Action of Sunlight on Glass (1867)

{ Extract from offprint.}










{ notes 2008: Quote from W Colné 1880 “Mr. Thomas Gaffield, of Boston, who, as an ex-manufacturer, now an amateur, and a true lover of the art of glass making, has contributed many articles to the literature of glass making.” … “It is with pleasure that I call attention to the researches of Mr. Gaffield, because of the fact that so far we have had comparatively nothing published in this country in regard to the scientific investigation of glass.” All footnotes placed in-line in different colour }

T. Gaffield on the action of Sunlight on Glass, 1867 1
[From the American Journal of Science and Arts, Sept. and Nov., 1867]



The great attention now given to all the phenomena connected with light and heat may awaken some interest in the experiments in which I have been engaged for the past four years on the subject named at the head of this article. Perhaps I cannot better commence my essay, than by quoting from the “Proceedings of the Natural History Society,” (vol. ix, p. 847) an account given before that Society, of my experiments in 1863, and after I had been engaged in them only a few months.

“He believed that his experiments in connection with the subject were original as to their method and their extent, although it had long been observed in Europe that colorless or light-colored plate-glass had turned to a purple hue by exposure to intense sunlight. One case {* Journal of Society of Arts for Feb. 15, 1854. is cited of a change to a gold color; and one experiment recorded by Dr. Faraday, {† Dr. Faraday’s Chemical Researches. London, 1859, p. 142.} some forty years ago, proving that a light purple changed to a darker hue after eight month’s exposure.

“Other experiments are on record showing the action of glass of different colors as media in the transmission of light and of heat; but none, with the above exception, showing the effect produced on the glass itself.

“An experience of some twenty years in the window-glass business had only presented a few isolated cases of supposed change of color from this cause, which were attributed to some obvious defect in an article of inferior manufacture; but, within

T. Gaffield on the action of Sunlight on Glass, 1867 2

a short time, he had heard of the change of color in an article of superior manufacture, in a quantity of white plate glass, of which some lights had been broken out of a window in which they had been exposed to the sun.

“This fact coming to his knowledge led him to try an experiment with several specimens of plate, crown, and sheet glass, during the month of July last; which proved that a month's exposure to a hot sun would change the best white French plate and all white sheet glass, such as is used for photographs and engravings, to a color containing more or less of a yellow hue. The dark green and dark blue or bluish green did not experience any change; but any hue which approached a white, whether bluish, greenish, or yellowish white, turned to a yellowish color.

A second series of experiments, commenced in July, and continued three months, on some thirty specimens from France, England, Belgium, Germany, and the United States, only confirmed the results of the first; and a daily examination at first, and afterward from week to week, and month to month, revealed the interesting fact, that, even after a single day’s exposure to a July sun, the change of color will, in some instances of the lightest hues, commence.

“So remarkable was the change in a week, affecting nearly all the light-colored glasses, that he commenced a third experiment on the 6th of August which should speak for itself. He then exhibited to the Society ten pieces of French white plate-glass, four by two inches in size (all of which were cut from the same sheet), one of which showed the original colorless glass, and the others exhibiting the change of hue towards yellow, after exposure respectively of one, two, and four days; one, two, and three weeks; one, two, and three months.

“The changes in the first four days were slight; but the last specimens were so yellow as to exhibit a contrast very marked, and excited the interest of all the members present. That the color permeates the body of the glass, and is not confined to the surface, or produced by reflection therefrom, has been conclusively proved by grinding off about one-sixteenth of an inch from both surfaces and the four edges of a duplicate exposed specimen, which, after repolishing, still exhibited the same yellow color.

“The glasses exposed were all what are called colorless window-glasses, although they varied in tinge and hue from the whitest French plate to the darkest green English sheet-glass.

“An experiment for four months, from July to November, on really colored glasses, red, green, yellow, blue, and purple, showed no change except in the purple, which became slightly darker.

“The experiments were carried on upon a rough plate-glass roof, nearly horizontal, and which received the rays of the sun

T. Gaffield on the action of Sunlight on Glass, 1867 3

during the greater part of the day. In all cases, strips corresponding to those exposed, and cut off from the same pieces, were placed in the dark, to be compared with the other specimens after exposure.

“It will be noticed that the dark green, blue, and bluish green did not change. The color of the Belgian sheet (called German or French by glass-dealers in America), a yellowish or brownish green, did not change; and these were the only exceptions. All plate-glasses changed, except an inferior blue quality, and a superior crystal plate of a greenish color, made in Germany, and at the only factory which has not given up the use of potash for soda-ash.

“It is possible that a longer exposure of a year, or of years, might change every color in some degree.

“His inquiries, since he instituted these experiments, have brought out some fine specimens of Belgian sheet-glass from a house built three years ago, which had changed in some instances to a golden and in others to the well-known purple hue.

“It is his intention to pursue the experiments farther, with a view to ascertain the effects of sunlight during each month and season of the year; and also whether exposure to heat, air, or moisture alone, out of the direct action of the sun’s rays, will produce any corresponding change.

