If the reader has grasped the full significance of interlocking, he understands that it makes it impossible to give a signal that would lead to a collision or to a derailment at a misplaced switch. The worst that a stupid, or drunken, or malicious signalman could do would be to delay traffic, if the signals were obeyed. Here comes in the failing case. The brake-power may be insufficient to stop a train after a danger signal is given. That is a rare occurrence, but may happen. The engineer may not see the danger signal because of fog, or he may carelessly run past it. Provision against a failure to see and to obey a signal may be made by placing on the track a torpedo, which will explode with a loud report when struck by a wheel. The use of hand-torpedoes in fogs, and for emergencies in places unprovided with fixed signals, is very common. These are little disks filled with a detonating powder, and provided with tin straps that are bent down to clasp over the top of the rail. A simple and very efficient torpedo machine, which has been used for some years on the Manhattan Elevated and elsewhere, is here shown. This machine has a magazine holding five torpedoes. It is connected to a signal-lever in such a way that, when the signal is put to danger, one torpedo is placed in a position to be exploded by the first passing wheel. When the signal returns to the clear position the torpedo, if unexploded, is withdrawn to the magazine. If the torpedo is exploded another one takes its place at the next movement of the signal-lever. One of these machines on the Elevated Road moves about five thousand times every day. In such a case a torpedo would soon be worn out if it was not exploded or frequently changed. When this apparatus is in operation, an unmistakable alarm is at once given to the engineer and to others if a danger signal is passed. On the Manhattan Elevated lines an engineman who overruns a danger signal and can show no good reason for it is suspended for the first offence, and discharged for the second. The torpedo makes it impossible for him to escape detection.

The second great class of signals comprises those which are intended to keep fixed intervals of space between trains running on the same track. These are block signals. The block system is used on a few of the railroads of the United States which have the heaviest and fastest traffic. Much the most common practice in this country, however, is to run trains by time intervals, and under the constant control of the train despatcher. In England the block system is almost universal. About ninety per cent. of all the passenger lines of that country are worked under the absolute block system.

When the block system is not used, it is quite common to protect particularly dangerous points, such as curves and deep cuts, by stationing watchmen there with flags or with some form of fixed signal. The watchman can notify an approaching engine-runner that a preceding train has or has not passed beyond his own range of vision; or can notify him that it has been gone a certain time. Travellers by the Philadelphia & Reading must have noticed the queer structures, with revolving vanes on top, looking like a feeble sort of windmill, which appear in positions to command a view of cuts, curves, etc. These are examples of the devices for local protection. The non-automatic block signal develops naturally from the protection of scattered points. Instead of placing watchmen at points of especial danger, they are placed at regular intervals of one mile, two miles, or five miles. Instead of the watchman looking to see that a train has disappeared from his field of vision before he lets another train pass, he uses the eyes of the next watchman ahead, who telegraphs back that the train has passed his station. Suppose A, B, and C to be three block-signal stations placed at intervals of two miles. When a train passes A, the operator at that point at once puts a signal to danger behind it. This signal stands at danger until the train passes B, and the operator puts his signal to danger, and telegraphs back to A to announce that train No. 1 has passed out of the block A B, and is protected by the signal at B. Then, and not until then, the operator clears the signal at A and allows train No. 2 to enter the block. Meanwhile train No. 1 is proceeding through the block B C, its rear protected at B; and the same sequence of events happens when it arrives at C as happened at B. This is the simplest form of block signalling. In the more elaborate form there are at each block-station three signals—the distant, the home, and the starting. The signals are often electrically interlocked, from one station to another, in such a way that it is mechanically impossible for the operator at A to give a signal for a train to pass that station until the signal at B has been put to danger behind the preceding train.

It is seen that no two trains can be in the same block and on the same track at the same time. If all run at a uniform speed, they will be kept just the length of a block apart. If No. 2 is faster than No. 1, it will arrive at B before No. 1 gets to C, but will have to wait there. The block system, therefore, while it gives security, does not always facilitate traffic. The longer the blocks the greater will be the delay to trains; but the shorter the blocks, the greater the cost of establishment, maintenance, and operation.

Various systems have been contrived to have block signals displayed automatically by the passage of trains. This, if it can be done reliably, will do away with the wages of part of the operators, and will also eliminate the dangers arising from human carelessness. But there are very great objections to relying solely upon the automatic action of signals, and automatic block signals are little used except as auxiliary to a system employing operators also. So used, they are of decided advantage, as they make sure that a danger signal is set behind every train in spite of the operator, and that it cannot be again set to the all-clear position till the train has passed out of the block. All this is accomplished by electricity.

