Early Railway Signals

Contents

1. Introduction
2. Vane Signals
3. Semaphore Signals
4. Electrical Signals
5. Light Signals
6. Motion Signals
7. References

Railway signals are a means of communication beyond the range of the voice. They may stimulate the eye, as optical signals, or the ear, as acoustic signals. The nature or appearance of a signal is its aspect, while its meaning is its indication. The most fundamental signals are the hand, or mobile, signals. They may be given and received by men on the ground, or on a train. Indeed, motions of the hand can be used. By day, they are emphasized by a flag (or an object held in the hand), and by night, with lights. A red flag or light is normally used for a stop signal, while a white flag or a clear light is used for general signals. When a "white" light is mentioned, what is meant is the light of the flame, mantle or filament that is the source of the light. Red and white were customarily used on the common roads at the birth of railways, and were generally adopted. Such hand signals are displayed by men on the ground or on trains for reception by the driver to indicate how he should handle his train. Optical signals must be noticed to be received. Motion or blinking can aid their recognition. Hand signals often use motion in their aspects. In America, moving the hand or lamp up and down vertically is a signal to proceed; moving them across the track is a signal to stop; moving them in a circle is a signal to back. Holding the hand or lamp out steadily is a signal to reduce speed and take caution. In Britain, holding an arm out horizontal signals all right; proceed. Holding an arm vertical above the head means caution, slow. Both arms held above the head means stop.

The driver himself has his whistle available, with which he makes requests and warnings, and acknowledges hand signals received by him. In America, he also has a bell, a reminder that early American trains often moved on tracks in city streets. The bell is purely a warning to the public; it has no role in train operation. The whistle is, of course, an acoustical signal, which can actively attract attention. The attention of the driver can be attracted by means of the detonator or torpedo, invented by E. A. Cowper in 1841 and relied upon since then for notice in emergencies. They are fastened to the rail head with flexible strips (originally lead), and are often required to be used in pairs, in case one of them fails. However, they are very reliable and have a good shelf life.

On the Liverpool and Manchester, a white lamp moved up and down was a signal to stop, and was a signal of Caution, if moved from side to side. A succession of short sounds of the whistle was the signal to apply brakes. In America, one sound of the whistle was recommended in 1867 for "Up brakes" (release brakes), two for "Down brakes" (apply brakes) and three for backing. Later, it was realized that it was much safer for one sound to mean apply brakes, and two to release brakes. Such contradictions in the meanings of hand and whistle signals were very common in early days.

The pyrotechnic flare is a very effective warning. They are available in red, yellow and green and various times of burning, from 5 to 15 minutes. The 10-minute red flare is commonly used by American railways, where it is called a fusee. This looks like a French word, and it is, but means rocket, or the end of an axle. The fusee is an allumette-tison in France. They were used in Britain in the mid-1800's, but later went out of use, probably because of the fire hazard. To light a fusee, the cap is taken off, and the match head on the end of the pyrotechnic mixture is scratched against it. In a short time, it is burning well and can be dropped off a moving train. If it is thrown upwards, it will land on the spike and stick in the ballast or a tie. Using fusees is an effective way to maintain the time interval between trains.

Before radio, the rear end of a freight train communicated with the head end with hand signals, for which the engine crew were instructed to look back for at frequent intervals. On passenger trains, communicating signals were available. They began with the bell cord that passed through the train (a rather unsatisfactory device), but later were pneumatically operated, sounding a whistle in the engine cab when a communicating valve was opened and let out some air.

All the signals so far mentioned are well called "mobile" since they can be displayed at any point. Some signals are, however, displayed from fixed locations and are called fixed signals. They include signs, which have only one aspect and one indication. A good example are speed limit signs. They consist of a warning sign placed a braking distance from the restriction, the sign indicating the beginning of the restriction, and finally the sign showing the end of the restriction. Track works may require temporary speed restrictions at any point, but these are not considered as mobile signals, but as temporary fixed signals.

More interesting are signals displaying two or more aspects. The first of these generally used was probably the one shown at the right. When trains are to stop, a red flag by day and a red lamp by night is displayed on a post reserved for that purpose. When trains can proceed, the flag and lamp are taken down. The advantage of a fixed signal is that its location is known and permanent, so it can be regularly looked out for. The earliest documented railway signal was an optical telegraph at Middleton Colliery. It was a black disc informing the operator of a stationary engine when waggons had been attached to the cable that would draw them to the summit of an incline. Let's review the places on a railway that may require such fixed signals.

First of all, we have stations. Here, trains may be standing, even on the main track, by which we mean the track used for the through movement of trains. There may be a number of routes available for entering or leaving trains. Trains may be in the process of being made up or taken apart, obstructing tracks. Therefore, a fixed signal may be needed to stop a train until it can safely be admitted to the station. Another may be required to inform the train when it can proceed on its journey. Arrival and departure signals are frequent uses of fixed signals, and are basic in French and German signalling. Junctions require that trains approach the junction with care from all directions. Trains must not simultaneously converge on one track, and must be permitted to proceed in the facing direction only when the route is properly set and the train is moving at a safe speed. Trains must not proceed onto drawbridges until the bridge is in place. Crossings of railways at grade require that only one train be allowed to cross at a time. Stations, junctions, drawbridges and crossings are places where fixed signals are usually required.

Consider the entrance to a station of considerable importance. The turnouts directing a train to its proper arrival track may be assigned to a switchtender for operation. There is little hazard when there is only a single switchtender, who walks from turnout to turnout and is always very aware of what he is doing. When a train arrives, it is stops outside the station while the switchtender identifies it and lines the points properly. When this has been done, he signals the train by hand to proceed. There is no difficulty here when there is only one train waiting for signals. When there is more than one, the proper train to receive the signal must be known. This can be made clear with fixed signals, since each train can have its own signal.

At junctions where several routes are available, it was often the custom for drivers to request the route with the whistle. Either special whistle signals were established for the junction, or some general rule such as as many whistle sounds as the number of the desired route, counting from the left or right.

Beginning in the 1850's, there was a constant stream of patent signals, few of which were even tried, and none of which were regularly used. These were usually mechanical automatic signals operated by treadles, and the idea was to have a train automatically protected by its own presence. After 1870, electric inventions predominated, for automatic or remotely operated signals. The inventors were usually ignorant of railway operating practices, and their inventions satisfied no real needs. These patent signals were often curious, and occasionally ingenious, but will not be given much space here.

Let's now consider the most common form of early railway fixed signals, consisting of a flat shape rotating about an axis, so that either the shape is presented full to the train, or its edge. What to call the flat shape is not clear. The usual French word is disque, or disc in English, and indeed most of the shapes are circular. Since many shapes are not circular, one may use "target" instead, which also implies a circular shape, but not as strongly. A good word is "vane", which comes from a word meaning a small flag. Since a weathervane can point in any direction, this is also appropriate for the signal. In the United States, a vane rotated a half turn, a target a quarter turn. I'll use vane or target here, unless the object is circular, when it will be a disc. Neither the shape nor the colour of a vane can be identified at a great distance under outdoor conditions, but at short distances vanes are quite practical. We shall see that shape should identify the vane, and its colour should be secondary, at best aiding visibility. A lamp must be used to display the night aspects. It is better to use the light of the lamp, rather than to illuminate the vane.

