin by intro-
ducing a kind of brown in the form of a disc.
This colour of medium depth is deepened by a
smaller half-disc of white near its centre, and
lightened by the contiguity of a corresponding
half-disc of black. Again, suppose we place
a small disc of full blue upon a larger disc
half white and half black, the black will make
the blue it touches less pure, that is, will
lighten it, while the white will affect the blue
in an exactly opposite way.

We now proceed a step further, and take
two colours instead of one. Using the same
reddish-brown colour as before, we place it on
a disc of deep blue, and find that it becomes
lighter in tone and less pure, while at the
same time it acquires a distinctly yellow cast,
having thus received a portion of the comple-
ment of blue. Similarly on an orange-red
(red lead colour) disc, lighter in tone than its
own, it appears deeper in tone, and assumes a
somewhat greenish-blue cast, though this
effect is but slight, owing to the addition
of this complementary mainly serving to
neutralise some of the red in the brown disc.

A third case may now be considered. We
select for apposition two colours which are
separated from one another by a compara-
tively small angular interval in the chromatic
circle, and which are both somewhat bright.
We find, as we might expect, that their
mutual action as to alteration of tone is but
slight, while their change of hue is more
decided, each tending to recede farther from
its companion by acquiring more of the com-
plementary hue. Here visual fatigue and the
warping of the judgment both conduce to the
same result.

Harmonies of Colour. — I must now ask
you to consider what are called chromatic
harmonies. Chevreul's classification of har-
monies under two headings — harmonies of

analogy and harmonies of contrast — is some-
what arbitrary. All harmonies are in varying
degrees harmonies of contrast. In every
chromatic harmony there is contrast, contrast
of hue, or of brightness, or of purity. Contrast
generally, therefore, does not afford a criterion
of classification. No fundamental difference
really exists between those harmonies where
change of tone or of hue is gradual and those
where it is abrupt ; it is a matter of degree,
and is quantitative rather than qualitative. I
do not see my way to improving the arrange-
ment which I proposed in my " Manual of
Colour" (pp. 134 to 139). I suggested arrang-
ing the various possible harmonies upon the
arc of a circle, placing the harmonies of close
analogy at one end, the harmonies of strong
contrast at the other: —

1. The passage, by insensible differences,
of the tints, shades, or broken tints of one
hue, from light to dark.

2. The gradation by small but regular,
definite, and perceptible steps, of the tints,
shades, or broken tints of a single hue, from
light to dark.

3. As in 2, but each step separated by a
neutral element, as white, grey, or black.

4. The passage, by insensible differences,
of one hue, or of its tones, into another
related hue or its tones.

5. The gradation by definite steps of one
hue, or of its tones, into another related hue
or its tones.

6. As in 5, but each step separated by a
neutral element.

7. The passage, by insensible differences,
of one hue, or its tones, into another chromatic-
ally remote hue.

8 The gradation, by definite steps, of one
hue into another chromatically remote hue.

9. As in 8, but each step separated by a
neutral element.

10. The collocation of different tones.

11. The collocation of chromatically distant
hues.

12. The collocation of chromatically distant
hues separated by a neutral element.

The idea of seriation or gradation becomes
more and more involved with that of decided
change as we follow the sequence of these
numbers ; analogy gives place to contrast.

This classification may be illustrated by a
few examples, but I have no time to develop
the subject completely. And it must be noted
that such definitions as I have offered are of
use in the way of suggestion, but must not be
allowed to fetter the imagination of the artist ;

'3

they cannot take the place of experiment,
observation, knowledge, and sensitive per-
ception of chromatic beauty.

I take three examples only ; they will serve
to illustrate Nos. 5, 6, and 7.

The diagram shows the gradation in three
steps only, of yellow to orange (Xo. 5 above ;
and also the same gradation when the neutral
elements, white, grey, or black, are inserted
(No. 6). These cases furnish examples of the
use of what has been variously termed " the
dominant hue" and "the small interval."
We may illustrate it by the budding foliage of
spring, the changing hues of autumn, and the
association of nearly-related hues in many
flowers, as in several kinds of daffodil.

My next diagram shows the passage of red
to blue 1X0. 7 above). Really it is only a
more extreme case of our No. 4, but there is
not here a chromatic element common to the
whole series of passage hues, for the red at
one end contains no blue, and the blue at the
other end no red.

Associated Colours. — Of pairs, or dyads, I
must say a few words. These are of three
orders, namely, complementary pairs, large-
interval pairs, small-interval pairs. Com-
plementary pairs, when their tones differ
considerably, as in the case of deep red with
bright sea-green ; or when they contain much
white, or are pale ; or when they contain
much black, or are of low luminosity ; or when
they are at once of low luminosity and contain
white (that is, are broken with grey), are all,
or may be, so adjusted as to be agreeable.
Large-interval pairs are usually more difficult
of management, and when both possess a full
degree of saturation, are often unpleasant.
Cases are afforded by red with yellow, orange-
red with blue-green, orange with purple,

orange-yellow with green, orange with green,
and greenish-yellow with turquoise. The dilu-
tion of one of the pairs with a neutral element
generally improves these combinations, but
some of them, though so strong as to require
moderation in use, are naturally excellent.
Amongst these may be named red with blue,
orange-red with blue, orange-yellow witli
violet, and blue-green with violet. In these
cases, however, there is normally a consider-
able difference in the luminosity of the two
colours associated in pairs. By beauty of
material and of pattern some of the simplest
dyads may yield exquisite effects, as in Italian
velvets and brocades of the 1 6th century. Such
dyads areyellow-greenwith medium violet, pale
olive green with deep indigo, leaf-green with
deep blue, and pale leaf-green with deep amber.
The simplest triads contain two colours and
a neutral. Suppose we desire to associate
violet and blue we may interpose with agree-
able effect between them white or a pale grey,
but not black. The blacks available have a
low degree of luminosity and so do not differ
enough from the full blue and full violet we are
using; moreover they become tinctured with
an unpleasant rusty hue, the complementary of
the blue or violet. Generally the poor and bad
dyads may be improved by the introduction of
a third element, while good dyads are often
spoilt thereby. Good triads may frequently be
secured by taking three hues rather widely
separated from each other on the chromatic
circle, and, as a general rule, two of these
hues should be pale or bright and one deep or
dark, or vice versa. Except in special cir-
cumstances it will be easier to manage these
triads when two of the members belong to the
"warm" side of the circle and one to the
" cold." I give a few good triads : —

r Amber,

< Cream,

\ Blue (medium).
Amber,
Blue (pale),
Crimson.