Brakes, interlocking, and the apparatus of signalling have been considered at length because they are very much the most important of all the appliances which go to increase the safety of operating railroads. They act chiefly to prevent collisions, but often prevent or mitigate accidents from derailments and other causes. Of all train-accidents happening in the last sixteen years, over one-third have been from collisions, and more than one-half from derailments.

After brakes and signals, the devices next in importance as means of saving life are those for the protection of highway crossings at the grade of railroads. In years to come, as wealth increases and as traffic becomes more crowded, we may suppose there will be few such crossings; but their abolition must be slow, and meantime the loss of life at them is great. The most accurate and complete statistics bearing on this matter are those collected by the Railroad Commissioners of Massachusetts. In 1888, of all those killed in the operation of the railroads of the State, seven per cent. were passengers, thirty-three per cent. were employees, and sixty per cent. were others. The others include trespassers, forty-seven per cent.; and killed at grade crossings, eleven per cent. More trespassers were killed than any other class; but the deaths at highway crossings considerably exceeded those among passengers. The difficulty of preventing this class of accidents is strikingly shown by the fact that, of all crossing accidents, forty-two per cent. were due to the victims' disregard of warnings given by closed gates or flags. It is evident that the efforts of the railroad companies to save people's lives at crossings are largely nullified by the carelessness of the public, and the lack of proper laws to punish those who venture upon railroad tracks when they should keep off them. Still, it remains the duty and the policy of the railroads to protect street crossings by all practicable means. The best protection is afforded by gates with watchmen, and of all forms of gate the most common, because it is the simplest and most convenient to operate, is the familiar arm-gate. This is usually worked by a man turning a crank, but it is also worked by compressed air. On this page is shown a group of gates worked from an elevated cabin by a mechanical connection. A bell fixed at a crossing, to be rung by an approaching train, is a very useful auxiliary to gates and to watchmen with flags, and is considerably used where the traffic does not warrant the expense of maintaining a watchman. There are several good devices of this sort, either electric or magneto-electric. One of the latter class has a lever alongside the rail, which is depressed by each wheel that passes over it. This lever is geared to a fly-wheel, which is set rapidly revolving and causes an armature to revolve in the field of a magnet, and thus generates a current and rings a gong, precisely as is done with the familiar magnetic bell used with the telephone.

About thirteen per cent. of the train-accidents in the United States, in the last sixteen years, were derailments due to defects of road. These include not only defective rails, switches, and frogs, but bridge wrecks. There are, however, few devices used in the track, other than those already mentioned, that can be called safety appliances. This class of accidents is to be provided against only by good material, good workmanship, and unceasing care. Many so-called safety switches and safety frogs are offered to railroad officers, but those actually in wide use are confined to a very few standard forms. The split-switch, which is shown in the engravings on pages 206 and 207, has gradually replaced the old stub-switch, as well as most of the "safety" switches that have been from time to time introduced; although the stub-switch is still in considerable use in yards where movements are slow, and in the main tracks of the less progressive roads. It consists of a pair of moving rails the ends of which are brought opposite to the ends of the main-line rails, or to those of the turnout, as the case may be. It follows that but one of these tracks is continuous at any one time, and a train reaching the switch by the other track must be derailed. The distressing accident which happened at Rio, Wis., in 1886, where seventeen people lost their lives, was a derailment of this sort. Since that time the railroad on which the accident happened has taken out all stub-switches on thousands of miles of main-line track. The split-switch provides against such derailments, for if the switch is set for the turnout, and a train approaches it from the main line in the "trailing" direction, the flanges of the wheels move the switch-rails to make the track continuous. The terms "facing" and "trailing," as applied to switches, are almost self-explanatory. If a train approaches toward the points of the moving rails, the switch is said to be facing. If it runs through the switch from the rear of the moving rails, the switch is said to be trailing. This will be made clear by reference to the illustration on page 206. If a train were coming from the bridge, the first switch reached by it would be a trailing and the second a facing switch. In the newspaper reports an accident will very often be assigned to one of two causes, failure of the air-brakes or spreading of the rails. The chances are that it will be found on investigation to be due to neither of these causes. Those interested to maintain the credit of the air-brake or of the track department are not often on the ground when the reporter gets his information, and the temptation is always great to shift the responsibility to the shoulders of the absent. Probably the displacement of the rail will have taken place after the derailment; but rails do sometimes spread. Loose spikes and rotten ties allow the outer edge of the rail-flange to sink into the wood, and the rail to roll outward enough to let the wheels drop. Sound ties are the first safeguard against such accidents. Metal plates under the rails are useful also; but one of the most efficient means of preventing displacement of the rails is the interlocking bolt shown above. These bolts cross in the timber, and slots cut in the two bolts engage with each other in such a way that when the nuts are screwed down on the rail-flange it is impossible to pull the bolts out. They can only be moved by tearing through the wood contained in the angle between them. This bolt is much used on bridges and trestles, where it is of vital importance that the rails should be held in place and no part of the floor broken.