The flag and lamp post was replaced by a permanent target or vane on a vertical axis. The axis was rotated by 90° to display the target, or else to turn it parallel to the track. For the night aspects, a lamp was placed on top of the axis or by its side, rotating with target. Some examples are shown in the figure. The signal invented by Woods was first used on the Livrepool and Manchester in 1834, but without the vane, as a support for the lamp. The vane was added around 1838, and a half-vane was sometimes used as the Safety aspect, so that a positive signal would be displayed. This signal became popular in Scotland, and lasted there for many years. It had a cast-iron post, like a street lamp. The Newcastle and Carlisle signal was a 4-ft disc applying to both lines. The triangular target was used on the Stockton and Darlington, and the double red spectacle on several roads. All these signals of the 1840's were quite similar, and are shown as painted red. Any bright colour would do as well.

Some supposed American vane signals are shown at the right. These are only impressions of what they could have been like, since no actual examples are known to me, except in words. The signal on the right displays the two sides of a vane, painted red and white, and turning 180°. In the shade, it would often have been difficult to distinguish the colours, and the shape was the same for both sides. On the left, there is a three-aspect signal, showing Caution in addition to Safety and Danger. Safety and Danger were the original names of the aspects which we would now call Proceed and Stop. Note that blue is the color of Caution in America. In a Memorial to Congress in 1867, the Franklin Institute proposed legislation on uniformity in railway signals, in which blue was the recommended Caution color. Until the introduction of British signals in the 1870's, this was quite typical. In the shade, red and blue squares both look like black, and would have been difficult to tell apart.

The Caledonian Railway signal shown at the left overcame the colour problem by making the shapes different for Danger and Caution. Danger had a tab on the track side, while Caution had the tab pointing away from the track. The aspect could be identified without seeing the colour. Safety was displayed with the target parallel to the track.

Brunel, on the Great Western Railway, designed excellent vane signals that displayed different shapes, so that colour played a subsidiary role. The main signals were the high disc-and-crossbar, used as station signals and at tunnels. The crossbar was 8 ft long and 1 ft wide, while the disc was 4 ft in diameter. Both the disc and the crossbar were painted bright red for visibility. Actually, against the sky background that was usually intended, both appeared little different from black except when the sun was directly on them. The "fantail" signal was used to maintain the time interval between trains. It was much lower, and usually used in conjunction with a disc-and-crossbar. The red arrow, pointing to the track, was displayed for 5 minutes, then the green arrow, pointing away for five minutes more. Finally, the vane was turned parallel to the track. Before the fantail was introduced, a frame displaying red and green flags was used, but did not prove satisfactory. The fantails actually outlasted the disc-and-crossbar, since they were used on time-interval lines to a late date. The disc-and-crossbar was characteristic of the broad gauge, as the semaphore was of the narrow (standard) gauge. The Bourne engraving of Box Tunnel shows a disc-and-crossbar and a fantail. By night, coloured lamps were used instead of the fantail, hung on a separate post.

On the London and South Western, Albinus Martin designed a unique disc signal that was rotated by a cord. Only one signal was used for both tracks, and could show that one track was blocked, both were blocked, or both were clear, the last by turning the disc parallel to the track. This signal was invented in 1840, and was said to still be in use as late as the 1870's. It was used chiefly to protect shunting movements. It was made with an iron frame covered by fabric or sheet metal. The colour shown is only presumed.

The signal at the right is a double spectacle, but instead of rotating about a vertical axis, the axis is horizontal. It can be operated remotely by means of a wire. The spectacle is automatically displayed when the operating wire is released, through the action of the counterweight. In 1846 it was in use as a distant signal, that is, one a distance from the protected point. A train approaching it had to stop at the signal, or at least be under control, and then approach the point at restricted speed, looking out for a train ahead. If it protected a junction, for example, there might already be a preceding train waiting at the junction. This could easily happen under the time-interval system. Of course, vanes with vertical axes could also be arranged to be wire-operated and restored by a counterweight.

The low signal shown at the left was called a "ground indicator" or "shunting signal". Sometimes it was operated directly by the points of a turnout, but usually by a lever in the signalbox. It could permit a movement to back through a crossover or to enter a siding. Before these signals were used, these movements were made on hand signals from the signalman. In any case, it was not intended for through train movements, and was purposely designed to be visible only at short distances. Such low signals are usually called dwarf signals, or colloqually, "pot signals" from their squat form, resembling that kitchen utensil. Sometimes, the even more obscure colours purple and blue were used instead of red and green. This signal displays a small blue "back light" when it is at Stop, so the signalman can verify its position. Green, here, means Caution, not Proceed. The white cross made the green target more visible. Green is a very bad colour for painting any signal; it is much less visible as a surface than as a light. This signal was the model for American switch targets, which also turned by 90°. Later, small semaphores were used as dwarf signals, and still later discs turning about a horizontal axis parallel to the track with a stripe across them like a semaphore arm. An American dwarf semaphore is shown at the right below. Its aspects were usually Stop and Restricting, if red and yellow lights were used (originally, of course, red and green). After the change to green for clear, a green light would display the aspect Slow-Clear. Before the change, red and white would probably be used, but I have not seen an example of this. Dwarf signals may also be used when there is insufficient space for a high signal; this is more appropriate with light signals than with vane or semaphore signals.

We have seen discs, targets or vanes that can rotate about any one of three mutually perpendicular axes. Vertical axes are the most common, with a 90° angle of rotation. 180° rotation is also found. German distant signals, which are yellow discs, also rotate about a horizontal axis. French vane signals, which survived until the end of mechanical signalling, always rotated by 90° about a vertical axis. See the articles on French and German signalling for examples of modern vane signals.

Ball signals are not vane signals, but may appropriately be mentioned here. They consist of one to several balls, perhaps of different colours and shapes, hoisted on yardarms. The balls, cylinders or cones may be of wicker or other light material. By night, lamps are hooked below the balls or hoisted by themselves. These are very nautical in context, and in fact were widely used as tide signals, to telegraph the depth of water over the bar to ships in the offing. They appeared with early railways in the United States, England and Germany, but were used for any time only in the United States, and there mainly in New England.

Brunel apparently erected several ball signals as station signals, at least at Southall and Reading where they are mentioned in rules. These were single balls raised to show Safety, and lowered to show Danger. The exact appearance of these signals is not known; the illustration in Rapier has no good basis, but is probably illustrative. These were in use in 1840, but were soon replaced by the disc-and-crossbar. MacDermot suggests they may actually have been discs, but there is even less evidence for this. Brunel may well have seen them on his visit to the United States, and imported them as station signals. At any rate, they were quite ephemeral.

In the United States, ball signals were used on the short New Castle and Frenchtown around 1832. Here, they were station signals in telescopic view of each other, and were raised and lowered to telegraph the movement of trains optically. This was a unique use, and it was not done elsewhere. These signals did not last past 1840. Balls were used on the Erie and even more extensively in New England. They were station and junction signals, as well as crossing and drawbridge signals. Most had only one or two yardarm ends, each which could raise one or several balls. In general, a raised ball was the signal to proceed. The term "highball" entered the jargon with this meaning. On the Erie, some balls were illuminated from inside. Instead of being lowered, a ball could be raised inside a cylindrical housing at the top, and then was considered "not displayed".