“Mr. Gaffield does not propound any theory to explain these changes of color, which, under our sunny skies, probably take place much more rapidly than in the different and less clear atmosphere of England.

“Some writers point to the presence of oxyd of manganese in the original composition of window-glass, and some to the oxyd of iron, as a chief cause.

“Some writers have peculiar theories about the different classes of the sun’s rays. Some may think the change referred to, a molecular or chemical one; and others, wiser than the rest, refrain from any explanation, waiting for a larger multiplication of experiments, and a greater accumulation of facts, before educing any satisfactory law of nature which governs these curious and interesting phenomena.

“Mr. G. makes no pretentions to any discoveries, unless it be to the very rapid change in glass observed in our climate in July, but only gives the result of his experiments, in the hope that the great interest now manifested in the subjects of light and heat may lead others to examine the matter, to repeat the same experiments in other countries, and to give the world the result of their researches, and enable the learned and scientific men of the age to explain this remarkable power and action of the sun’s rays.

“It should be remembered that he submitted his specimens to

T. Gaffield on the action of Sunlight on Glass, 1867 4

the most severe tests by placing them where they received reflected as well as transmitted light and heat. The change in glass, when glazed in the windows of our dwellings and stores, is so much more gradual, that it very rarely attracts the attention of observers, except in the marked variation from white to purple.”

In accordance with the intention above expressed, I have continued my experiments on this interesting subject, and under different heads will now give some account of their method and results.

My first experiment was with pieces of glass four by six inches, placed in a sash six by sixty inches long, in the grooves of which the specimens were placed, the sash being fastened together by wooden pins, and placed on a nearly horizontal rough plate-glass roof, which received the direct rays of the sun during the greater part of the day. In my next experiment, I placed some of the specimens directly on the rough plate-glass. These were carried on in summer and autumn. It was necessary to make a different arrangement for winter, when the fall of snow and the formation of ice might interfere with the full success of the experiments. I concluded that the best size for specimens was four by two inches, and I made some wooden boxes about 3/8 inches deep, 4Πinches wide, and of a length to fit the sills of windows facing the south, in the upper story of a Boston house. One of the windows was three-sided, looking east, south, and west. In these boxes, painted white, my specimens were placed in a nearly horizontal position, side by side, and, (after the loss of a few pieces, blown away by hurricanes and squalls), were secured in their positions by cords of twine or slight copper wire fastened at each end. I have arranged boxes on the roof in front of this window also, some being of greater depth in which to place pieces of plate and rough plate about one inch in thickness. All of these boxes are provided with covers, which are placed over the glass, and fastened by buttons, on the occurrence at any time of a snow storm. At all other times, the glass is exposed. Holes in proper places in the boxes allow the rain which falls to pass off easily. Every piece of glass is carefully marked by a diamond in an upper corner with its name, and if necessary, with some abbreviation descriptive of the experiment to which it belongs. This precaution is necessary to prevent mistakes, when a comparison is to be made of several kinds and colors, of exposed and unexposed specimens. In all cases, where an experiment is to be made with any description of glass, the pieces (4x2 inches in size) must all be cut from the same sheet, as there is frequently a slight difference in the shade of sheets from the

T. Gaffield on the action of Sunlight on Glass, 1867 5

same factory, arising from the difficulty of having the materials in their manufacture, and the circumstances attending the melting, blowing or casting precisely the same. Two pieces or more should be laid aside to show the original color, and to compare with the others, and thus to show the changes produced by exposure to sunlight for days, weeks, months or years. At the commencement of an experiment, say for instance, the exposure of white plate-glass for from one to twelve months, a neat paper box is provided, 4Œ inches long, 2 1/8 inches deep and wide enough to hold fourteen pieces, (two of the original color, and twelve exposed specimens) and a description of the contents is marked on its cover on an adhesive label. At the end of each month, a piece is withdrawn from exposure, carefully cleaned, and marked either with a diamond or by an adhesive label, and placed in the box. As “order is heaven’s first law,” it is peculiarly necessary in all observations upon the actions of heaven’s brightest luminary. By observing the above directions, much time will be saved, and at the conclusion of an experiment, everything is in shape for exhibition to friends at home, or students at a lecture room.

Of course a perfect arrangement could only be made when a perfectly flat roof or platform in an open field could be provided, and the sunlight could act with full force during every hour and minute of the day. But mine was sufficiently near this point to show very interesting results.

In one of my earliest experiments, I kept a record of the changes going on in the various kinds of glass, at first from day to day, and afterwards from week to week, and month to month. The following is a specimen:

Kinds of glass. Original color before exposure. Color after one-day’s exposure. Here were inserted other columns to show effect of exposure for 2 and 4 days, for 1, 2 and 3 weeks, for 1 and 2 months. Color at end of exp’t. (3 mos.)
French white plate, English crown glass. French white sheet. Belgian sheet. American sheet.      

I might give the names of all the different kinds of glass which I have exposed to sunlight, but I refrain because for the reason given previously, I have found different specimens from the same manufactory of plate, of crown, and of sheet-glass, sometimes to differ in shade, sometimes in result of exposure, and sometimes in both.