In 1853 an express train went through an open draw at South Norwalk, Conn., and forty-six lives were lost. This, one of the most serious railroad accidents that ever happened, is still remembered as an historical calamity. The bridge which stands on the same site is shown opposite. In May, 1888, a west-bound express train, consisting of an engine and seven cars, was derailed just as it was entering the draw-span. The train ran three hundred feet on the sleepers before it was stopped. Then it was found that all of the driving-wheels of the engine had regained the rails, but all the other wheels were off, except those of two sleeping-cars in the rear. This was a remarkable escape from a bad accident, and much of the credit of it has been given to the interlocking bolts with which the rails were fastened. They are supposed to have prevented the rails being crowded aside, and thus to have made possible the rerailing of the engine. Besides, they helped the oak guard-timbers to hold the ties in place. The destruction of a bridge in an accident frequently begins by the ties bunching in front of the wheels and allowing the wheels to drop through and strike the floor-beams below. For this reason guard-timbers, notched down over the ties, should always be used.

The traveller will have noticed, on all bridges of various roads, two rails placed inside the track-rails, and curved to meet in a point at either end of the bridge. These are known as inside guard-rails, and their function is to keep derailed trucks in line till the train can be stopped. Besides the bunching of the ties, there is danger in a bridge derailment that a truck may swing around and strike one of the trusses. Then the bridge is very likely to be wrecked. A further provision for the protection of bridges is the rerailing frog invented by the late Charles Latimer, whose name is dear to railroad men all over America. This consists of a pair of castings combined with inside guard-rails, designed to raise the derailed wheels and guide them on to the rails. There is no doubt that it has prevented several wrecks, although it has never been widely used. The subject of bridges should not be left without a word of explanation of the stout timber-posts often seen at either end placed in line with the trusses. These are designed to stop any derailed vehicle which might otherwise strike against and destroy a truss.

There is one track-fixture that has no duty or value except as it promotes safety. It helps only one humble class of railroad employees. That device is the foot-guard. At all places where two rails cross or approach each other, as at frogs and guard-rails, dangerous boot-jacks are formed by the rail-heads. The overhang of the heads of the rail makes it easy for one to so fasten his foot in one of those boot-jacks that it is hard to get it out. If a man finds himself in this position in front of an approaching train, he sometimes has the alternative of standing up to be struck by the engine or lying down and having his foot cut off. Fortunately this class of accidents is comparatively rare; probably not more than two or three per cent. of all deaths and injuries to passengers and employees is caused in this way. Nevertheless, the means of guarding against accidents of this class is so cheap that it should be more generally adopted than it is. It consists simply in partly filling the space between the rail-heads by putting in wooden blocks or strips of metal, or even packing with cinders, gravel, or any sort of ballast. Various wooden and metal foot-guards have been patented. They are all too simple to require description.