Although Galileo demonstrated a practical telescope in 1610, it was nearly 200 years before optical telegraphs were introduced by Claude Chappe (1763-1805) and his four brothers in revolutionary France, where Napoleon made it an important military tool. Chappe made the first test of the télégraphe in 1793, and the first line, Paris-Lille, began operation in 1794. The T-telegraph had a central arm, the regulator, 4.62 m x 0.35 m taking 4 positions, and two shorter arms at each end, the indicators, 2 m x 0.33 m, each taking 6 positions. This apparatus was the result of investigating the most visible way to display the signals. It reflects the unusual sensitivity of the eye to the difference in directions of the two segments of a broken line. Telegraph stations were from 3 to 11 km apart, about 8-10 km on average. The telegraph was used by day only, and was interrupted by bad weather. Night use was studied, using oxyhydrogen limelight, but was never adopted.

Although the Chappe's investigations of the visibility of optical telegraphs are applicable to railway signalling, the railway semaphore was derived rather from Depillon's coast telegraph, called the sémaphore. This was a vertical post with several, usually three, arms jointed to it. Introduced in 1803, it eventually covered the French coast, with 97 stations from Flushing to Bayonne on the Atlantic, 32 stations on the Mediterranean, and 6 stations on the Algerian coast. These semaphores did not communicate with each other as a telegraphic line, but were used for communication with ships in the offing and costal surveillance. They were connected with the Chappe telegraph by messengers, and later with the electromagnetic telegraph. Some stations survived as late as the 20th century. These stations were the favourite target of Admiral Cochrane in the Napoleonic Wars, and came to the notice of the Royal Navy. Admiral Popham's bridge telegraph was suggested by them, as well as General Pasley's field telegraph, designed about 1815 and built at the Royal Engineer establishment at Chatham.

In 1840, General Pasley became an Inspecting Officer for railways. When he observed the operation of the busy junctions at Corbett's Lane and other places on the London and Greenwich, Croydon and South Eastern railways, he suggested the application of his telegraph for the control of trains to Charles Hutton Gregory of the Croydon (later London, Brighton and South Coast Railway). An example was installed at New Cross in 1841, and proved very satisfactory. These first signals were Pasley field telegraphs as designed at Chatham, with two arms and a slotted post, that Gregory quickly adapted to railway use. This telegraph was designed by Robert Howe, Clerk of Works at the Royal Engineer's Establishment, Chatham, in the 1820's. His arms were tapered to a smaller width at their ends, but the Admiralty telegraph had parallel arms.

The arms of Chappe's T-telegraph were of slatted construction. This feature, like the holes in vane signals, chiefly reduced the weight of the structure and made it easier to operate. It is often remarked that this reduced wind pressure, which it did to a minor degree. Indeed, the designers may have thought this, but it is probably mainly an invention of writers thinking of some reason for it. Slatted semaphore arms were widely copied, especially in early years. Solid arms of wood or pressed sheet metal replaced slatted arms, notably in Britain. Austria retained slatted arms until after World War II.

The semaphore was a day signal (as, of course were vane signals as well). By night, the aspects were presented by coloured lights of lamps. The colours were produced by moving coloured glasses in front of a light. We'll call the moving frame a spectacle, though it more commonly has three, or only one, apertures or roundels. Fortunately, coloured lights are easily perceived and recognized in darkness, so that even an ordinary oil lamp gives sufficent light. The perception of colour is not a simple thing, and colours are susceptible to modification by atmospheric conditions and the state of the eye. This was recognized quite early, together with the effects of anomalous colour vision ("colour blindness"). The intensities available with oil lamps are not sufficient for good daytime visibility (except in tunnels). Electric lamps, however, can be sufficiently bright to be easily visible in daylight. Eventually, the night aspects of semaphores became the day and night aspects of colour-light signals, the most popular form of railway signals in the present day. This change happened around 1920 in the United States, but semaphores were still common in many countries until the 1950's.

The day aspects of a semaphore are best displayed against a sky background. In this case, a semaphore can be sighted and read at distances of a mile or more. When not against a sky background, the range of the signal is greatly reduced. Sometimes, a painted white background is used for contrast with the arm. A signal should be distinctly visible and recognizable at a distance of, say, 300 metres. As an example of good sighting, on the London and North Western line from Shrewsbury to Crewe, over flat and open country, the signals of one block post could usually be seen from the preceding block post against the sky, and were continually in view. The high arms were complemented by low arms and the spectacles for night aspects that were more easily visible when close to the signal, or in a fog, called "co-acting" arms.

The spectacles may be separate from the semaphore arms and operated by a mechanical linkage, or they may form a part of the arm itself, and move with it. Separate spectacles allow the semaphore arm to be placed high with a sky background, while the lights are placed low, at the height of the driver's eye and usually with a dark background. Some early signals used rotating lamps, operated by bevel gears.

The colours of the night aspects is a subject of some complexity. Even in the early days, a good red glass was available, and red and white (clear) were the colours used in signalling. In January 1841, faced with Parliamentary investigation and possible legislation, railway managers met at the Queen's Hotel, Birmingham (later the Curzon Street station, and still standing) to discuss safety issues. Henry Booth of the Liverpool and Manchester was the driving force at this meeting, and contributed the results of 10 year's experience on that line. The recommendations were essentially the practice of the Liverpool and Manchester Railway. Hand signals were standardized, and the signal colours of red, green and white were adopted. Blue signals stopped a train for traffic, and black flags were used by track workers. Red indicated Danger, White indicated Safety, and Green indicated Caution, Go Slowly. Green was introduced by this conference, and became generally adopted for Caution in all countries. As we have already mentioned, blue was used instead in America, though the adoption of green was soon noted. Green competed with blue, but was did not completely replace it until after British signals were imported in the 1870's.

A good signal green glass was not available. Green glasses tended to be rather dark. There is not a great deal of green in oil lamp light anyway, so the greens used generally included a good deal of yellow, and were distinctly yellow-green. Blue and purple glasses gave good colours, but there was even less blue light available, so they were not very visible. The Chappes found that red glass dropped the visibility to 1/3, green glass to 1/5, and blue to 1/7 that of a bare flame. The result of all this was that there was no good yellow glass, and blue and purple were only used when a limited sighting distance was desired.

Signal colours were first investigated scientifically and with respect to colour blindness in 1855 by Dr. G. Wilson of Edinburgh. He concluded that red and green were poor choices, because: (1) red in shade tended to black; (2) red and green were complementary and did not contrast strongly, and each was the fatigue colour of the other; and (3) about 1 in 50 males could not distinguish them, especially in dim light. He proposed white, black and sky blue as the best triad, but of course there was no chance of adopting them. Yellow was the most visible colour, but could easily be confused with white. However, he suggested that it be given a trial, since the available colours were so few. He advocated replacing green for Caution by sky blue, but only for flags and targets. The difficulty of choosing signal colours made him recommend the use of form and motion instead, which indeed was done, as well as the use of pyrotechnic lights. Dr. Wilson originated the tests for colour perception by the use of yarns.