S Flesh-red,
Blue (normal),
Olive-green.
Brick-red,
Indigo,
Sea-green.

Leaf, green,
Puce (deep),
Rose-grey.
Terra-cotta,

■ Leaf- green,
'iolet,

■' Maroon,

Sage-green.
I Maroon,

Bronze-yellow,
' Olive-green (dark).
, Bluish-green,

Violet,

Red-ochre.

/ Lea
Vio

\ Salmon.

Ypricot,

Crimson,
Gold-brown.

(

/ Red (normal),

Gold,
\ Blue (normal).
/ Lavender,
i Turquoise,
I Blue (deep).

Other examples of triads, and of more com-
plex colour-combinations, taken chiefly from

historic and national chromatic arrangements,
are —

•4

i. Blue (deep), or Lavender (pale)

2. Gold-amber

3. Vermilion-red in fine lines on 2

1. Lavender (pale)

2. White

3. Black

4. Red-brown (medium) )

1. Chamois-leather

2. Raw Sienna J on ground of 4, black

3. Venetian Red

on ground of

5, orange yellow

6, ochre-yellow

(Ancient Egyptian)

(Assyrian)

1. Blue (medium) "j

2. White

3. Red

4. Greenish grey )

with outlines of 5, black

1. Silver

2. Gold

with outlines of 3, puce, all on ground of 4, pale emerald green

1. Silver

2. Lavender

3. Emerald Green (medium)

with 4, white in bands
and 5, black in lines

Silver

Gold

Red edgings

Black ground

(area =
(area =
(area =
(area =

A)

T'o)
T'o)

Lavender
Lemon-yellow
Crimson (medium)
White (in strap work)
Green (medium)

Bluish-green (full)
Yellowish olive (medium)
Orange (pale) 6.

Crimson (medium) 4.

Salmon 5.

Yellow (pale) 6

r- with as lines and ground, 6, black

4. Blue (full)

5. Purple (medium)

6. Crimson

Yellow green (medium)
Yellowish olive (medium)
Maroon

(Greek)

(Moresque)

(Indian)

(Irish MS.)

(Irish MS.)

(Irish MS.)

And now I desire to draw your particular
attention to one of those characters of good
colour which is of supreme importance. You
will understand the nature of this " quality "
in colour when I have described and analysed
illustrative examples. I begin by choosing
a few cases in which it is either absent or
irreco.gfnisable. A surface of matt paint of one
uniform tone and hue, and illuminated by a
direct and colourless light, is confessedly pro-
saic and uninteresting ; so also is a piece of
self-coloured "surface" paper or a plain
printed cotton cloth. All the materials are
flat, and they exhibit no variations of tone or
hue, or at least no perceptible variations. Of
all the hues that can be chosen as examples of
this point, perhaps the full red, the full green,
and the full blue are the most telling ; but the

full orange and the full violet come very near
to them. If the generally accepted theory of
the three primary colour-sensations being red,
green, and blue be accepted, we shall perhaps
be able to suggest a partial explanation of the
unpleasant appearance of a considerable sur-
face of any one of the three above-named hues
when alone in the visual fatigue of one set of
optical nerve-fibrils which they severally cause.
Orange and violet are, however, assumed to
affect the red and green fibrils and the red and
blue fibrils respectively. Yet in the case of
these two colours (orange and violet), the
stimulation of the two sets of fibrils which
they each cause is not equal, for the orange
has an excess of red in it, and the violet has
an excess of blue. If we select other hues in
which the two components are present in

'5

equivalent, if not in equal, proportions, the
force of our argument is, I think, increased.
Such hues are yellow (red + green), seagreen
(green + blue), and purple (blue -f- red). Now
it will be generally conceded that the contem-
plation of an unmitigated yellow is less dis-
pleasing to the eye than that of orange or red,
seagreen than green or blue, and purple than
violet. Such comparisons are, I confess, very
difficult to make, and, without special con-
trivances for excluding all disturbing elements,
they are quite inconclusive. With the pig-
ments commonly employed for such compari-
sons, the varying amount of white light they
reflect, and the presence of chromatic elements
other than those we desire to compare, do
more than imperil the accuracy of the result.
And then the presence of backgrounds and
surroundings modifies the effects produced ;
these depend also very much upon the manner
in which the coloured substances used are
illuminated. Even with the spectrum-colours
themselves a just comparison is by no means
easy. I am reluctant to illustrate my argu-
ment by citing the case of certain sounds,
because the analogy between colour and sound
has been frequently pushed to unwarrantable
lengths ; yet I think I may say that an abso-
lutely pure loud note of uniform pitch, without
the simultaneous presence of those harmonious
constituents which combine to form its timbre,
would be no more pleasing to the ear than a
pure full colour of elementary character is to
the eye.

