The great signal color shift at the turn of the century
The Chicago and North Western Railway used semaphore arms with spectacles of a distinctive shape on three-aspect upper quadrant signals. The reason for this distinction goes back to the days when green was being adopted for Proceed instead of white, and three-aspect upper quadrant signals were new. In later days, the spectacle had red, yellow and green lenses, but this was not the case originally. In fact, originally there were no colored lenses in the spectacle at all! The key to all this was a new "distant signal lamp" that was used for more than distant signals, and the choice of the C&NW to use the simultaneous display of red and green for Caution. The reasons for these developments will be explained in this article.
By 1890 the electric incandescent lamp was appearing in general use, especially in and near cities where alternating-current electricity was available, even outdoors in street lamps and signs. Railway signal lamps were oil lamps, visible enough in the night, but no competition for bright electric lights. Oil lamps were retained because of their reliability and the lack of electricity in the countryside. The signal colors had long been standardized at white (clear lens) for Safety, green for Caution and red for Danger, using the words then in vogue for the indications. White was the most visible, red was less visible but distinctive, and green less visible than red, and a rather variable and poor color. The possibility of confusing a dim signal white light with an extraneous white light had been appreciated in Britain for some years, and green was required for Safety after 1892 by the Board of Trade. The use of green for Caution simply fell out of use with the elimination of this concept in operating rules, since the meaning of Caution could not be usefully defined. If it meant run slow, or run at restricted speed, then why not simply say that? However, Caution still meant more in the United States, though definitions were still poor.
The lights providing the night aspects of semaphore signals were mainly for distant sighting of the signals. When the train was close to the signal, the headlight illuminated the semaphore arm well enough that its position could be clearly seen. The rules provided that if the light were out, then the day aspects would govern. It was observed by E. C. Carter that enginemen looked for clear signals when moving at speed, and if they did not observe them, immediately became cautious, rather than moving confidently until stopped by a red signal. Therefore, positive proceed signals are essential. This is an important signalling principle that should be clearly recognized; I think it deserves to be called the Carter Principle. This principle emphasizes the importance of a distinctive and obvious clear signal, which the white color assisted. The adoption of green would make the signal dimmer, but the color contrast now assists.
The Chicago and North Western was perhaps the first major American railway to adopt green for Safety. This change appears in the 1893 Rule Book, green replacing white for Safety in distant, home and dwarf semaphore signals, antedating the change on the New York, New Haven and Hartford in 1895, often mentioned as the first. The Pennsylvania and the New York Central did not change until after 1916. The new green was then 50% brighter than the old. In Hall enclosed disc signals, green replaced white in block (home) signals and in clear distant signals. However, a green disc with a white X was retained for the Caution aspect of the distant signal. Distant signal arms were still painted green with a red chevron. Green was an appallingly bad choice of color for a semaphore signal arm, but green arms were standard in the United States for distant signals. Earlier, the C&NW had used blue instead of green as the color of Caution, mainly in flags and signs, but this had been given up by 1893.
The semaphore and Hall distant signals displayed a green light and a red light simultaneously, side by side, when indicating Caution. This was the new feature in C&NW distant signals, and these lights were provided by a special signal lamp. The display of double lights was not new, even in the United States. The Pennsylvania had used double red lights to distinguish the signal for the high-speed route in the days of "signalling from the left" at junctions where a separate arm was used for each route, stacked one above the other, and the main route could be somewhere in the middle. It was later to use double red lights to distinguish manual block signals from interlocking or automatic signals. Two lights of different colors displayed simultaneously had been used in France, and were later to be widely used in Belgium. In all cases, one light was direct and the other reflected, so that it was impossible for just one light to fail. The Pennsylvania's block limit signs displaying red and yellow lights also used the same principle.