At the end of the 19th century, with the spread of bright electrical illumination, a movement began to replace white by the more distinctive green for Safety. This change happened by about 1890 in Britain, but not in France until 1937. Since there was little need by that time of a Caution aspect, this aspect was simply dropped. The Chicago and North Western very effectively used the simultaneous display of red and green, produced by a lamp and reflector, but this idea was not taken up by other companies. See the description in the article on this topic. Where a colour for Caution was thought desirable, as in France, yellow was eventually adopted for this purpose. Green and yellow signal glasses were reformulated to make them more distinct, the green becoming less yellow and the yellow more orangish. The work with coloured glasses added a new signal colour, lunar white, a milky bluish white easily distinguished from clear. Then, red, green, yellow, blue, purple and lunar were the available signal colours. Different glasses were used for oil or electric light, because of the different spectral natures of the two illuminants.

The arms of the early semaphores shown at the left are operated by endless chains from the foot of the post. The catch for the handle is not shown. The rotating lamp, not part of the original Pasley semaphore, is driven by bevel gears. The arm fits between the side boards of the post, a mechanically advantageous construction. Many of these semaphores had two arms side by side, with operating machinery on both sides of the post. The back of the arm was typically painted white. Posts were originally framed through the roofs of the signalbox, and several pairs of arms were mounted on each post if necessary. In Britain, an arm pointing to the left governed a train approaching the signal. Drivers had to learn which signals applied to which track, and where to stop.

Any uncertainty was soon eliminated by moving the semaphore to trackside, placing it at the point where it applied. A signal at Danger was not to be passed. Where trains ran to the left on double track, as in Britain, France and Belgium, the signals were placed on the left of the track with the arms pointing away from the track. If trains ran on the right, as in America, Germany and Holland, the signals were placed on the right. The signals themselves were enantiomorphs, or mirror images, in the two cases. Today, Belgian light signals used on the right are mirror images of those used on the left, an excellent practice. In America, on the Chicago and North Western, trains ran on the left on double track. The signals were exactly like those used on the right, but in this case were located to the left of the track, so the arms pointed toward the track. Semaphores with blades pointing across the track were used in electrified districts of the New York, New Haven and Hartford for better visibility. Some semaphores in Spain are also arranged like this.

The usual way to operate a semaphore signal is shown at the left. The signal arm is pushed "off" by the up-and-down rod. It falls to Danger, or "on", on its own because of the weight of its connections, or its own balancing. Here, a counterweight is shown as providing the weight together with the weight of the up-and-down rod. This signal has a separate spectacle that is operated by a linkage. If it were incorporated in the arm, or mounted on the other side of the post, it would add its weight to that of the counterbalance. This spectacle is a monocle with one red roundel. The counterweight normally rests against a stop. When the wire is pulled, it is raised, and the arm and spectacle are moved. The arm is depressed, and a white light is shown by night. This signal can be operated at distances of 1000 yards or more by pulling on the wire.

A typical lower-quadrant semaphore with a spectacle as part of the arm is shown at the right. When pulled off, it can be arranged to assume either a 45° depression or a 60%deg; depression. This requires that the wire be pulled the full anticipated amount. If it is too little, the arm is not depressed enough. Although many companies used 45°, the Great Western preferred 60°. This gives a much stronger aspect, and is a good idea. American practice used 60° almost exclusively. The 45° and 60° aspects cannot be used together as a three-aspect signal, since the difference between them is not sufficient. Lower-quadrant three-aspect signals were used in America around 1900, but the positions were horizontal, 45° and vertical. This diagram also shows the appearance of a signal in the United States, where the arm points to the right. When green was adopted for Proceed, a second roundel was added to the spectacle, which now indeed had two lenses.

The methods of signalling a facing junction are shown at the left. By a facing junction we mean one at which different routes can be taken by means of facing points. In this example, suppose the main route is straight ahead, while the branch is a divergence to the right. One method is to put one arm on the post for each route; here there are two arms. The upper arm refers to the leftmost divergence, the lower arm to the rightmost, and any intermediate arm to the route in the same order as the arm. In this method, the arm referring to the main route is not always in the same place. If the main route can be taken at a higher speed than the other routes, this is inconvenient. One solution to this problem is to emphasize the arm referring to the main or high speed route by changing its shape, and perhaps by displaying a double light by night.

A second solution is to erect one post for each route, side by side, as shown by the middle two signals in the diagram. If one route demands a slower speed, its arm may be mounted lower than the arm of the other. Here the two arms are at the same level, which may imply that the routes are equal in speed. A less expensive third solution is to use one post or doll for each route on a bracket. Here, the main arm is on the main post and is higher than the other arm on the doll to the right. This indicates that the divergence is to the right, and it demands a lower speed than the main route. If the two arms were the same height, then the two routes would be equivalent. This method of signalling by giving the route was always used in Britain. The driver's route knowledge supplied the appropriate speed on the route indicated.

The first method, using multiple arms and called "signalling from the left" (or from the right if another convention was employed) was originally used everywhere. In France, the semaphore became indelibly associated with junctions, though all other signals were rotating vanes. Ultimately, semaphores were only indicators of direction, other signals having the duties of stopping, starting and warning trains. These signals began prescribing the speed applying at the junction, whether by rule or by sign. French signals still specify the route, by white lights in a horizontal row rather than by multiple semaphore arms, as well as the speed.

In Germany, giving route information was abandoned. Instead, the main signals displayed either Proceed or Slow Proceed (Langsamfahrt). Reduced speed was defined as 30 km/h, appropriate for most pointwork. Since many junctions can be taken at a higher speed, ways were found to specify different reduced speeds. This has become the usual way to handle junctions everywhere now, by giving warning of a divergence and specifying the appropriate speed over it.

In the United States, signalling from the left or right with multiple arms was generally used at first. Usually there were no more than three or four arms at most. Then it became the practice to put the arm for the highest-speed or main route on top, and those for slower routes below in some order. Although this convention may still be used, with the arms for the different routes specified in the time table, it became standard to use at most three arms, as shown in the diagram. The top arm applied to the main route, the second arm to medium-speed routes, and the lowest arm, usually with a short blade, to slow-speed routes. A full-length blade is 3' 6" long, and from 7 to 9-5/8" wide. The short blade is 2' 6" long. from 7" to 8-7/8" wide. The lights for the three arms were aligned vertically. The roundels are 8-3/8" diameter. Arms are separated vertically by 7' 0", the short arm from the long arm above it by 5' 0". The lowest arm was usually 22' 6" high on a ground mast. On a signal bridge, it was 3' 6" above the base of the mast. The highest arm was 4' 2" below the peak of the pinnacle. These were the 1910 RSA standards, not followed by all companies. This was pure speed signalling. Medium speed was the speed at which certain crossovers and junctions could be negotiated, at first 30 mph, and later 40 mph or even 45 mph. Sometimes a permanent sign specified the medium speed applying to the aspects of a certain signal. Slow speed was from 15 to 20 mph, and included restricted speed, which did not guarantee an unobstructed track.

Most current books of rules simply use the term "diverging" instead of "medium speed" and specify some way of determining the appropriate speed. There is no way of distinguishing different diverging routes, but this is seldom necessary.

When a signal consists of several dolls, in America each doll refers to a separate track, never to different routes for the same track. On quadruple track in semaphore days, signals to the right of the track had two elevated dolls, one for each of two tracks in the same direction. These were called bracket signals. Sometimes the arm for the less important track was lower than the other, and sometimes both were at the same height. If, due to lack of space, a signal had to be located beyond a parallel track, the parallel track was represented by a doll with a marker light, usually blue or purple. The same was done if the signal had to be placed to the left on double track, where trains kept to the right. A doll with a marker light then referred to the track for the opposite direction. If a signal had to be placed to the left instead of to the right of the track, the arm still pointed in the same direction. In Britain, semaphores could also be placed on either side of the track for best sighting. The Great Western often put signals on the right side so they were easier to see from the right-hand driving position.