I will, however, give below, (not naming the particular manufactories), a general description of my experiments in 1863, when an exposure of thirty-three specimens for a few months in summer and autumn showed the following results.

T. Gaffield on the action of Sunlight on Glass, 1867 6

The first table shows that twenty-four kinds were found to be easily affected, and exhibits the time at which the change in color was first observed. The second table shows nine kinds which did not change in color during the same time.

Table of glasses easily changed.

Kinds of Glass. Original color before exposure.
5 of French white plate. White, with light bluish green tinge.
2 of German white plate. White, with light bluish green tinge.
1 of English white plate. White, with light bluish green tinge.
5 of English plate. Yellowish green.
1 of Belgian rough plate. White, with slight green tinge.
3 of French white sheet. White, with light bluish green tinge
1 of Belgian white sheet. White, with light yellowish green tinge.
4 of English white sheet. One light yellowish green and
three light bluish green tinge.
2 of American white sheet. White, with light blue green tinge.


The time at which change of color was first observed. Color after exposure.
Days. Weeks. Months.
1 2 4 1 2 3 1 2 3

1   4             Yellow or yellowish green.
1 1 Yellow or yellowish green.
1 Yellow or yellowish green.
4 1 Deeper yellow tinge
1 Yellowish green tinge.
3           Yellowish green tinge.
1 Yellowish green tinge.
1   1 1   1     Yellowish green tinge.
1 1 Yellowish green tinge.

5 6 1 3   1      

Table of Glasses not changed in three months.

Kinds of glass.
Original color of glass.
1 kind of German crystal plate. Light green.
1 kind of German plate. Dark blue.
1 kind of English rough plate. Dark green.
1 kind of English crown. Light green.
1 kind of American crown Bluish green.
1 kind of American sheet. Dark bluish green.
1 kind of English sheet. Dark green.
1 kind of Belgian sheet. Yellowish or brownish green.
1 kind of French sheet. Yellowish or brownish green.


Subsequent experiments with five of these kinds, (all which I could conveniently obtain), showed that an exposure of a year, or even less, would change all but an ordinary kind of American sheet, which was of a dark bluish green color.

The experiments which speak for themselves are the most satisfactory ones, that is to say, where one has not only the

T. Gaffield on the action of Sunlight on Glass, 1867 7

record made at the end of each month, but a piece of glass taken in and laid aside at the same time to show the actual color produced, and the truth of the records. It is very interesting to witness any one of these series of specimens, showing, as in one of white plate, a gradual change, commencing in a day or a few days in summer, from greenish or bluish white to a yellowish white, or light yellow, a deep and deeper yellow, until it becomes a dark yellow or a gold color ; and in some Belgian sheet specimens, a gradual change, commencing in a few weeks in summer, from brownish yellow to deeper yellow, yellowish pink, pink, dark pink, purple and deep purple.

There are several kinds of glass in which no perceptible change took place in three months, which were very sensibly affected by an exposure of a year.

Experiments. — I have given a general account of my first experiments in 1863, and a portion of the tables kept in my journal at that time. I might have given names and results in full and shown the actual effects and shades of color produced by exposure for a few months, on some thirty kinds of glass. But in my case, as in many novel and original investigations, the results of first experiments, and the theories based upon them, were modified by subsequent ones. I supposed that many kinds of glass not changed in three months would not change at all: that all which changed would take a yellowish color, unless by exposure of many years: that no color but some shade of yellow or pink would ever be produced in any kind by exposure to sunlight.

The experiments of 1864 and the two following years, proved to me that nearly every kind of window glass I had exposed, could be changed in one year; that a rose or pink color, (or some tint approaching them) could be produced in various kinds in a few months; and that some kinds of greenish white glass would, after exposure, assume a bluish tint or bluish white.

It may seem singular for one who has been a glass dealer and manufacturer like ourselves, thus to advertise what may be called a defect in his own wares. It might seem unkind to other manufacturers to expose the defects of their productions, literally before the light of day. But my scruples have been all removed, when I have noticed in a late communication of Pelouze (see Comptes Rendus, Jan. 14, 1867) the following statement: “I do not believe that there exists in commerce a single species of glass that does not change its shade in the sunlight.” As all manufacturers are in the same category, it will do no harm for me to repeat what Pelouze says, as the result of my experiments, and to affirm that a longer or shorter exposure to the direct action of the sun’s rays will probably change in some degree the color of all or nearly all kinds of window glass.

T. Gaffield on the action of Sunlight on Glass, 1867 8

I subjoin an account of an experiment carried on for one year with nine different kinds of glasses. These represent plate, crown and cylinder glass, the manufactures of both hemispheres and almost every shade and color of what are called colorless glasses.

I name only the nationalities, and not the particular manufactories of the glasses in any of my tables. I have this information recorded in my journal, and shall be happy to communicate it to any who may desire it.

Memorandum of nine different kinds of glass
exposed from Jan. 12, 1866, to Jan. 12, 1867.