Hitherto, I have been speaking of full
colours, colours, that is, which are intense
and saturated, and also pure or unmixed with
white. The current theory of colour sensation
does, however, assume that there is some
stimulation of all the three kinds of optical
fibrils caused by each colour, however pure.
Still the sensation, say, of green is caused by
an immensely preponderating stimulation of
the green fibrils over that of the blue and red
fibrils. A corresponding statement may be
made with regard to the sensations of blue
and of red. In these instances, the fibrils
not necessary to the production of the two
sensations in question are even less stimu-
lated than in the case of green, a difference
which has been attributed to the median
position of the latter hue in the spectrum and
therefore to its intermediate wave-length. It
is argued that the green nerve-fibrils which
respond to the vibrations of green light are
capable of responding, to a slight extent, to
the vibrations of light on either oide of green,

but that the red nerve-fibrils are somewhat
less affected by the impact of the vibrations
at the other end of the spectrum, and vice
versa. In confirmation of this view I may
mention the curious observation that the last
colour to disappear when a continuous and
normal spectrum is gradually reduced in
luminosity is the green. That it appears
finally as a grey, having but a faint tinge of
green, is quite in accordance with theory.
And it may not be out of place to observe in
this connection that the full, nearly normal
green, as represented by the pigment emerald
green, is much more luminous than the full
red as represented by vermilion, probably in a
ratio approaching 2:1.

One reason, then, why pure white and the
infinite series of tones of grey do not offend
the eye as do certain pure, full, and luminous
colours, may be traced to the equal stimulation
of the three sets of optical fibrils which they
cause. Tints of any colour— the full colour
lightened with white ; shades of any colour —
the full colour darkened with black ; broken
tints of any colour— the full colour dulled with
grey, come into the same category ; for the
stimulations they cause, though not equable
like that of white and of grey, either approach
equality, or are less energetic and one-sided
than those of the full normal hues.

We are now in a position to offer some ex-
planation, inadequate though it be, of the
peculiar satisfaction afforded to the eye by
one of the characteristics of beautiful colour.
We call it " throbbing," "pulsating," "pal-
pitating." This quality may be imparted to
it in several ways, by which its dead uniformity
is broken, or its criant effect is subdued. The
fluctuation in question may be caused by-
Light and shade from inequalities of surface
or irregular scattering of light.

Varying surface-reflections.

Passage from translucency to opacity.

Varying depth of hue.

Variations of hue limited to the "small-
interval."

Several of these causes often conspire to
produce throbbing colour in one and the same
case, and they will be best illustrated by a few
concrete examples.

The case of rough paper which has received
a uniform wash of water-colour furnishes us
with an illustration of the effect of light and
shade as modifying hue. We will assume
that the pigment has been so equally dis-
tributed that the hollows of the surface have

i6

not retained more than the elevations. .Then
we shall find, in the case of many pigments of
full colour, that the hue in the hollows differs
from that in the prominences. For instance,
if French ultramarine has been used, the
hollows will show a colour tending slightly
towards violet, while the pigment on the
elevations will exhibit a purer blue. In prac-
tice the effect will be augmented by the
varying amounts of pigment which have
lodged in different parts, the hollows retaining
more than the elevations. The same wash of
the same pigment on smooth paper will pre-
sent a different appearance ; and in order that
the effect of the pigment on the latter may
rival that on the rough, it will be necessary to
introduce a second pigment, perhaps even a
third. That the difference in question is not
one of light and shade only may be proved by
comparing the two kinds of coloured sur-
face in a suitable instrument constructed
on the principle of Lovibond's tintometer,
and admitting of the examination of small
isolated points or patches of colour — the
chromatic analysis of these points will be
found to vary not merely in depth but also
in hue.

That the majority of painters in water-
colours should prefer a comparatively rough
paper for their work to the smooth sameness
of Bristol board is not surprising. One admits
of happy accidents of hue, and even involves
them ; the other demands the introduction of
laboured effects, while the result, at best, is
inferior.

Our second cause has been called " varying
surface-reflections." To illustrate it we will
take a piece of porcelain. The glaze may
vary in thickness ; parts may be more perfect
in gloss than other parts ; it may even exhibit
iridescence. A colourless glaze may produce
modifications of the hue of the colours beneath
it by reflecting from its surface in var) ing
amounts the coloured rays which fall upon it
from surrounding objects, or by itself decom-
posing the white light by which it is seen. In
both cases its colour will be variously changed
even when that colour was originally flat and
uniform. I have in my mind, at the moment,
a vase of old Chinese egg-shell porcelain of
an amber-yellow. This colour is due to lead
antimoniate, a substance which has become
partially incorporated with the colourless glaze
in the kiln. The surface reflection varies in
intensity ; in some places it reflects the in-
cident white light unchanged, in others it
shows iridescent colours, chiefly purple and

greenish blue. The fluctuation of colour is
produced by the mingling of the light which
has passed through the yellow pigmentary
layer and has then been reflected to the eye
from the white subjacent ground with the light
directly reflected from the glossy surface.
From different portions of the surface we
receive light of at least four different hues,
namely, the original yellow; the same colour
diluted with white ; pale orange-yellow, where
the original yellow hue is modified by com-
mixture with the purple of the iridescence ; and
pale greenish yellow, where it is similarly
mingled with the greenish blue rays from the
glaze. Of course, in some positions, the
iridescence alone can be perceived, just as
in others the original amber-yellow alone is
visible.