In the days of time-interval operation, trains that were stopped outside a station or at a junction presented a danger that they would be run into before protection could be appplied. Also, if they were properly protected against trains running a full speed, flag protection would require considerable time. This problem was solved with the distant signal, placed at least the stopping distance in rear of a danger point. The first was erected in 1846 at a junction just east of Edinburgh. Semaphore distant signals were used on the Great Northern Railway from its opening in 1852. Drivers were often required to approach a distant signal prepared to stop in case it were "on". They could, however, proceed at restricted speed on to the danger point, looking out for trains ahead. If the distant were "off', they could proceed expecting signals to be clear. This was the theory behind the French disque rouge and similar "advance" signals which proved very useful. [An advance signal in Britain and America is one beyond a point, not one approaching it.] Even in Britain, the distant signal was first used in this way, and did not differ in appearance from a normal stop signal. A little later, it was distinguished by a notch at the end of the blade, but the blade was still red, and a red light was displayed by night. The notch seems first to have been used on the London, Brighton and South Coast around 1870.

Signals that can display stop have several different names in Britain, depending on how they are used. Typical signal layout on one track at a station is illustrated at the left. Not all of these signals may be present, but the home and starter are basic. An intermediate block post may have only a home signal. These signals are duplicated for the track in the other direction. The stop signal at the end of a block that protects a train that passes it is the home signal, and this is a general name for a stop signal. The stop signal that authorizes entry to a block is called the starting signal. The home and starting signals may be the same, but usually the length between them is station limits, controlled by the signalman alone. Additional stop signals beyond the starter are called advanced starters because they are "in advance" of the starter. The furthest one of these is the actual starter. Additional home signals in rear of the home signal are called outer homes. The reason for all these signals is the strict requirement in Britain that a clearance space of 1/4 mile must exist beyond (in advance of) a stop signal before a train can be accepted. If the station shown had only a home and starter, a following train could not be accepted while a train was standing at the platform. If an outer home were provided, then the clearance would be available, and a second train could be accepted. If the train at the platform could not be accepted into the block in advance, then an advanced starter would allow the train to pull up to it to await acceptance, again pemitting a following train to be admitted to the block. The distant signal is not pulled off unless all the stop signals in advance of it are off, assuring the train a clear passage. Otherwise, the signalman waits until the train approaches ready to stop, and clears the signals in time to allow the train to proceed at slow speed to the first signal that must remain at stop.

In America, only a home signal was generally used, which was also a starting signal. If a starting signal was provided, it was called an advanced signal, and its purpose was to permit a train to pull past the home signal to the station platform without being authorized to enter the next block. A block could not be reported clear unless the rear of the train had passed a certain distance beyond the block signal.

With the block system in effect, the strict precautions at the distant were not necessary. Every block signal served as a distant in that sense for the block signal in advance even in the case of the permissive block. Under the block system, the distant signal served only as a repeater for the home signal, warning the driver of a stop required ahead. If the distant were "off", then he could expect the stop signals ahead to be "off" as well, so he could proceed at normal speed. If it were "on", then he had to prepare for a stop ahead, but expected the track to be clear that far. This became the sole purpose of the distant signal, and it was then desirable to make it more distinctive. At first, there was no difference between the home and distant signal, the driver's route knowledge supplying the identification. Later, the signal was clearly distinguished by notching the end of the blade and painting the blade green with a red chevron. After yellow had been adopted instead of green for Caution, the blade was painted yellow with a black chevron. When signal yellow glass became available, the red lens was replaced by a yellow one, so the distant signal could be distinguished by night as well. Before this, in a few places, the Coligny-Welch lamp was used that displayed an illuminated chevron to the side of the spectacle. The Chicago and North Western displayed red and green simultaneously, as mentioned above, keeping the green arm with red chevron. The use of yellow for Caution eliminated all of these variations.

Some examples of British distant signals are shown at the right. A distant arm can be used alone on its own post, placed about 1000 yards from the stop signal. The exact distance depends on the maximum train speed and gradient. If blocks are short, the distant arm for the signal in advance can be place below the stop arm. These arms are interlocked so that if the stop arm is horizontal, the distant arm will be horizontal as well. This two-arm signal displays three aspects: Stop, red over yellow; Approach, green over yellow; and Proceed, green over green. American lower-quadrant signals were exactly the same, aside from being mirror images, and were used as automatic block signals as well as interlocking signals. At the right is a "splitting distant" that gives advance information about a junction in advance. The high arm refers to the main route, the lower arm to the diverging route. Splitting distants are now used for diverging routes with speeds of 40 mph or more. At one time, they were more generally used. For lower speeds, a fixed distant arm is used that displays a constant yellow light.

Distant signals created another need. Until they were used, signals were generally in view of the signalman, so their operation was easily verified. The operation of a distant signal out of sight, had to be proved to the signalman by other means. It was particularly important to know that the signal had returned to Caution. The earliest way was by means of a return wire operating a miniature signal in view of the signalman. Later, much superior electric repeaters operated by contacts on the signal were used. With electricity, it was also possible to prove that the lamp was lighted by means of a thermocouple-operated switch heated by the flame.

The three-aspect signal was reinvented in the United States to economize by using only one arm instead of two. The first three-aspect signal, invented by Gray of the Pittsburgh, Ft. Wayne and Chicago (PRR), moved in the lower quadrant, assuming horizontal, 45° and vertical positions. In the vertical position, the arm lay alongside the post. It was first used between Pittsburgh and Homewood in 1900. Three-aspect train order signals had already been used, but this was the first automatic block three-aspect signal. In 1903 Patenall and Loree designed the three-aspect upper-quadrant signal that later became standard. The first of these signals was installed by the PRR between West Philadelphia to Elwyn in 1906. The three aspects were interpreted as Stop, Approach and Proceed, the same as given by the two arms of a two-aspect lower-quadrant signal. Belgian railways adopted three-aspect signals in 1919.

The Santa Fe used the three aspects of the upper-quadrant semaphore, but interpreted them differently. The 45° aspect was not Approach, but Restricting. Apparently, if encountered at high speed it meant that speed should be reduced immediately to restricted speed. In effect, it had the same indication as a French disque rouge. At interlockings, it was used for all routes except the main route. This meant that the Santa Fe could use only one arm on any signal, even at an interlocking. This practice was still in effect in the 1927 Rule Book. By 1952 the 45° aspect with a number plate (as in automatic block signals) meant Approach, but if there were no number plate (as at interlockings) it still meant Restricting, or Approach-Restricting as stated in the Rule Book. The Santa Fe painted signal arms white, black or striped white and black, depending on the background.

The presence of a number plate turned a Stop into a Stop-and-Proceed aspect. Sometimes, a pointed blade end identified an automatic block signal giving Stop-and-Proceed, while a square end identified an interlocking signal giving Stop. If there were two arms, the lights were staggered vertically. These three identifying characteristics are illustrated at the right. On double track, a train could proceed at once at restricted speed after stopping at a Stop-and-Proceed aspect. On single track, a flagman had to be sent ahead before the train could proceed after 5 minutes, to a point where the next signal could be seen. This was to allow for the chance that the Stop signal was caused by an opposing train. Unexpected stop aspects were usually due to a train in the same direction immediately ahead, or else to a fault in the signal system, not to an opposing train, which would mean a breakdown of the time table and train order system. Number plates identified automatic block signals by milepost and tenths. Even numbers were used for one direction, odd for the other.