Kind of glass. Color before exposure. Color after exposure.
French white plate. Bluish white. Yellowish color.
German crystal plate. Light green. Bluish tinge.
English plate. Light green. Yellowish green.
English crown. Light green. Light purplish color.
Belgian sheet. Brownish yellow. Deep purplish color.
English sheet. Dark green. Brownish green.
American crystal sheet. Light bluish white. Purplish white.
American crystal sheet. Lighter bluish white. Light yellowish green.
American ordinary sheet. Bluish green. No change.

The colors named above are given from an observation of the glass edgewise, by which one can see a body of color two or four inches in depth, whereas the usual thickness of the glass varies from one-fourteenth to one-quarter of an inch, and shows its color easily only by placing a white curtain or paper behind it.

I have tried several experiments showing the effect of sunlight during each month and each season of the year. At the end of the year, by the comparative depth of yellow or purple color produced in the various specimens, one can see the comparative actinic power of the rays during each month and season. The results proved that the actinic effect increased from January to July, and decreased after that month. The greatest effect during any season was observed in the summer, the least in winter, and that in spring and autumn was about alike, and midway between that of summer and winter.

Crystal or lead glass, and a piece of optical glass having probably very little, if any, manganese, changed not in two years. Perhaps a longer exposure may produce some change.

A rough piece of light colored window glass metal changed to a yellowish color in a year.

Colored glasses after two or three years’ exposure showed no perceptible change in any instance except a slight one in a single specimen of purple. Perhaps an exposure of many years may make a change in some other colors.

I have made experiments with artificial heat on glass in various ways, from exposure to the heat of a cooking range oven

T. Gaffield on the action of Sunlight on Glass, 1867 9

to that of a glass stainer’s kiln, without any change of color in the common colorless window glasses, while the same or similar specimens exposed to sunlight have been nearly all changed in a few months.

Specimens exposed in hot water for a month, in doors and out of sunlight, experienced no change of color, while similar ones exposed during the same length of time in the bottom of a dish filled with two or three inches of water out of doors, and to the direct rays of the sun, experienced a decided change, though only about half as much as when directly exposed, out of the water.

Being convinced that air, moisture, and artificial heat do not make any change of color, our experiments indicate that the change is effected by the actinic rays of the sun alone.

This actinic effect is cut off in some degree by every medium, by water as stated above, and even by clear glass, as a specimen exposed inside of a window or under another piece of thin colorless glass shows only about one-half as much change, as that exposed outside of the window or with no covering of glass over it. The amount cut off by colorless glass and by colored glass differs greatly with the difference of color.

The comparative power of glass of different kinds to transmit the actinic rays I have tested, by placing underneath pieces of each kind, pieces of easily changing glass, (white plate or Belgian sheet glass,) exposing them one year, and noticing, at the end of that period, the comparative depth of the yellow or pink color to which the under pieces had changed. The result of my experiments proved that the most easily transmissive of the colorless glasses were the English crown, French plate, two kinds of white crystal sheet made in Massachusetts, (from the celebrated Berkshire white sand,) the New Jersey sheet glass, one kind of English plate, and one kind of Belgian sheet, and about in the order which I have named them.

Of the colored glasses, the blue transmitted the most, the purple less, the red and orange the least, the glasses under these two and the yellow and green showing little or no change.

This last experiment proves the propriety of the preference given by photographers to blue glass for skylights, because it transmits the blue rays, which exert the most actinic power. But it may be added, that a colorless white glass, or bluish white, — if one which will not change by sunlight to a yellow or rose color, owing to the presence of manganese, or any other cause, — is equally good, as it will transmit all the rays, and among them, the actinic or blue ones. In proportion as any kind changes to a yellow or rose color, it will lose its power of

T. Gaffield on the action of Sunlight on Glass, 1867 10

transmission, and its value as photographic glass. I have seen specimens of the two kinds of white crystal sheet made in Massachusetts, before alluded to, which answered the demands of photographic artists. Of foreign glass, I have noticed a fine bluish white sheet, made lately without manganese, from a certain excellent manufactory in Belgium, and one kind of English crown glass.

Should plate glass be required, the most permanently enduring, or least likely to assume a yellow color, are a superior kind of white plate, made by the French and Belgian Plate Glass Companies, and an excellent quality of German crystal plate, made at a long established factory in Hanover.

I desire to say here, however, that it is not the place where any glass is made, which determines its good character, but the actual constituent materials and the superiority of its manufacture.

Manufacturers are frequently changing their mixture or “batch,” so that any results given with one set of samples might differ from those made with another set, from the same manufacturers. For this reason, in noticing any differences which may occur in experiments made by any of our readers, this fact should be considered as an explaining cause.

I have seen specimens of glass from a factory which changed to a yellowish tinge in a few months, others which changed to a purplish hue, and still others from the same factory, which hardly changed at all. A difference in the mixture, (or batch, as it is termed), makes a difference in the tinge of the specimens from the same factory, both before and after exposure to sunlight. The chief points for photographers are to get glass made from as pure materials as possible, of as light a color as practicable, and free from oxyd of manganese. A glass like either of those named above, as most easily transmitting the actinic rays, might be good for one year or more, and then become very much injured for photographic effects, by the change of color to yellow or pink by sunlight.