When a colour varies in transparency or
opacity we get a third case of fluctuation.
This variation is generally accompanied by
variations in depth and hue, but it alone
suffices to. bring about the quality we are con-
sidering. Let us take an illustration from oil-
painting. We shall require two pigments, one
transparent, the other opaque. It will be diffi-
cult to match them exactly as to hue, but
vermilion and a kind of madder red will answer
sufficiently well. We spread a very thin layer
of the latter upon a white or colourless ground;
some distance off we prepare a similar patch
of the vermilion ; between the two we make a
gradated mixture of the two pigments, and
thus obtain the passage of a transparent colour
into an opaque colour ; yet the whole coloured
area is a red, and a red which can be repre-
sented by one and the same numerical expres-
sion when referred to the wave length of the
corresponding hue in the spectrum. But in-
stead of employing these two pigments in the
diagrammatic fashion, let us use them in
pictorial or decorative painting by placing
the two pigments side by side, or glazing
with the translucent pigment, or scumbling
with the opaque upon the same background.
Only in some such way is it possible to suggest
the throbbing colours seen in many natural
objects, feathers, precious stones, marbles,
&c. Of course in representing these we are
not restricted to the use of a pair of colours of
the same hue and depth ; the limitations we
imposed upon ourselves were intended merely
to simplify the consideration of this third
source of pulsating colour. Let me add that
the fluctuation here is between two red lights,
one of which reaches the eye after passing
twice through a translucent red medium, while

■7

the other is produced by selective reflection
from the surface of an opaque red solid. I
say opaque, although in point of fact the inci-
dent light, or rather a part of it, plunges to
some depth amongst the red particles of
vermilion and passes through them, a chro-
matic selection being thereby made. And it
is really in this manner that the light finally
reflected to the eye by this pigment, having
lost in selective absorption some of its coloured
constituents, is red. I might cite many ex-
amples of the fluctuation of colour produced
by the passage of an opaque colour into a
transparent one or by the juxta-position of two
closely related colours, one of which is due to
surface-reflection, the other to reflection from
an appreciable depth : the sea, for instance,
when it shows the local blue colour of the water
itself, and, associated with this, the reflected
blue of the sky. The effects produced in cer-
tain textiles, such as linen and silk, damasks
and velvets, partly belong here, but they are
also connected with the variations due to
inequalities of surface and to differing powers
of reflection.

The fourth and fifth causes of the throbbing
colour are to be traced to variations in depth
of one colour, or to slight variations in hue ;
they may be fitly considered together, for in
actual examples they are generally associated.
Again, we may have recourse to Chinese
porcelain for an illustrative example. There
is a beautiful ceramic colour known as dragon's
blood ; its hue varies much in differeut speci-
mens, and even in the same specimen, but the
red element always largely preponderates. The
red sometimes approaches a red-purple, that
is, has a small proportion of blue in it ; more
often it oscillates between a crimson and the
colour known in France as garatice doree. The
crimson varies in depth. Sometimes it is a
full and saturated colour, sometimes it passes
into a paler tint, a kind of rich rose-pink. The
garance doree is a translucent red, which,
when diluted, has a somewhat golden-red or
orange-red hue. When we examine a good
specimen of this kind of porcelain, we may
sometimes detect the presence of all the colours
I have named, as well as of their intermediate
or passage tints. These hues are separated
from one another in the chromatic circle by a
very small number of degrees ; they illustrate
the effect of the collocation of colours differing
by what is called " the small interval." J f the
circle be divided into ioo*, their extreme range
will amount to about io°. And the most ex-
treme colours are not in juxtaposition, for one

slides into another by imperceptible gradations
both of tone and of hue. And a range so
extensive as io° is by no means necessary in
order that the prosaic flatness of ordinary
colour may be avoided ; possibly a range of
2° or 3° may suffice, even if it be unaccom-
panied by variation of tone. Coloured glass,
of good quality, may be chosen as affording
an excellent illustration of the points now
under consideration, but I reserve it for
fuller consideration later on, although I may
observe, in passing, that the fluctuations in
colour of the transparent enamels on porcelain,
and of glass, arise from the same causes,
only that, as we look at porcelain and through
glass, surface iridescence in the latter plays
either no part, or a very small part, in the
production of the phenomena in question. Of
course I exclude from consideration glass in-
tended to be looked at, that is, those vessels
of glass, coloured or uncoloured, which owe
their beauty of hue to the iridescence of their
surface.

I have now shown, I trust, that throbbing
colour is produced in more than one way, and
that it is agreeable to the eye, even when we
are unconscious of the elements that concur in
its production, at least in part through physio-
logical causes. I do not think these causes
furnish an adequate explanation of the pleasure
such colour affords, for its charm is doubtless
connected with certain fluctuations of our
judgment when any element of contrast in
colour or in tone comes in, with the agreeable
associations it recalls, and with its complexity
and mystery. We do not see it all at once,
nor do we instantly grasp its full meaning.
Its effect may be likened to that of an intricate
architectural vista, or of a range of distant
mountains, or of the cadences in the exquisitely
chosen language of a beautiful poem. I can-
not but think that throbbing colour is as neces-
sary to chromatic decoration as to pictorial
art. You are familiar with the numerous and
most useful coloured designs which illustrate
Owen Jones's "Grammar of Ornament."
Translate these flat and terribly prosaic figures
from the uniformly coloured paints with which
they have brin printed into those precious
materials— silk, and marble, and glass — which
present delicate variations of tone and of hue,
and at once the chromatic combinations are
ennobled, changed from awkward prose into
refined poetry fraught with varied interest. It
is just in the same way that the sympathetic
painter of landscape treats the clouds and the
sky. He is not satisfied witli white and grey for

|8

the former, and with blue for the latter. He
does not merely gradate his tints, but suffuses
his whites and greys with faint suggestions of

amber and apricot, and rose ; and varies his
azure with hints of lavender and of aqua-
marine.

LECTURE III— Delivered March 31, 1890.

SYLLABUS. — Balance. Proportion, and Distribution of Colour — Interchange and Counterchange of Colour — Colours of some
Natural and Artificial Materials— Decorative and Pictorial Colour— Colour in Relation to Architecture and Sculpture.