At the time that green was adopted for safety, there was no agreement on the color to be used for Caution, if any. One suggestion was violet; another was two white lights side by side. In 1899, the Standard Code retained only red for Danger, leaving the colors to represent Caution and Safety up to the individual companies. At this time, the terms Safety, Caution and Danger were replaced by Proceed, Proceed with caution and Stop. As late as 1928, the colors for Proceed and Proceed with Caution were still not agreed upon, although markers were illustrated with yellow and red instead of green and red. By that time, however, the use of green for Proceed and yellow for Proceed with Caution were nearly universal. The C&NW's suggestion of red and green for Caution was actually a good one, but was not widely adopted. Prior to the introduction of semaphore signals in the 1870's, red and white were nearly the only signal colors used in the United States. If a color for Caution was required, blue was as likely a color as any. Finally, Caution disappeared and was replaced by Approach, or proceed prepared to stop at the next signal, of which yellow was the color.
At this date, one may wonder why yellow was not just adopted at once. The reason lies in the spectral distribution of the light of a flame (color temperature about 2055K). There is a considerable amount of red, but very little blue or green. To make a colored light, one removes all the light that is not of the color desired, so the light is necessarily weaker. If a good cobalt blue glass or a rich green glass is placed in front of a flame, the transmitted light is quite weak. Blue is all but impossible, perhaps 1/7 the brightness of the naked flame, while green gives only about 1/5 the brightness. An acceptable signal green is, therefore, allowed to contain a considerable amount of yellow, its spectral neighbor, and appears distinctly yellow-green. It proved to be difficult to distinguish this color from the yellows then available, especially under difficult atmospheric conditions, which made any yellow even more greenish. In fact, a signal yellow was eventually created, the Nels or Baird yellow from Corning, that was actually amber, or yellow-orange, different enough from green, while not easily confused with red. This glass was first used by the NYNH&H in 1899. Therefore, even with flame illumination it was eventually possible to use red, amber and green successfully as signal colors. Purple was used when the light was specifically not meant to be seen from a distance, as in dwarf signals. Purple was produced by a bluish-red glass of small transmissibility.
Electric illumination is a very different story. The color temperature (2848K) is high enough that a reasonable amount even of blue is present. Now, not only would the blue be acceptable, but the signal green could tend to bluish, contrasting better with amber, which could then be made a purer yellow. It must be appreciated that the same glass colors cannot be used with both an incandescent filament and a flame if good results are desired. Not only was a good signal yellow developed for electric lighting, but a pale yellow "fog penetrating" light was also developed by Dr. William Churchill at Corning Glass around 1915. Actually, no color penetrates fog much better than any other, but this was an excellent bright color that was used in the Pennsylvania's position lights, producing good visibility with small consumption of electricity. [These lights are certainly not "amber," as is sometimes stated.] It is not commonly recognized that these signals used less electricity than the popular color-light signals, where most of the light is filtered out by the colored glass. Another useful color was "lunar white," a milky bluish light that was also easy to see and recognize. Possibly unfortunately, lunar white was never used for Proceed in place of the old clear white. It was used, however, for Proceed at restricted speed, or for Permissive Block, where it is distinctive. With incandescent light, signal colors of lunar white, blue, green, yellow, red and purple are all available.
In the absence of a suitable yellow, the C&NW chose to display a green and a red light side by side. At a distance, these lights combine to create the sensation of yellow, until they can be separately distinguished. The lights were produced by a modified signal lamp, designed by E. C. Carter, the C&NW signal engineer, in 1890. This lamp had two lenses at 90°, one red and one white. The red light was produced directly by the lamp, while the green light was produced by reflection in a 6-5/8" x 8-3/8" x 5/16" mirror at 45° in front of the white lens. A green roundel 6-7/8" in diameter beside the standard 5-3/8" red lens colored the light. The separation of the centres of the lights was 9-1/4". The angular sensitivity of the human eye is about 2' of arc, so the lights could be resolved at a distance of about 1325' or 440 yd, and seen as separate red and green lights. I do not know what an engineman's first impression of the light was, whether a red, or yellow. It was, however, certainly not green, and would receive immediate attention.