Any photographer can make these observations practical, by testing the action of sunlight for six months or a year, on all the specimens offered him for sale. And all manufacturers can make them practical, by making their glass of pure materials, which will not have to be “doctored,” to use the glass-makers’ term for the use of manganese; or by allowing the glass to assume its natural color, even if it be a little blue or green, rather than to run the risk of its subsequent change to yellow or purple by exposure to sunlight.

In the Comptes Rendus for January 14th, 1867, Pelouze says, (and we believe he is the first and only writer who has made this observation):

T. Gaffield on the action of Sunlight on Glass, 1867 11

“Exposure to red heat decolorizes the glasses which have been made yellow by sunlight, or to speak more exactly, they retake the light green shade which they had before exposure. A second exposure to sunlight produces a second coloration, similar to the first, and a red heat makes it disappear again. These phenomena can be reproduced indefinitely. The glass preserves its transparency and does not give place to any striae or bubbles.” He also says:

“I possess specimens of glass rendered violet by sunlight. All present the property of being decolorized by heat. A temperature of 350 degrees is not sufficient. It is necessary to have that employed in the reheating of glass in general, and that is in the vicinity of red heat. The glass decolorized by heat when exposed to sunlight retakes the amethyst color which it acquired the first time, loses it anew when it is heated; and these curious phenomena can be reproduced without cessation.”

In confirmation of this most interesting statement of Pelouze, I have exposed in a glass stainer’s kiln, several specimens of glass which had been changed by the action of sunlight, some to a yellow and some to a purple color. The exposure to an extreme red heat made the glass assume, some a white, some a yellowish white, and some a green color, which were probably the original colors. These specimens were taken from windows where they had been exposed from a few years to more than half a century. Further experiments, which I have already commenced, will show whether we can reproduce the exact original colors by heat, after being changed by exposure to sunlight.

We have in the same kiln exposed some dozen original and unexposed specimens of what are called colorless window glasses of different kinds and shades of color, and found them unchanged in the slightest degree by the action of great heat, while similar specimens have been changed in a few days, weeks or months, by the simple action of the sun’s rays. Fifteen specimens of really colored glasses, (red, green, yellow, &c.,) have been exposed in the same way without any change of color, except a very slight one in a few specimens which were burnt or over-heated.

In Poggendorffs Annalen, Berlin, of May 1st, 1839, is recorded the following interesting fact by A. Splittgerber:

“I would mention a curious fact, in which the sunbeams have, if I may say so, done something in the art of penmanship; not only on the surface, but by inscribing characters through the body of the glass; and, though the matter is based upon causes well known by experience, yet there has probably never before been so striking an instance of their effect known. I am in possession of a plate of glass which was used as a window

T. Gaffield on the action of Sunlight on Glass, 1867 12

pane for more than twenty years, and on which was an inscription in gold letters. This inscription was taken off by grinding the plate on both sides, and polishing it so as to have a new surface. When the glass had been polished, the inscription could again be clearly seen. The parts which had been under the letters remained white, while the remainder of the plate had assumed a violet tint, in consequence of the manganese it contained, a coloring which permeates the whole mass, as the grinding of the surface proved. The uncovered part of the plate, especially when laid upon a white background show the clearly readable characters.”

The same or a similar instance is related by Dr. Herman Vogel in the Photographische Mittheilungen, Berlin, of Sept., 1866.

Desiring to produce a similar result, we made an inscription on a piece of Belgian sheet glass, in part with gold and silver leaf, and in part with black and white paint. The gold and silver leaf were soon washed off, but the black and white painted letters remained, and being removed after an exposure of nearly two years, the words stood out in clear contrast and full proportions, the inscription being in the original color of the glass, and the surrounding portions having been changed by the action of the sunlight to a purple color.

A very interesting experiment can be made, to show the gradually increasing effect of the sunlight on glass, by taking a piece of easily changing glass, say 4x20 inches, painting black a strip 4x2 inches at each end to preserve the original color, and then exposing the strip to sunlight. At the end of one, two, four, six, eight and ten months, one, two and three years respectively, cover with black paint a strip 4x2 inches, and at the end of three years remove all the paint, and you will have, in a single piece of glass, the original color and all the gradations of change effected by exposure from one to thirty-six months. I have made a similar one with Belgian sheet glass exposed nearly two years. It is one of those interesting experiments which speak for themselves, and defy suspicion or contradiction.

I have made an experiment for one year with two kinds of easily changing glass out of doors, and out of the direct rays of the sun, and found that they were both slightly affected, and changed toward a yellowish color. I did not expect any change, but can, perhaps, properly account for it, on the ground that it was the result of the action of diffused sunlight. It is barely possible that the sun may for a few minutes in some days of the year have cast some reflections when I was not present, in the dark corner in which I placed my specimens.

It may be, that the action of the sun’s heat produced the slight effect noticed. If so, it would be an interesting confirmation of Tyndall’s experiments, and of his theory of the correlation

T. Gaffield on the action of Sunlight on Glass, 1867 13

of forces. I do not consider my single experiment as entirely conclusive, and shall make others, which will give us more material for proper theories and conclusions.