Distribution ami Balance of Colour. —
When the constituents of a set of colours to
be associated together have been decided on,
it still remains to determine their sequence
and the relative areas they shall occupy.
Such triads as were given in the last lecture
may be made disagreeable or pleasant by
altering their distribution or their proportions.
The triad red, gold, blue, affords as good an
illustration of this fact as we could wish. If
we make the red and blue come into contact
throughout the design in equal areas, and if
we then spread over the surface a spray of
gold foliage in such a way as to employ, in
masses of considerable extent, just as much
gold as red and as blue, the result will be
crude and unsatisfactory. But we may develop
the varied beauties of which this group is
susceptible by modifying the arrangement and
proportions of its constituents in scores of
different ways. Let us devise one such way,
of which the plan shall be — separation of the
blue and red by gold, and the use of these three
chromatic elements in the ratio 7 : 3 : 1 : so we
make our blue ground cover 7-nths of the area
to be decorated ; our arabesque and its curves,
or our spray and its foliage, will be of red,
3-itths, bordered throughout with gold, i-nth.
I need not tell you how much we may modify
and improve the result by the beautiful curves
of our spray, and by half entangling patches
of the ground within its borders. And this
end may be reached by many other methods
which I cannot stay to describe.

One of the tetrads named in the last lecture
will serve to illustrate these matters of arrange-
ment and proportion. The example is taken
from an early Irish illuminated missal. The
four elements are silver, gold, red, and black.
The proportionate areas are 4:4:1:1. Silver
and gold quatrefoils alternate. They are

bordered by red in the form of a line of
uniform breadth ; the residual space is covered
with black, but this is broken up into such
small portions that it does not look heavy,
while the continuous red outline serves at once
to bind the whole design together, and to
prevent too strong a contrast of tone between
the black and the gold or silver. You will
be able to think of many an arrangement of
these four elements which would prove less
felicitous, and, doubtless, of others equally
good.

I have before spoken of the principle of the
"dominant hue." It will be easy to carry
out this principle by modifying the area of the
elements in our triads. In this way one of our
"cold" triads — such as bluish-green, violet,
red ochre, may be made " warm." In such a
case we shall have to increase greatly the area
occupied by the red ochre, and to proportion-
ately diminish that of the other two elements
of the group. We may, for instance, prepare
a large design of foliage of red ochre upon a
ground of bluish green, separating these two
colours by a contour line of violet, adding
perhaps some fine veins of violet to our large
leaves. But it must be noted that such
arrangements often become easier and more
pleasing when we are at liberty to introduce a
fourth element, even if that element be but
white or other neutral. A fine old Rhodian
tile in my collection affords a case in point.
It has a ground of a red ochre hue occupying
5-ioths of the whole area. Upon this a design
of white foliage, highly conventionalised,
spreads in a connected and symmetrical
fashion, and covers 3-ioths of the surface. The
remaining 2-ioths are shared between a full
blue and a slightly bluish green in nearly
equal proportions. These colours are dis-
tributed upon the white portions of the design

19

chiefly in the form of veinings. There is also
a fifth but inconspicuous element in the chro-
matic scheme, namely, a delicate contour line
of deep grey. The effect of the whole arrange-
ment depends upon the dominant red hue of
the ground, but this is refined and lightened
by the net of white foliage, which, in its turn,
gives strength and purity to the graceful curved
forms of blue and of green which it encloses.
The uses of colour-schemes, in which a do
minant hue is present, are frequently of the
utmost importance in hangings, papers, and
wall-decoratinns. These have to perform the
office of a background, and to set off one or
more conspicuous chromatic elements in the
room. If you want to display properly some
fine pieces of blue and white porcelain, your
walls must not be so coloured as to interfere
with their effect, but rather to enhance it both
as to tone and hue. A flat buff or grey may make
your porcelain too conspicuous ; a bright blue
and a pure white will clash with it ; a strong
red will overpower it. A scheme of colouration,
which in its totality produces the effect of a
greyish olive green, half way in depth of tone
between the white and the blue of your vases,
will probably be found to answer well.

When it is not desired to direct attention
specially to the colours of the objects in a
room, then it is allowable to gather into
certain parts of your wall decoration the more
telling of the chromatic elements which enter
into its composition. You accentuate the
design by focussing the more saturated and
luminous colours in the more important parts
of the design.

From what I have stated thus far, I think it
will be evident that, even if we could attain
such a chromatic balance as would produce a
true " neutralised bloom " on our walls, or in
the carpets on our floors, we should still have
left out of our scheme the more important
chromatic elements belonging to our move-
able furniture and ornaments. In fact, the
system of contrasts of hue and tone — contrasts
more or less subdued — is the only one by
which a real unity of effect can be reached ;
at the same time 'that due prominence and
relief are secured for those parts or objects
where prominence and relief are needed.

Returning once more to our triads, I would
mention a principle of arrangement often
adopted with success, namely, the separation
of two nearly related tones by a tone which is
darker or lighter. Sometimes this sequence
brings together chromatic elements which
clash, and, of course, it must then be avoided.

Still, the rivalry of two colours having the
same degree of brightness is frequently un-
pleasant, especially where they are both
strong. The triad of full red, gold, full blue,
illustrates these observations. The use of a
triad of three full colours is rarely successful,
and can no more be tolerated than the
presence of three trees of equal size and equal
prominence in a landscape picture. Where
three colours having three decidedly different
tones are used in one scheme, we are almost
compelled to arrange them in the order of
their depth — dark, medium, light.

You will, I think, excuse the elementary
and fragmentary character of all these re-
marks of mine on distribution and balance of
colour ; no one can be more conscious of their
imperfection and of their commonplace cha-
racter than I am.