The two-arm automatic block signal could display additional aspects. Yellow over yellow was Advance Approach, indicating proceed prepared to stop at the second signal, providing four-block signalling and allowing higher speeds with the same signal spacing. Alternatively, yellow over green was Approach Medium, indicating approach next signal at medium speed, as an alternative fourth aspect, or in the approach to an interlocking. If the second arm was not required, a marker light might replace it so the signal would be uniform with those of two arms, and would be identifiable as an automatic block signal by night.

A type of signal related to the semaphore is the tilting bar signal, where the arm is pivoted at the centre. A typical example is shown at the left. Some customary semaphores may have arms pivoted toward the centre, as the somersault signals of the Great Northern Railway, or some Belgian semaphores, so that they balance better, but what we have in mind here is the signal invented on the Pittsburgh, Ft. Wayne and Chicago well before the 1870's to protect railway crossings at grade, and its descendants. Economics and the American prairies led to many level crossings of railways, since it was expensive to carry one line over another. The tilting bar signal was erected at a crossing so that it was visible from either line. When horizontal, it appeared horizontal from both, and when inclined or even vertical, it appeared so from both lines. Therefore, if the horizontal arm allowed one road to proceeed, and an inclined arm the other, conflicting signals could never be given. The value of this is evident, so these signals were widely used east of the Mississipi and north of the Ohio, though not spreading far beyond these geographical limits.

These signals were called crossing targets, though they were not targets at all. They were used on main lines, where crossing gates would have been inconvenient, and were operated by the telegraph operator. Although it was not theoretically necessary for trains to stop if the signals were right, in many cases state laws required a stop at crossings protected with crossing targets, which were not considered "interlocking". By night, lamps, usually red, were hung from the ends of the arms. This provided a distinctive "position light" aspect. In some cases, an ingenious arrangement of the ropes allowed the lamps to be raised and lowered from the ground. Otherwise, they were reached by a platform or ladder. Later crossing targets had electric lights. Such signals are still in service.

An unusual example of a tilting bar signal with two arms is shown at the right. This signal was erected by the Central Vermont at Willimantic, CT to govern a junction with the New Haven. The date 1852 is quoted in my source, but may not be correct, as it seems much too early. Upper arm diagonal, lower horizontal was for the NH Hartford line; upper horizontal, lower diagonal was for the Air Line; upper arm vertical, lower horizontal was for the CV. Both arms horizontal was Stop on all routes. The arms were operated by rods moved by levers at ground level. There was a platform just below the arms to reach the lamps. The figures on the upper arm were painted on, not apertures.

Many attempts were made to operate signals remotely or automatically by means of electricity. The aim was to overcome the limit of 1000 yards or so for mechanical operation, or the unsatisfactory performance of mechanical treadles. The effective solution to automatic operation was the closed-circuit track circuit of William Robinson, demonstrated in 1872 (U.S. Patent 130,661, 20 August 1872). The use of the rails as electrical conductors was shown by Steinheil as early as 1838. However, for signalling open track circuits were generally used, from Bull in 1860 in England (British Patent 2660, 31 October 1860), to Frank L. Pope in the United States in the 1870's.

The chief problem was the design of a suitable signal, since the electrical currents were then not strong enough to operate a regular semaphore signal. One course was to resort to weights, whose descent was electrically controlled, and the other was to use electromagnetic forces alone. These signals, of whatever form, were usually controlled by track instruments, or electrical contacts operated by the wheels of a train, if they were automatic. Examples are Rousseau's automatic block signals of the 1870's, or Pope's remotely operated "semaphores", which were not semaphores, but a disc signal.

Rousseau's signals were rotating coloured panels, driven by a weight inside the post and electrically released. These signals were operated by momentary currents that released the rotating panels. If standing at Clear, operation of a track instrument at the entrance to the block pulled on an armature that released the signal, which then stopped after a quarter-turn, displaying red. When the train operated a track instrument at the block exit, the rotor was again released, and went back to white. The switching was done by contacts opened and closed by the rotor. It is evident that there was a great danger of the signal's getting out of phase if any impulse failed, and this fault was fatal to the system.

Oscar Gassett of Boston designed a much superior weight-driven signal, called the "clockwork" signal (U.S. Patent 228,455, 8 June 1880). When current was supplied to it, it displayed Clear, and when the current was interrupted, it went to Danger. Instead of the rather obscure red and white silk panels inside a box, Gassett used targets at the top of the signal. A circular white disc in a black border indicated Clear, while no disc indicated Stop. Gassett advocated black backgrounds for the signal. The disc had hinged slats to reduce wind pressure. There was also an enclosed signal head, like the Rousseau signal, and a more open one, like a switch indicator, with red and white targets. The light was originally beside the mechanism at the bottom of the disc, but was soon placed above the disc on an extension of the axis. Gassett and Robinson went into partnership to form the Union Signal Company in 1878, which had reasonable success in their area, placing signals on the Boston and Albany, Providence and Worcester, Fitchburg (Boston to Waltham), and Old Colony, among others. After Westinghouse absorbed Union Signal in 1880 to create Union Switch and Signal Company, the clockwork banner signal was tepidly promoted (the pneumatic was their pet), and was supplied from 1883 to 1894, 1055 signals in all. The best thing about the signal was its track-circuit operation, and much was learned about track circuits and insulated joints in its use. The circuits were usually arranged so that the engineman approaching a clear signal could observe its changing before he passed it. If the weight ran down, the signal assumed the Stop position. The first use of track circuit overlap was with these signals on the Eastern Railroad, in 1880m from Salem to North Beverly. The overlap was 1200 ft.

The absence of a disc or something red for a stop signal worried railway authorities, who did not appreciate theories. One solution was to paint the disc red and use it as the Stop signal, leaving the other aspect for Clear. Examples from the New York, Ontario and Western Rule Book of 1913 are shown at the left. There was also a banner Caution signal, as illustrated. Sometimes a positive Clear signal was given by a target, as shown at the right. This target was usually white, but could, of course, be painted green after the change in signal colors. This signal was called a disc signal by the NYO&W; more generally, it was known at a clockwork or banner signal. It was not called a banjo signal.

Thomas S. Hall was disturbed by the frequent accidents at switches in New England, which encouraged him to the design of an electrical signal (U.S. Patent 103,875, 7 June 1870). This was a disc signal operated by a tractive armature enclosed in a protective case. The earlier patent referenced (62,414 of 1867) was actually for a vibrating bell that rang when a switch was moved from the main track. Hall should have referred to 89,308 of 1869, which was a disc signal to be used as a remote switch indicator. An early Hall signal house may have been used by Robinson in his experiments on the Philadelphia and Erie in 1872. By night, a lamp shone through the light red disc that fell in front of it when the circuit was open. When the circuit was complete, the disc was pulled aside, and the lamp was displayed. The development of a track instrument allowed the application of this signal as a block signal. By 1871, it was in use on the Eastern Railroad (16 miles) and on the New York and Harlem. In 1872-1873 the Eastern Railroad installed the first continuous automatic block system, between Boston and Beverly.