The experiments which I have carried on for four years embrace one specimen of optical glass, a few kinds of flint glass and glass ware; sixteen kinds of French, Belgian, German and English plate glass, four kinds of American, English, French and Belgian rough plate, two of American and English crown glass, ten kinds of American, Belgian, French and English white sheet glass, four kinds of American, Belgian and English ordinary sheet glass, fifteen kinds and shades of English colored glass, four of opaque, white enamelled and ground glass, and one piece of the rough metal of American sheet glass; in all, about sixty varieties.

I have watched and recorded in some experiments, the results from day to day, in others from month to month, and season to season. I have now commenced a series, in which I may record results from year to year, for ten years or more. In these, it may be found that specimens of what are called colorless glasses changed to a yellow color by exposure for a year, may by much longer exposure be turned to a yellowish pink and a purple. And some which have been entirely unaffected, may be affected by an exposure for ten or twenty years. Perhaps some of the colored glasses may show signs of a change of hue or shade.

These new experiments include rough and polished plate, crown, cylinder, ground, enamelled and colored glass. I have also begun to expose under several of these kinds of glass, pieces of easily changing glass, which I shall take in from year to year, these under pieces showing the power of these glasses above them to transmit the actinic rays.

The most easily changing glasses are a certain kind of white plate, which changes from a white to a yellowish color, and a certain kind of Belgian sheet, which the manufacturers used to make of a brownish yellow, (they now make it of a bluish or greenish hue, and it is not so easily changed,) which changes to a flesh color, or a pinkish hue. I have accordingly taken these two kinds, for my under glass experiments. Under each of several kinds, to be exposed from one to ten or twenty years, I have placed pieces 4x2 of the white plate. I shall take in one piece at the end of the first, second, third, fourth, sixth and tenth year. These six lights will show the increased action of the transmitted rays from year to year. By comparing the different series with each other, one can perceive the comparative actinic power of each kind of glass, or rather, their comparative power of transmitting actinic rays.

Another interesting under experiment is the following. I have placed under one piece of each kind of glass exposed, a piece

T. Gaffield on the action of Sunlight on Glass, 1867 14

of easily changing glass, which I shall take in at the end of the year.

At the beginning respectively of the second, third, fourth, sixth and tenth years, I shall place under the same piece, another strip of 4x2 inch glass, taking in each piece at the end of the year of its exposure. This series will show the diminishing or increasing power of the glasses under which they have been exposed, to transmit the actinic rays; in other words, will show whether exposure to the sun increases or diminishes the actinic power of the glasses exposed, and renders them better or worse for photographic purposes.

I have taken a piece of 4x18, of easily changing white plate, painted with black paint two inches of each end, to preserve the original color, and exposed the piece. At the end of the year, I shall paint over two inches more of the glass. At the end respectively of two, three, four, six, and ten years, I shall paint over two inches more. At the end of this time, or a longer term, I shall remove all the black paint, and on one light, I shall have all the grades of changed color and shade produced by their different lengths of exposure. I shall lay aside one piece of 4x18 white plate, taken from the same sheet with the exposed light, in order to compare the original with the changed specimen.

I have painted and exposed, just in the same manner as above described, a piece of 4x16 of easily changing Belgian sheet glass.

To show a speaking proof of the painting power of the sunlight, I have taken a piece of 4x6 Belgian sheet and covered it with a thin plate of brass, having the following letters cut out of it: T. G., Jan. 1, 1867. I have taken another piece 4x6 Belgian, and stuck on with gum shellac the two letters T. G. After exposure of one year or more, the removal of the brass plate and letters will show in the former case, rose or purple colored letters on a brownish yellow ground, and in the latter, brownish yellow letters on a rose or purple colored ground.

A similar experiment as the above, I have commenced with two pieces of white plate, and the simple letters T. G., without the date. The result of the experiment in a year or more, will be to show in one case, yellowish letters on a light colored ground, and in the other, light colored letters on a yellowish ground.

I have thus given, as briefly as possible, and yet as fully as desirable, an account of my past and present experiments. New ones are suggested from year to year. I trust that this interesting field for observation and experiment may be worked in other countries. There is ample room for research in the application of chemical knowledge, of qualitative, quantitative, and spectral analysis, and of photogenic tests, to discover the

T. Gaffield on the action of Sunlight on Glass, 1867 15

exact action and causes of the interesting effects of the sun’s rays, which have here been noticed.