Counterchange and Interchange. — In
heraldic colouring the principle of counter-
change is frequently adopted. Ermine is
white spotted with black ; counter-ermine,
black spotted with white. Suppose a shield
of arms divided diagonally from right to left,
and bearing a star. Where the field is azure
the star is argent, where the field is argent
the other half of the star is azure. In the
latter half of the 17th century the same
arrangement was adopted in textiles and
embroideries, great skill being shown in the
passage from a design, say, of red upon a
blue ground to the same design in blue upon
a red ground. I have noticed that in the more
successful of these counterchanged patterns
the areas occupied by the pattern and the
ground are about equal. When these designs
were executed in "cut" work there was no
loss of material, for the pattern cut out from
one part of the design formed the ground of
the next. In the cut-cloth-work of Resht in
Persia, a similar mechanical method of pro-
cedure was adopted, but the counterchanges
were not effected in the same piece.

The principle of interchange may be illus-
trated by the alteration of a set of tones of
greenish yellow with a corresponding set of
tones of violet. These tones may be so
arranged that the series begins with the
deepest tone of each colour alternated, the
rest following in regular sequence, or the
deepest tone of one colour may be followed by
tin lightest tone of the other, and so on ; but
a satisfactory effect is difficult of attainment.

Colours of Minerals. — I have time to refer
to a very few only of the characteristics of
coloured minerals. The clours of some of

20

the most beautiful precious stones are in a
measure dependent upon the phenomena of
dichroism which they exhibit. The ruby,
sapphire, emerald, and tourmaline may be
named in this connection. Quite apart from
the prismatic decomposition of light which
these gems effect when facetted, they show,
with more or less distinctiveness, the twin
colours of dichroic bodies. The ruby exhibits
in the same stone a pure red and a crimson or
carmine-red; the sapphire a straw-yellow and
a pure blue ; the emerald a bluish-green and
a pure green ; and the tourmaline broken
tints of yellowish-green and reddish-brown.
No monochroic substance, such as paste,
enamel, or glass, can give rise to the variable
fluctuations of colour seen in these dichroic
gems. The peculiar chatoyancy of Labrador
spar has also not yet been imitated in any
artificial material. The beauty of lapis-lazuli
and of many agates and jaspers is no doubt
partly dependent upon the pulsing of the
colours they exhibit, partly upon a certain
measure of translucency which they possess.
This quality of trauslucency is also seen in
many marbles, especially when they are
polished ; its absence from some artificial
materials (such as terra-cotta) is perhaps one
of the reasons why it is difficult to associate
the two classes of materials together with
satisfactory effect. When both are in small
pieces, and especially when they all receive
together the same polish, the incongruity in
question does not attract notice. The tessel-
lated Roman pavements found in this country
often afford excellent illustrations of this har-
monising of natural with artificial materials —
marble and stone, with pottery and even glass.

The picturesque coloured markings and
veinings of many marbles are best developed
by polishing, and best seen in surfaces of con-
siderable size, If these surfaces are not plane,
their curvature should not be complicated with
flutings or other sculptured treatment. A
carved ornament in such marble is a mistake,
at least whenever the coloured markings are
decided. Man's art in relief spoils nature's
decoration in colour. For this reason the
carved work in Derbyshire alabaster, for which
the late Sir Gilbert Scott showed so much par-
tiality (witness many a reredos which he de-
signed) is most unsatisfactory. Over and over
again a dark veining makes a prominent part
of the carving appear to recede, while a bright
and pale patch throws a hollow or recessed
detail into conspicuous relief.

Colours of Animate.— The most brilliantly

coloured of animals — birds— naturally attract
our attention. I do not pretend to champion
all the chromatic arrangements of nature. Yet
two circumstances must be taken into account
in considering the colours of animals and
plants. These colours and colour-associations
must not be considered apart from their natural
entourage and the conditions of their existence.
And, secondly, if we are to follow the leadings
of evolutionists, it cannot be maintained that
the colouration of plants and animals in a
scate of nature is connected with the chromatic
tastes of man. Of course, under domestication,
the hues of both are greatly modified by pro-
cesses of artificial selection and of treatment.
Let me direct your attention to the colours
of certain humming birds as typical examples
of peculiarly powerful colours, many of them
belonging, in the intensity of their regular re-
flection of light, to the metallic group, but
being more strongly coloured. One of these is
a Central American species, and is known as
the "garnet" ( Lamprolcsma Rhami). It is
commonly spoken of as showing a brilliant
patch of garnet-red upon its throat and breast.
But in reality this hue is not invariable and
uniform. In the constant movements of the
living bird it must be ever changing with its
changes of position. As you examine a
mounted specimen you will see at first only a
metallic crimson ; but when your view-point is
altered, all the passage-tints between a rose-
madder and a greenish yellow follow in their
regular spectral sequence. This sequence,
which will be observed, I believe, in all similar
cases, is connected with the mode of produc-
tion of these metallic colours. They do not
arise directly from the presence of actual
colouring matters or pigments, but from the
minute optical structure of the web. Another
hummingbird, a " train-bearer" (Cynantlius
forficatas), has its conspicuous tail-feathers
adorned with metallic violet ; but this violet
passes in some positions and in some parts
into a splendid greenish blue, every inter-
mediate hue being present. So with one of the
"comets" (Sappho spargaiuirus), the tail-
feathers sparkle with a golden bronze, ranging
from orange to red-orange. Another "comet"
(Sappho phaon), is similarly decorated, but
the colours range from madder-red to red-
purple. Another bird, the " ruby and topaz "
(Chrysola)iipis mosquitus), has on its head a
patch of metallic crimson, passing into orange
when viewed from some positions ; similarly,
the fiery metallic orange of its breast passes
into yellowish green in some positions.