The burner (after 9 March 1912) was an Adlake #100 using a 5/8" flat wick, called a "long time" (i.e., more than one day) burner. The wick was replaced monthly before it became gummy and clogged. The capacity of the font was about 34 cu. in. (about 19 oz.), which would give a burning time of about 4 days at a rate of 5 g per hour, and produce about 1.4 cp. A glass chimney about 2" high and 1" in diameter was used around the flame. The wick was at 45° so that each lens saw the same size of flame. No reflectors were used, though they would have improved the lamp considerably. A 1" glazed aperture was provided for looking at the flame when adjusting it. There was no provision for a backlight. The diameter of the lamp body was 6-1/2", and the Fresnel lenses had back focal lengths of 3-1/2". The lamp was provided with a carrying bail at a point of balance. The lamp was tested with a 40 psi compressed air jet held 2" from the lamp and blown at all parts.
The application of the lamp to distant signals is shown in the figure. At the left is the Hall disc signal. The case was white on the front and yellow on the back (home signals were black on the front and yellow on the back), and an X of 1-1/2" white lines was painted on the green disc for better visibility. An extra hole was provided to accommodate the second light, which was used from the first. When at Clear, a blnder obscured the red light, leaving the green light displayed. There was a disc aperture on the back as well as the front, so light could be seen through the signal when the disc was withdrawn, a valuable feature. A small part of the disc was still visible then, to show that it ws still there.
On the lower-quadrant distant, the lens that previously had been clear was now replaced by a blinder, while the green lens was simply removed so that the red light from the lamp could show through. This simple modification was easily carried out. The new upper-quadrant distant used a spectacle with one blinder and two empty spaces. In the Clear aspect, as shown, the arm obscured the red light, leaving the green to show. In the Approach aspect, with the arm at 45°, the red light was shown in addition through an aperture in the spectacle. The aspect with the arm horizontal was not used. Note that the distant signal lamp was used as a marker at the location where a lower arm would be. In upper-quadrant territory, there were always two arms, or an arm and a marker, and this was maintained for distants. The marker light was white, though after 1953 it was red.
We observe that the lamp itself produced red and green lights, and could not produce just a green light with no red. It was not used only in distant signals, but wherever an Approach aspect was required. When used in an automatic block signal, the spectacle plate was specially shaped, as shown in the figure. It contained no colored glass, only an opaque screen and two open apertures. When the arm was horizontal, the opaque screen on the top roundel blocked the green light, so that only the red showed. At 45°, an open aperture allowed the green to be seen as well. It was not possible to display just the green light in this position, since it was derived from the red. At 90°, the green light showed through an aperture, while the red light was obscured. The curious shape of the semaphore arm outlasted the signal lamp. It was still used with red, yellow and green lenses many years later, as shown in the article title. The semaphore arm with red/green instead of yellow was retained in the 1919 and 1929 Rule Books, but had disappeared by 1953. Arms of the same shape, but with red, yellow and green lenses, were used instead. The red/green signal lamp was a practical and long-enduring solution to the night aspect problem, though never adopted by other companies.
On the C&NW, semaphore arms pointed to the right as a train faced them. The signals were to the right of the track on single track, as is the customary practice in the United States. On double track, C&NW trains ran on the left-hand track (the reason for this was some peculiarity of the stations or traffic in the suburban district around Chicago). The signal masts were placed to the left of the track, though the arms still pointed to the right. In 1893, the C&NW was being very progressive in signalling. The Westinghouse electropneumatic system with track circuits, and lower-quadrant signals, was used between the Wells Street Depot (the site of the Merchandise Mart) and West 40th St, and Wells Street and Deering. Hall disc signals with track instruments were first installed between 40th St. and Turner (West Chicago), 25 mi., Clybourne Junction and Barrington, 29 mi., and Deering and Waukegan, 32 mi., 86 mi. in all. Later in 1893, Hall signals were placed on 11 mi. near Milwaukee. They also soon replaced clockwork signals on 4 blocks near Dixon. By the end of 1893, the C&NW had some 200 miles of wire-connected Hall signals. They still appeared in the 1929 Rule Book, but by that time had long been converted to track circuits or replaced by semaphores.