Theories. — The interesting phenomena of which I have given an account, have given rise to many theories to account for their cause. Some attribute them to the presence of oxyd of iron, and some to oxyd of manganese. Exactly how the change takes place is a question on which writers differ, although it is my opinion that the precise explanation can only be given after a multiplication of experiments, and a thorough examination of exposed and unexposed specimens of glass by quantitative and qualitative analysis, and perhaps by spectral analysis and observation of photogenic effects, or photogenic tests.{* Since writing the above, by the kindness of Mr. John A. Whipple, the distinguished photographer of Boston, I have been enabled to show, by the comparative darkening of sensitive paper under several exposed and unexposed specimens, the effect of exposure to sunlight for one year. The loss of actinic power, or power to transmit the actinic rays, was in proportion to the change of color. This was in some varieties of glass quite perceptible, but in all will be more so after an exposure of several years.}

We will briefly state the part which the oxyds of iron and manganese play in glass making. In almost all kinds of window glass, and in some poorer qualities of flint glass, and glass ware, materials are used which are not perfectly and chemically pure. The sand, the carbonate or sulphate of soda, and the lime, one or all, contain slight impurities of iron, the protoxyd of which gives glass a green color. To correct this, after the batch is partially melted, a little oxyd of manganese, called glass-maker’s soap, is put into the crucible or glass pot; some of the oxygen of the manganese flies off to the iron, and converts the protoxyd into peroxyd of iron. The peroxyd gives a yellowish color to the glass, and this, being complimentary to the natural pink of the manganese, is neutralized, and the glass is thereby made of a light color. When the sunlight acts upon glass thus made, the nice equilibrium between the oxygen of the iron and the manganese is disturbed, and sometimes the yellow, and sometimes the pink or purple color is produced. I have produced all shades of the purples, running from pale lavender, into the lilac, mulberry, flesh, amethyst, rose, violet, pink, and deep purple. I have produced, or seen specimens, showing all shades of the yellow, from the brownish yellow, up to the brightest gold color; and I have several series of specimens, in which the green has gradually changed into the yellow, and the yellow gradually run into the pink and purple.

Pelouze, in an article in the Comptes Rendus, of Jan. 14th, 1867, sets forth the following theory.

“There is in glass colored yellow in sunlight some protoxyd of iron and sulphate of soda. Light provokes between these matters a reaction from which results peroxyd of iron and sulphid

T. Gaffield on the action of Sunlight on Glass, 1867 16

of sodium. The heat brings about an inverse reaction and reproduces sulphate of soda and protoxyd of iron. From thence comes the return of the glass to its primitive color. Analysis comes to the aid of this theory in demonstrating in glass rendered yellow by sunlight, the presence of an infinitely feeble, but still very sensible proportion of a sulphid, whilst the reactions do not show the slightest trace in the same glass before their exposure.”

“It may be asked, why glasses colored by the reduction of the sulphate, or the direct introduction of a sulphid into their mass, resist an equal or superior heat to that which produces the decoloration of glass become yellow in sunlight. Here is the answer. In glass made yellow at a high temperature by the reduction of sulphate, the iron is found in a state of protoxyd, which cannot react in any degree upon the sulphids. Therefore the glass remains colored. In glass made yellow by sunlight, the iron is peroxyd and in consequence, in a condition to change the sulphid into sulphate, when we expose the glass to the action of heat.”

In reference to glass made violet color by sunlight, he says, “The coloration seems to be due to the fact that the peroxyd of iron gives up a part of its oxygen to the protoxyd of manganese conformably to one of the two following equations.


or Fe2O3+2(MNO)=2(FeO)+MN2O3.

The reheating of glass, that is to say, the action of a temperature of red heat, produces an inverse reaction, which explains the decoloration. In which we have

2(FeO)+Mn2O3=Fe2O3 + 2(MnO).”

Bontemps, in an interesting article in the Comptes Rendus of Feb. 4th, 1867, attributes the changes in color to the presence of oxyd of manganese. He also remarks that he thinks that the violet color occurs only in glass in which the silicates have a base of potash, and the yellow in cases where soda is used. I doubt the correctness of this opinion, as I am quite sure that I have several specimens colored violet, which contain no potash at all in their composition.

I have also specimens which are colored both yellow and violet in the same piece, the yellow portion in one case having been produced by a certain length of exposure, and then having been covered with black paint, while the violet portion was produced by longer exposure.

As before stated, I have noticed changes in what are called colorless glasses, from light colors approaching white, to yellow and pink or purple. I have noticed also a change in a few specimens from a light green to a bluish shade. The former

T. Gaffield on the action of Sunlight on Glass, 1867 17

may be accounted for by the presence of manganese, a very minute proportion of which oxyd will have a sensible decolorizing effect in a crucible of melted glass metal. Pelouze’s theory of the peroxydation of the iron may have some weight in determination of the cause of the yellow or pink color by the action of sunlight. But I know not how to account for the change of one specimen each of plate, crown, and sheet glass in my possession, from a greenish white to a bluish tinge, not mingled with either yellow or purple.