I cannot refrain from citing the case of
another humming-bird, one of the "sapphires "
{F.ucephala cceruled). Here the head is de-
corated with a colour not of the metallic order,
but still very rich, almost the colour of French
ultramarine blue. But on the breast of this
bird a metallic colour re-appears ; it is a fine
blue, passing into an equally fine green. I
have said nothing of the brownish-black back-
ground which shows off all these splendid
hues, because I want you to notice more par-
ticularly that these dazzling metallic colours
are never really uniform and flat, even though
their tones may be equivalent, for they show
large ranges of hue, not capricious, but follow-
ing the orderly sequence of wave-lengths in
the spectrum. I must omit all reference to the
colouring of butterflies and shells, about which
I had much to say, in order to pass on to the
vegetable kingdom.

Colours of Plants. — The texture and trans-
lucency of flowers and leaves greatly modify
the hues of the colouring matter they contain,
and of the light which they reflect. The
peculiar glistening of the cell walls in the
coloured tissues of flowers is a case in point ;
it is sometimes erroneously spoken of as crys-
talline. The distribution of the various colour-
ing matters of flowers in the cells, when the
coloured tissues are examined under the micro-
scope, is seen to be much less regular than
might be imagined, and is doubtless one of
the causes of the peculiar chromatic beauty
of many flowers. In speaking of the colours
of plants some reference must be made to the
green of foliage. If the fully developed leaves
of a forest tree in summer sometimes tend
towards a certain heaviness and monotony of
hue, still there is, even in this case, a degree
of variety present. Some leaves are in shadow,
some transmit the incident light, some reflect
the blue of the sky or the grey of clouds, some
display in perfection their local colour, and
some show in different parts the various hues
just indicated. And then, too, it must not be
forgotten that the green colour of foliage varies
with its age, and that it always possesses a
singular characteristic which distinguishes it
from other greens ; for if we place a spray of
green leaves upon a piece of green cloth or of
green paper resembling it in general hue, we
shall yet find that the natural pigment differs
materially from the artificial in its chromatic
constituents. To prove this you need not have
recourse to prismatic analysis ; you have only
to illuminate both surfaces with red light, or
to inspect both through a piece of red glass,

to see a very striking difference, especially
noticeable with the yellowish-green foliage of
young branches.

I regret that time fails me, or I should have
liked to direct your attention to the colours
and textures of wood, and of various vegetable
fibres.

Colours of Artificial Materials and of
Fabrics.— I have already named glass as
affording illustrations of the throbbing or
vibrating of colour. In the examples of
coloured glass which I now show you by means
of the electric lantern, we may observe the
various causes which conspire to produce the
artistic charm of really beautiful glass. They
include not merely fluctuations in hue and
tone, but the presence of bubbles and blebs, of
stria;, and of solid and comparatively opaque
particles. All these things serve to prevent
the direct emergence of uniformly-coloured
beams of light from the material. I have
mounted some of these specimens of glass in
pairs, putting side by side, for instance, a
piece of builders' blue glass and a piece of
the beautiful blue glass made by Messrs.
Rust and Co., of 353, Battersea-park-road,
S.W., to whom, indeed, I am indebted for a
series of choice specimens, illustrating the
poetry of coloured glass as distinguished from
its most ordinary prose. In this connection,
it is interesting to observe that the good old
stained glass, like these fine samples of Messrs.
Rust and Co.'s manufacture, is immediately
discriminated from the common-place kinds in
a strong light. When the sun or the electric
light shines through the former it becomes
itself illuminated, but throws no definite
coloured images upon a white screen suitably
placed ; while the optically perfect but wholly
uninteresting flat-coloured modern glass re-
produces by transmission its chromatic design.
By this test the modern " restorations " in an
ancient window may frequently be recognised.
I have no time to tell of the colours of other
artificial products, of enamels and of porcelain,
for example. But I cannot refrain from direct-
ing your attention to the colours and patinas
of certain metallic alloys. The Japanese are
masters in this particular line. Their shibuichi,
or grey bronze, consisting of copper alloyed
with considerable but varying proportions of
silver, gives a series of yellowish greys ; their
shakudu, copper containing a small pcr-cent-
age of gold, when heated in suitable pickling
baths, is susceptible of receiving a blackish
patina, sometimes violet-black, sometimes
bluish-black. Their bronzes, into which lead

11

and antimony, as well as tin, often enter, fre-
quently present very beautiful effects of colour
upon their surface. But perhaps the " red
copper" of the Japanese metal-worker is as
beautiful as any alloy. It is pure copper, the
surface of which has become coated with a
strongly adherent and coherent film of red
sub oxide. This film is tough and strong, and
varies in colour in different specimens from a
crimson-red to a reddish brown. This patina
is often translucent, if not transparent, and
shows a fine gloss like that of lacquer. Messrs.
Christofle, of Paris, and Messrs. Tiffany, of
New York, have achieved a measure of success
in their attempts to reproduce some of these
beautiful Japanese patinas, but there remains
a large field for their employment in artistic
metal-work, in the form of flat inlays and
reliefs.

I cannot do better, in order to illustrate the
tone and colour effects which may be produced
by the association of different textile fibres,
than show you in the light of the electric
lantern a series of Japanese brocades and of
Persian embroideries. The "shot" silks, the
varying sheen produced by the different direc-
tions of the threads, and the association of
lustrous with comparatively dull materials — all
conduce to the beauty of the effects.