The signal and associated apparatus were subject to continual improvement, notably by Hall's son, W. P. Hall. In 1888, improved signals, with track instruments, were used on the Boston and Albany. After 1890 the later Hall disc signal was perfected, using track circuits (since the patents had expired), and the "normal danger" system, which economized on batteries (U.S. Patents 535,102 and 535,103, 5 March 1895). In this model, the night aspects had been separated from the day aspects, using a separate roundel, and when the disc was withdrawn, there was a circular aperture completely through the signal, which gave very good visibility even when the sun was behind the signal. The aperture was larger than the disc, so a white annulus showed around it. When the disc was withdrawn, a bit of it still could be seen at the edge, to show that it had not fallen off. This signal is illustrated at the left.

The design of a good electromechanical operating mechanism was a subject of rather wide interest and application. With any tractive armature, the width of the air gap decreases as the device operates, and this leads to a very rapid increase in tractive force, which is undesirable. The ideal is a more constant force over a good distance. This can be accomplished to some degree in a solenoid by using a cylindrical plunger sliding in a thin-walled iron tube. The movement of the plunger changes the effective area of the magnetic circuit, giving a force over a larger distance. The 1888 Hall signal used an improved solenoid, still with a tractive armature but making use of a flexible link chain to move the disc. Finally, around 1890, the Z-armature was discovered, which was not only simple but gave a good torque over a large angle of rotation. The inventor appears to be H. E. Boothby (U.S. Patent 537,268, 9 April 1895). The air gap is nearly constant in this solenoid.

The popularity of the Hall disc signal led Union Switch and Signal to offer their own disc signal, which replaced the clockwork signal in their line in 1895. It was designed by Jens G. Schreuder (U.S. Patent 550,535, 26 November 1895). The patent is remarkable in that there was almost nothing to be claimed but that the rear aperture in the case was smaller than the front aperture, which was hardly an improvement. The disc was enamelled aluminum, and the red lens was an oiled paper, for lightness and resistance to moisture. The motor was a Z-armature rotary solenoid. The case was about 4 ft in diameter, and very thin. Both it and the motor housing were hermetically sealed. The case rotated on the mast for ease of maintenance. When out of service, it could easily be rotated parallel to the track. There was a ladder, not shown in the figure.

This signal was notably used on the single-track automatic block of the Cincinnati, New Orleans and Texas Pacific, on the 132-mile "Rat Hole" division. This was the first considerable installation of automatic block on single track. Signals were first installed on this line in 1891 to protect tunnels as isolated blocks, so there were some clockwork signals as well as Hall discs in addition to the US&S signals. Bicycles were carried in baggage cars for the flagman who had to go ahead if a signal was encountered at Stop. The disc signal had a rather short day, since motor semaphores soon replaced disc signals entirely.

Throughout the later 19th century, many patents for electrical railway signals can be found. They usually appeared subsequent to a significant patent, hoping to interfere with it, but were often spontaneous and independent. Very few, if any, were ever even tried in practice, and many were simply impractical. These include Le Grande's signal of 1883 with a radial "sunburst" display, Stuckey's disc signal operated by a treadle of 1892, with no evident way of clearing the signal again, Hughson's electric flagman who waves a flag and rings a bell of 1877, Conklin's disc signal of 1876 operated by treadles at the ends of the block, but whose electromagnet seems too delicate to rotate the disc, and nothing to use out-of-doors, and Harrington's semaphore of 1890 that moved mercury to and fro with an electric valve and track instrument to balance the arm one way or the other. These are simply a few of the more reasonable patents from the American literature. Similar patents appeared in Britain.

Vanes and semaphores can only be seen by day, or in the beam of the headlight. Lights for the night aspects were required from the first; indeed, they may even have preceded day aspects in fixed signals. Lights produced by the illuminants of the time were not visible by day, so signals necessarily had a dual nature. Not until electrical lamps became available was it possible to use the same aspects by day as by night, and these were, of course, the night aspects. This change depended more on the availability of electricity than on any other factor, so light signals could not be generally used until after about 1920.

Lights can give information either by their colour, by their position, or by blinking, and all of these means have been used. Most used has been colour. The majority of modern signals use red, green, yellow, lunar white, blue or purple lights, mostly the first four. Until the early 20th century, white, which was the untinted flame, was also used, and yellow was not used. White was far more visible than any other colour, but could not compete with extraneous electric white lights. Then, as we mentioned above, green replaced white for Clear, Proceed or Safety, and yellow replaced green for Approach or Caution.

With respect to the use of colour, it is interesting to quote the "Observations on Nocturnal Signals in General" of 20 January 1823 by the Royal Engineers: "Many attempts have been made from time to time to use coloured lights for night signals, all of which have failed, for the colour of a luminous point or line cannot be distinguished at any distance." Not only do atmospheric conditions contribute to this conclusion, but the colour sense of the eye itself is not reliable, depending on intensity, contrast and fatigue as well as on anomalous colour vision, which is common in men. The eye is relatively insensitive to red or blue light, which are at the ends of the visible spectrum. In spite of this, colours have been extensively used as signals, and have been satisfactory where the demands are not too great. For semaphores, the different colours are produced by moving coloured glasses in front of a flame.

The perception of position, and in particular the directional contrast of lines, is very sensitive. If lamps are hung at different points on a semaphore arm, distinctive night aspects result. Before 1889, the Boston and Albany hung two white lamps on semaphore arms. By night, two lights in a horizontal line indicated Danger, while two lights in a vertical line indicated Safety. The Old Colony erected similar signals for its 84-lever interlocking at Mansfield, Massachusetts. Two green lights in a vertical line indicated Safety, in a diagonal line Caution, and two red lights in a horizontal line indicated Stop. These were called "Hardy" or "B&A" signals. The idea is also used with crossing targets. These are all position-light aspects, and could have been used more widely than they actually were.

Outside of crossing targets, the first idea for a position light signal was the illuminated semaphore arm. This was a night aspect, of course, made by reflecting a light at the axis of the arm from the blade, perhaps with a change of color of red to white as the blade went from horizontal to depressed. Around 1887, V. K. Spicer and J. G. Schreuder of Union Switch and Signal designed a blade with a corrugated linear reflector, 32" x 2-1/2". These were used on the Old Colony and on the Pittsburgh, Ft. Wayne and Chicago, and with short arms on the Chicago and North Western. Dr. Herschel Koyl about the same time designed a semaphore arm as a corrugated parabolic reflector 5' 9" long. The lamp, at the focus of the parabola in front of the signal, showed red and white lights. In both cases, the corrugation was to spread the beam of light about 16° on either side of the axis.

It is curious that while these blades were distinctive at short distances, at long distances they became simply a blur of light, and it was not possible to discern their position easily. With the Koyl semaphore, the change in color might have been a help. We'll see that when replaced by a row of a few point sources, the direction is more easily distinguished.