A full three-arm upper-quadrant signal is shown in the figure. The lights for the two upper arms were in a vertical line for interlocking signals, 7 ft. apart. The Stop aspect then indicated just that, Stop (and stay). In automatic block signals, the lights for the two upper arms were staggered, as shown here. This aspect indicated Stop and Proceed. Putting the light on the left of the mast required a different shape for the spectacle of the lower arm, but the application of the distant signal lamp was equally satisfactory. Where the second arm was not needed, it was replaced by a white marker light, on one side or the other of the mast as required. Signals like these, but with lights in a vertical line on the right of the mast, were very commonly seen as interlocking signals, for example on the lines near the Chicago terminal. Sometimes, the second arms were replaced by marker lights when there were no medium speed routes to govern. These signals survived at least into the 1960's.
The lowest arm, with a short blade, indicated Proceed at Restricted Speed when cleared to 45° and showing green/red. This aspect was called Restricting. At 90° it indicated Proceed at Slow Speed, and the aspect was Clear-Slow. This arm was low on the mast, and could appear by itself as a dwarf signal, giving the same indications. When the arm was horizontal, in some cases both lights were masked, so that no light was shown. Two-position lower quadrant dwarf signals showed Stop and Restricting. The Clear-Slow aspect was probably not very useful, but the Restricting aspect was very often used, applying to all routes at an interlocking except the main ones. Dwarf signals and the lowest arms later displayed only red and lunar white, and the short arms were moved up to a normal 7 ft. spacing from the second main arm or marker light.
Three-position semaphore signals were invented by Loree and Patenall in 1903. It was a few years before the final arrangment of right-hand, upper quadrant operation was established. The first of these were installed on the Pennsylvania between W. Philadelphia and Elwyn in 1906. It is clear that the adoption of these signals was rapid, since they appeared in the C&NW Rule Book in 1919. Their fame even spread to Belgium, where they were used after 1919 (but not in the Loree-Patenall form until 1931). They were used to a very limited extent in Britain, as well.
In the 1919 Rule Book, red/green over green indicated Approach Next Signal at Restricted Speed, while red over green indicated Proceed at Restricted Speed. The 45° aspect of the short lowest arm indicated Porceed at Slow Speed Prepared to Stop. The distinction was that in the first case, it applied to a main route, while in the second, to a subsidiary route. This shows that at that time, crossovers and junctions had to be negotiated at slow speed (15 mph). By 1929, medium-speed turnouts had been installed, and now the second arm indicated medium speed (30 mph) instead of slow speed. Red/green over green now was Approach Medium, and red over green was Clear-Medium, as was becoming standard. The lowest arm could now be reserved for Restricting. In every case, a yellow or lunar white light simply replaced a red/green signal lamp. It seems that yellow was used for new work and replacements, but the red/green lamp was not immediately banished. Color-light signals first appeared in the 1929 Rule Book, while Hall disc signals were still in use. These first color light signals were also distinctive in that the three lenses were in a horizontal row instead of vertical, another C&NW peculiarity. These GRS Type E signals were first used in 1924. There was a wonderful variety of signals at this time, a very interesting state but not a desirable one. In 1926-27, the 511 miles from Chicago to Council Bluffs was equipped with 2-indication continuous cab signals with automatic train stop, and the lineside signals were removed.
Hall Automatic Block Signals on the Chicago & North Western, Railway Gazette, 13 January 1893, p.22-23. Also 24 November 1893, pp. 845-846.
C&NW Ry., Signal Drawing 1692, 22 August 1912. The lamp.
C&NW Ry., Signal Drawing 1647, 4 March 1915. The automatic block signal.
C&NW Ry., Rules and Regulations for the Government of Employes of the Operating Department, taking effect 1 August 1893, pp. 21-31.
USRA, C&NW RR, Rules and Regulations for the Government of Employes of the Operating Department, taking effect 1 June 1919.
C&NW Ry., Rules for the Government of the Operating Department, taking effect 1 December 1929.
Composed by J. B. Calvert
Created 30 June 2004
Last revised 2 July 2004