Pictorial and Decorative Art. — It has often
been pointed out that colour is an end in
decorative art, a means in pictorial. This
almost amounts to saying that decorative
colour is without meaning ; expresses neither
the ideal nor the real. I cannot but think that
this view is far from being correct. Let us
contrast and compare the colouring of a land-
scape and that of a decorative fabric. In the
picture great use is made of gradation of tone
and hue, in order to represent atmosphere, the
play of light and shade, and different planes ;
the artist, moreover, is restrained in his use of
full and saturated colours, nor does he attempt
to represent all he sees, but makes such a
selection of the materials at his disposal as,
without violating nature, shall best serve to
realise his impression of the scene. Form is
with him of equal importance with colour,
while symmetrical arrangement and repetition
of similar elements are never made evidently
conspicuous. The designer of a fabric, one,
say, for use as a curtain, avoids the use of
those contrivances by which the flatness of the
surface would be destroyed ; gradation of hue
and tone, though admissible, plays a quite
subordinate part in his work. He is at liberty
to use the most saturated and intense colours,

provided only he can so employ them as to
produce a pleasing harmony, and at the same
time be in accord with the nature of his
material and the purpose to which his design
is to be put. Moreover, the decorator selects,
as does the landscape painter, but he rejects
much more than he absorbs. To nature he
goes for motives, but individualisation assumes
the last place, generalisation the first. Form
and colour are not with the decorator of equal
moment, sometimes the former being par-
amount, sometimes the latter. Usually, the
less pronounced his colour the more depend-
ence does he place upon his form, although in
complex chromatic sshemes also he has often
to rely greatly upon trenchant contours in
order to avoid confusion. The notion that
bad or weak drawing is permissible in deco-
rative art ought not to be accepted for one
moment. The very fact that the boundaries
between two colours in ornamental designs
frequently need to be sharply accentuated by
means of well-defined outlines should suffice
to demonstrate the necessity of thorough
draughtsmanship. And, lastly, the decorator
makes large and frequent use of symmetrical
and repetitive arrangements.

If I am right in this comparison and con-
trast, then the distinction between decorative
and pictorial work is, like many another dis-
tinction in the domain of art, one of degree
rather than of kind, quantitative rather than
qualitative. The same elements are at the
disposal of the decorative and of the pictorial
artist, but they must be employed in different
proportions. And here I am in accord with
the view that there are an infinite number of
possible transitions between the colouring that
is most fully pictorial and picturesque, and the
colouring that is most purely decorative or
conventional. Much, but by no means all, of
Japanese decorative work in colour illustrates
some of these transitional forms.

Colour in Relation to Sculpture. — In the
very few words on this subject for which I can
find time I should wish to draw your attention
to these two points, namely, the surroundings
of sculpture in relation to its colour, and the
material of sculpture in relation to its colour.
It seems to me that when sculpture is destined
to form an integral part of an architectural
scheme, its tone and its hue must not be such
as to interfere with the unity of the plan, but
rather to enrich and vary it. And it may
easily happen that an association of sculpture
in marble with an edifice in stone may be
harmonious when both marble and stone are

*3

fresh, but may in course of time accord less
happily, as the latter darkens with age much
more considerably than the former. The same
effect is produced when a glaring white maible
monument is intruded into a building deeply
toned with the rich warm grey of centuries.
So, on the other hand, a statue in bronze may
be too deep in tone to harmonise with new
clean cut stone, but, as years pass, the
startling difference between them will abate.
The case is different where the building and
the sculpture it contains are not intended to
fuse into one organic whole. Under such
circumstances, the building must be sub-
ordinated to and even contrasted with its
contents ; it becomes a background so
coloured and so arranged as to emphasise
the statuary it protects ; and in considering
the vexed question of the artificial colouring
of works in sculpture, it will be well to glance
for a moment at the four chief kinds of material
out of which such sculpture is generally
wrought. These materials are terra-cotta,
bronze, stone, and white marble. It will be
owned that the critical eye accepts without
hesitation the refined colouring which the
exquisite terra-cotta figures from Tanagra still
so frequently retain, as well as the varied hues
of the patina on statues in bronze. In the
latter case, we go so far as to dislike what
may be called the natural hue of the metal,
unless it appears to a limited extent only in
such parts of a figure as might be supposed
to have lost their artificial or altered surface
colour by a process of attrition. But when
we leave artificial materials and pass on to
those which are natural our attitude as to the
problem of colouring differs. While there
seems to be an innate congruity between arti-
ficial materials and artificial colouring, natural
materials lend themselves less readily to chro-
matic treatment. Perhaps there are here two
causes at work, one being our satisfaction with
and appreciation of the natural beauty of the
tones and hues of native substances, the other

. being the difficulty of bringing our added
colouration into harmony with that which is
natural. This view is supported by the ob-
servation that the addition of colour to the rarer
and choicer materials, such as statuary marble,
is more generally resented than the similar
treatment of the commoner and less interesting
kinds of stone. But is there not a third and
more potent cause ? Does not pure white
marble, with its slight translucency and its
beautiful crystalline texture, lend itself so per-
fectly to the presentation of ideal forms that
even the faintest suggestion of realistic colour

j may look like sacrilege, and may only too

' easily lapse into vulgarity ?

In architecture, too, the poorer and less

' interesting the material, the more useful does
the addition of artificial colour become. Yet
it may serve, on the one hand, to emphasise
the poverty and weakness of bad contours and
mouldings, though, on the other hand, it may
be made to accentuate and to ennoble forms
which are in themselves beautiful. Its dis-
tribution and arrangement are as important as
its tone and hue. It may destroy the unity of
an architectural scheme, or it may serve to
fuse discordant or fragmentary elements into
one harmonious whole.

Excuse, ladies and gentlemen, the all too
casual and imperfect discourses which I now
conclude. Yet I trust my disjointed utter-
ances may not prove to be wholly valueless,
for they may serve, if no other end, that of
directing your attention and guiding your
observation to some of the infinite resources
and intricate delights of chromatic combina-
tions. I venture, moreover, to express my
confidence that the scientific study of the
glorious sheaf of spectral colours which we
have been making together will not lead you
to agree with the poet Keats when he sang —

" Do not all charms fly
At the mere touch of cold philosophy ?
There was an awful rainbow once in heaven :
We know her woof, her texture— she is given
In the ilull catalogue of common things."

1'EIN-TIiD BV W. TROUNCE, 10, OOCOH-SCjrUJlE, FLEET-STaEET, LOMDOX K.O.

9