Of all accidents to employees the most numerous are those which arise in coupling and uncoupling cars. In Massachusetts, in 1888, the employees killed and injured were 391; of these casualties 154 occurred in coupling accidents. The commissioners of other States, especially of Iowa, have for years published statistics showing nearly the same ratio. Fortunately accidents of this class, although numerous, are not proportionately fatal. Far the greater part of them result in the loss of part of a hand; but they are so frequent as to have caused much discussion, legislation, and invention. Several States have, one time and another, passed laws requiring the use of automatic couplers; and two or three years ago there were on record in the United States over four thousand coupler patents. The laws have been futile because impracticable; and most of the patents have been worthless for the same reason. It was obvious that the business of supplying couplers for the one million freight cars of the country could not be put into the hands of some one patentee unless his device was manifestly and pre-eminently superior to all others. It became important, therefore, to select as a standard some type of coupler general enough to include the patents of various men, and at the same time so definite that all couplers made to conform to the standard could work together interchangeably. Those who read Mr. Voorhees' story[21] of the wanderings of a freight car will understand that any one freight car in the United States or Canada should be prepared to run in the same train with any other car. A few years ago a committee of the Master Car-builders' Association was appointed to choose and recommend a type of coupler to be adopted as the standard of the association. After prolonged and careful study of the subject, the committee recommended the type of which the Janney is the best known example, and that has now become the standard of the association. This action does not give a monopoly to the Janney company, as there are already half a dozen couplers which conform to the type. This coupler is shown by diagrams in the article by M. N. Forney, page 142. A perspective view is herewith given. This device couples automatically, and thus does away with the necessity for the brakeman going between the cars. It can also be unlocked by the rod shown extending to the side of the car, and the locking device can be set not to couple, to facilitate switching and yard work. The mechanical principles of this coupler are a great and important improvement upon any form of link-and-pin coupler; and the coupler question has now come to this point: A type of coupler has been selected by a technical body representing most of the railroads of the United States. It is general enough to avoid the evils of a patent monopoly. It promises to be economical in operation, and will certainly do away with the terrible loss of life and limb which results from the use of the non-automatic coupler. The railroads are adopting it with reasonable speed, perhaps, but not as rapidly as simple considerations of humanity would dictate.

Closely related to the coupler is the vestibule, which within the last two years has become so fashionable. The vestibule is not merely a luxury, but has a certain value as a safety device.[22] The full measure of this value has not yet been proved. Occasionally lives are lost by passengers falling from or being blown from the platforms of moving trains. Such accidents the vestibule will prevent, and, further, it decreases the oscillation of the cars, and thus to some degree helps to prevent derailment. It is also some protection against telescoping. A few months ago a coal train on a double-track road was derailed, and four cars were thrown across in front of a solid vestibule train of seven Pullman cars approaching on the other track. The engine of the vestibuled train was completely wrecked. Even the sheet-iron jacket was stripped off it. The engineer and fireman were instantly killed, but not another person on the train was injured. They escaped partly because the cars were strong, and partly, doubtless, because the vestibules helped to keep the platforms on the same level and in line, and thus to prevent crushing of the ends of the cars.

The number of passengers burned in wrecks is greatly exaggerated in the public mind; but that fate is so horrible that it is not wonderful that "the deadly car-stove" should be the object of persistent and energetic attacks by the press and in State legislatures. The result has been the development, in the last three years, of the entirely new business of inventing and trying to sell systems of heating by steam or hot water from the locomotive, and even by electricity. In fact, the manufacture of such apparatus has already become an industry of some importance, several thousand cars being equipped with it. This whole matter of steam-heating is still in a somewhat crude state, and it does not seem desirable to force it by legislation. It has been demonstrated that it is the cheapest way of heating trains, and the most easily regulated; and it has become a good advertisement to attract passengers. Consequently the whole subject may be safely left in the hands of the railroad companies, and allowed to develop itself naturally in a business way. There is not yet any system of continuous heating so perfected that a railroad company could without hardship be compelled to adopt it for all its passenger equipment.

Fires in wrecked trains have originated probably quite as often from kerosene lamps as from the stoves. The danger of fire from this source, and the desire to give passengers the luxury of sufficient light, have led to methods of lighting by gas and, more recently by electricity. Lighting by compressed gas ceased years ago to be an experiment. In Germany it is almost universal, but in this country it has been brought into use very slowly. The system is almost absolutely safe, not unreasonably expensive, and may be made to give satisfactory and even brilliant illumination; but the ideal light for railroad trains will probably be found in electricity. It is even safer than gas, and is the most adaptable of any known method of lighting. Some sleeping-cars that have been recently put in service on the Chicago, Milwaukee & St. Paul Railway are provided with small electric lamps in the sides of the car, between each two adjoining seats, so that the occupants can read comfortably either when sitting in their seats or lying in their berths.

It is not to be supposed that so large a subject as that of safety appliances can be exhaustively treated within the limits of one article. It has been thought best, therefore, to give most of the space available to the two or three devices of greatest and most useful application. There remain various others that are in daily use, and that have important offices, which have not even been mentioned. If the reader has gleaned from these very incomplete notes some clearer notions than he had before of the means by which the power of the locomotive is guided into safe and useful paths, the writer's object has been accomplished.