Position-light signals are easily made by switching electrical lamps, and all mechanical elements can be eliminated. This was first done by A. H. Rudd in 1915 on the Pennsylvania Railroad, who remembered the illuminated semaphore and its fate. Experiments were carried out near Philadelphia, and the first signals were installed on the newly electrified line from OB to Paoli. Although these signals used four lights in a line, it was soon found that three lights were equally well perceived. These signals could be read at large distances. Even when the individual lamps could not be distinguished, the direction was evident, because of the remarkable sensitivity of the eye to the direction of a luminous line. Incidentally, fewer than three light points do not produce this distinction reliably, nor do more than four. The lights used are lightly tinted yellow, so little intensity is lost; they are essentially still white lights. This makes the signals economical with electricity. In fact, these are the only light signals that can be economically operated with primary cells, using approach lighting. They use less electricity than a colour-light signal, since they do not lose intensity in absorption and are easily perceived against their dark backgrounds. They are, however, more expensive in first cost.

The aspects are shown above. These are the 3-light signals adopted in 1921. The lights are 5" diameter on an 18" circle with centre 17' 0" above the ground. Arms are separated by 7' 0". The backgrounds are 4' 4" diameter. Clear-block and Permissive-block are block signals. Clear is not authority to enter a block. Caution is a venerable aspect peculiar to the PRR that is used to protect a turnout in advance, perhaps in the middle of a block. Approach-medium and Medium-clear were adopted in 1943; previously, they had been Approach-restricting and Clear-restricting, but "restricting" had been half maximum speed but not over 30 mph, so this was just a clarification of terminology. Approach-slow was added later; it indicates proceed at medium speed approaching next signal at slow speed. If the lower arm is flashing in Slow-approach, the aspect is Medium-approach instead. Note that Approach-slow and Slow-approach have quite different indications. Two forms of dwarf signals are shown; the upper ones were used after 1930 to give added clearance. The lights are 4" diameter frosted lunar white on an 8" circle. At the right are the PRR pedestal signals, introduced in 1930. they are 7' 0" high, 1' 4" wide, and the light groups are 2' 2" apart. They are used in tunnels, and when clearances are restricted.

An amusing curiosity was that some official in the 1950's thought red lights should be used for stop, and ordered that the row of three lights for Stop be replaced by two red lights at either end of the row. These signals were much less visible than the unmodified signals. From a distance, a signal going to Stop seemed to disappear, where previously the horizontal row of white lights was easily visible.

The Baltimore and Ohio developed colour-position signals whose main arm showed two coloured lights, in a vertical line for green, a 45° line to the right for yellow, horizontal for red, and 45° to the left for lunar white. These are very attractive signals, especially by night, when they are very bright. However, the use of two coloured lights is really superfluous, and these signals have none of the advantages of the position-light signal. They are expensive and use a lot of electricity. They were never adopted anywhere other than on the B&O and its subsidiaries.

These signals did, however, introduce a new and valuable idea. The aspect of the main arm is modified by marker lights above and below it. A lighted marker above means normal speed, a lighted marker light below means medium speed. No marker light implies slow speed. Other marker lights to the right or left above indicate medium or slow speed at the next signal. Red lights always mean Stop with these signals; a driver never has to pass a red light when he should proceed. Further aspects using flashing lights have been added.

The combination of a position-light main arm with these marker lights would create an economical signal not depending on colour. This was never done.

In motion signals, we must distinguish between the illusion of motion produced by blinking lights and physical motion. In either case, the attention of the eye is actively attracted.

Blinking lights are easy to obtain with electric lamps. They are generally used to make an aspect less restrictive, since the failure of the blinking apparatus usually results in a steady light. The availability of solid-state blinkers without mechanical parts makes them even more attractive. It is less well known that acetylene lights could be made to blink by modulating the gas supply, and this can probably be done with propane as well. The illusion of motion can be produced by blinking lights, for a signal that attracts the attention. This is evident in the two flashing red lights of a road crossing signal.

Physical motion can be used for a signal that actively attracts attention. These signals have not prospered, except for road crossing signals that must attract attention and operate only intermittently. An example is the "wig-wag", one form of which was invented by D. H. Wilson (U.S. Patent 533,938, 12 February 1895), driven by a tractive armature. This became a familiar crossing signal as a circular target with a black cross and a red light in the centre suspended above the road, that oscillated from side to side, widely used by the Santa Fe. This signal was motor-driven.

The Bezer rotating signal (U.S. Patent 602,792, 19 April 1898, filed 3 December 1895) was actually tested on the Delaware, Lackawanna and Western at Kingsland, NJ in 1896, where 3 signals protected a drawbridge. It was a semaphore arm revolving about its centre at 10 times a minute, with a revolving lamp with four lenses, alternately red and white, geared to it above, driven by a small electric motor (by Lundell, 4V, 2.5A). The lamp rotated at twice the speed of the arm. Normally, the arm was locked at horizontal. When a train approached, and the block was clear, the arm would begin to rotate, and the lamp would flash red and white at 1.5 Hz. If the block were occupied, the arm would remain horizontal and a red light would be displayed. The signal was originally designed as an ordinary 3-aspect semaphore. No further mention of this signal was found.

J. D. Taylor invented a similar signal, with four arms (U.S. Patent 516,558, 13 March 1894), proposing to use two mechanisms, as home and distant signals, on the same post. The arms alternately eclipsed a light and its reflection. Taylor used a motor-generator to supply the large currents required, and the signals were not normal-danger, so they rotated constantly. This current requirement and the normal-clear operation doomed the system. J. H. Farrar (U.S. patent 369,929, 13 September 1887) used a lamp inside a cage rotated by a small motor. Rotation again implied Safety, and the signal was to be activated by a track circuit. When the track circuit was occupied, current was removed from the signal, and it would stop. Application to a grade crossing was suggested, with manual control of the signals. In this case, the signals would normally be at stop, and would only operate intermittently, saving the batteries. All the small motors then available used far too much current for constant supply by primary cells.

R. C. Rapier, On the Fixed Signals of Railways, Min. Proc. Inst. C. E., 38, 142-247 (1874).

American Railway Signaling Principles and Practice, Chapter I, History and Development of Railway Signaling (Chicago: AAR Signal Section, 1955). Most of the illustrations of early signals are only artist's imaginings, and many dates and explanations are incorrect. For example, Ashbel Welch's signal is confused with the different and later PRR banner signal, and the date is incorrect. This is a much better source for later developments, after about 1890.

The New Haven Railroad Historical and Technical Association, The Shoreliner, Vol. 23, Issue No. 1, pp. 18-21 (1992).

R. J. Cook, Fostoria...All Trains Stop!, Railroad Magazine, Vol. 49, No. 3, pp. 24-31 (August, 1949).

E. T. MacDermot, History of the Great Western Railway (London: Ian Allan, 1964; orig. published GWR, 1927). Vol I, pp. 309-328.

E. Waytel, The First Position Light Signals and Subsequent Developments (PRTHS, The Keystone, Vol. XIV, No. 4, pp. 4-17, 1981).

Memorial Asking the Establishment of an Uniform Code of Danger Signals (Proceedings of the Franklin Institute, 84, pp. 135-137, 1867).

Rules and regulations, proposed to be observed by Enginemen, Guards, Policemen, and others, on all railways, recommended by the Railway Conference held at Birmingham, England, Jan. 1841 (Journal of the Franklin Institute, 33, pp. 239-242, 1842).

G. Wilson, On Railway and Ship Signals in Relation to Colour-Blindness (Trans. Roy. Scottish Soc. of Arts, 3, 127-152, 1855).

Composed by J. B. Calvert
Created 25 July 2004
Last revised 15 August 2004