I have been pleased to find the interest in these experiments by photographers, who have long noticed that they could take better pictures under a newly glazed skylight, than under one which had long been exposed to sunlight. The cause of this change is, that the slightest yellow color interferes with the transmission of the actinic rays, and a very deep shade will cut them off in a very great degree. My experiments with glasses under other glasses proved which was best for photographers’ use, information which all can gain by exposure of the specimens of various manufacturers which may be offered them. The most pure glasses of light green, or bluish white color are the best for photographers, and when I say pure glasses, I mean those most free from oxyd of iron or manganese, but especially of the latter, which, I think, is the cause of nearly all the changes which I have observed. Mr. J. W. Osborne, of New York, the gentleman who has done so much to bring the art of photo-lithography to perfection, and into practical use, writes as follows:

“I believe your researches will prove of much practical importance, and I wish the glass manufacturers could be got to give serious attention to the subject. It takes but a very slight tinge of yellow to cut off twenty-five per cent of the actinic rays. I am forced to work under glass, because of the protection from the wind and weather, but in doing so, I sacrifice in any case much of the light. I have to increase the time of exposure proportionately. With the best of glass, the loss from reflection and absorption is considerable. This may be fearfully increased by the color, and if that undergoes a continual change for the worse, the state of things is exceedingly unpleasant. I was warned by a photographic friend in England, to take care of a certain kind of cheap French glass, made for glass rooms, which turned yellow; but I had no notion that the evil was so general as you appear to have found it. The subject is of such vital importance to photographers, that I intend drawing the attention of that friend to what you have done.”

I am indebted to Mr. Osborne for bringing our researches to the notice of his friends at home and abroad, who have contributed interesting articles and valuable information on the subject to the Philadelphia Photographer, (October, 1866,) the London

T. Gaffield on the action of Sunlight on Glass, 1867 18

Photographic News, (August 2,1866,) the London Photographic Journal (Aug. 15, 1866,) and the Photographische Mittheilungen (Sept. 1866).

It may seem singular that so long an experience in the window glass business, had not at an earlier period drawn my attention to the subject of this article. But my experience is not singular. In conversation with many glaziers and glass dealers, I have seldom found one who was aware of the great change of color effected by sunlight. Few have supposed that specimens were to be found in any other windows in our city, than those on Beacon street facing Boston Common and the south, and exposed to the full force of the sun’s rays. But having my attention now particularly directed to the subject, I never pass a window without detecting where any considerable change has been effected. I have found them in all portions of the city, and most generally in those positions which face the east and south. The color is most easily detected, when the glass has been ground or enamelled, or where a white window curtain forms a good contrasting back ground. Many people suppose that the very distinctly marked purple plate glass in Beacon st. was imported of this color, and that it is now no longer made; and hence the reason why the windows looked like checkerboards, when broken lights were replaced by those of the usual light greenish color of plate glass.

On this point I have convinced many who had held contrary views, by showing pieces from which the putty on the edges has been removed, and displaying beneath the original color. When the putty covered the glass, the sun’s rays could not reach it, and the color was unchanged. Such specimens are quite interesting. I have had many conversations with old glaziers, dealers and consumers of glass, and also with those who stain and enamel it. I have thus gathered specimens of various kinds and colors, and of differing lengths of exposure from old windows. The oldest specimen was one of crown glass set in a church in Lexington, Massachusetts, in 1794, from which the windows were removed in 1846, and since used as covers for hot-beds. The original color, found by removing the putty from the edges, was a light green, and the present, after seventy-three years exposure, is a purple color. I have never yet met any one who has seen glass in original imported packages, of the purple color made by exposure to sunlight, and until I do, I shall adhere to my opinion, that all purple or rose colored glass which is seen in our city windows, was made so by said exposure. I have very fortunately found an octogenarian, who has furnished me with some glass which was imported, he thinks from some part of Germany, which is of a light green or yellowish green color. The glass was imported more than thirty years ago. Much of

T. Gaffield on the action of Sunlight on Glass, 1867 19

that which has been set in his windows facing the south is now purple. An experiment with the original glass, commenced this summer, showed a perceptible change in color in one day, and in two weeks, the change toward purple was so marked that I have no doubt that this color will be distinctly visible in less than a year. If two years were occupied in the erection of the Beacon street houses, or any others in which the plate glass purpled by sunlight is found (as I am informed was the fact in some cases), then the result of the single experiment named above is a sufficient reason for the mistaken belief of many occupants and owners to-day, that the glass was purple when imported. It was probably changed from yellowish green to purple before the houses were finished, and the owners had taken possession.

The action of sunlight which I have spoken of in this article, must not be confounded with that called “rust” or “stain,” which is occasioned by exposure to the weather, and manifests itself in two ways; first, by a disintegration and roughening of the surface, sometimes producing all the effects of ground glass; and secondly, by an apparent formation of an infinitesimal coating of oxyd on the surface, on which the play of light gives all the colors of the rainbow, as with the action of light on the infinitesimal grooves of the surface of mother-of-pearl. This is simply surface action, whereas the action of sunlight permeates the whole body of the glass, wherever the rays directly strike it.

The writer of this article, (compiled in the midst of the busy duties of mercantile and official life,) makes no pretension to accurate scientific knowledge, but gives the results of his observations and methodical experiments on a well known phenomenon, in the hopes that they may add some mite to the sum of human knowledge, and may stimulate and aid those who are better versed in scientific studies, to ascertain the causes and exact operations of this interesting power of the sun’s rays to paint the products of art, as they do so beautifully and wonderfully the works of nature on the mountain, in the forest and field.

Boston, Mass., July, 1867.