Two excellent books have recently been published on the railways of Indiana, one by Simons and Parker covering all lines, and one by Watt specifically on the Pennsylvania Railroad. Some technical information is presented, which is welcome, but it is, unfortunately, not well illustrated. Signaling and operation are, as usual, very lightly treated. There is, however, a photograph in Simons and Parker, p. 40, showing an LS&MS bootjack signal as well as dwarf semaphores at La Porte, Ind., and a few train order signals are shown in station pictures. Alas, the typical "crossing target" or balance-beam signal is not shown in either book. One of the final chapters of Simons and Parker is entitled Disasters, and discusses the most notable accidents occurring in the state from the beginning of railways there. This listing is a convenient basis for a brief discussion of safety and the causes of accidents, as it seems to offer a representative selection of incidents. The exact details of the individual accidents must be sought in the accident reports, where such exist, and where they do not exist, from the more doubtful contemporary sources. Here, I shall discuss them as examples of types, and try to give some insight into the factors entering such kinds of accidents in general. Where these accidents are discussed in popular works, the words used sometimes imply a rather shallow understanding of railway operation and signaling.
Simons and Parker list 18 accidents, all of which involve passenger trains. The more numerous accidents involving freight trains are not mentioned, which is quite normal in newspaper reports and in the public consciousness. A freight-train accident was something internal to the company, perhaps very costly in employee casualties, property loss, and disruption of service, but not directly involving the public. Passenger train casualties, however, were a public matter to which everyone was exposed. Newspaper reports were usually exaggerated and contained snap judgments based on little evidence and less knowledge, and were often merely attacks on the companies. Accurate and universal reporting began only in 1901, when accident reports were required by the ICC. Some state governments had had such requirements for many years, and some observers, such as C. F. Adams, Jr., and B. B. Adams, editor of the Railroad Gazette, had tried to make proper and balanced analyses. Watt adds one more accident to the list of Indiana accidents in Simons and Parker. Some, but not all, of the accidents mentioned in Simons and Parker are also at least listed in Shaw, but often with different casualty figures. These "sampling errors" must always be taken into consideration when discussing railway accidents.
Of the 18 accidents, 6 were due to failures of bridges, track or equipment. The single derailment, which occurred in 1864 on a bridge, was sabotage. An accident in 1907 was the result of an explosion of gunpowder on a train on an adjacent track. The 1947 accident at Walton, Ind., not mentioned in Simons and Parker, with 4 passenger fatalities, was caused by vandalism. The remaining 12 accidents were operational errors, and it is these that mainly concern us here.
One interesting observation concerns the relative safety of operation of the companies involved. In Indiana, the New York Central System and the Pennsylvania Railroad eventually controlled just over 20% of the state's railway mileage each, and the majority of the traffic of the state between them. The New York Central System comprised the Big Four (Cleveland, Cincinnati, Chicago and St. Louis) centered in Indianapolis, and the east-west main lines of the Lake Shore and Michigan Southern and the Michigan Central in the far north of the state. The Pennsylvania System comprised the Panhandle Route (Pittsburgh, Cincinnati, Chicago & St. Louis) and the Pittsburgh, Ft. Wayne & Chicago, the main line to the East. Of all the accidents listed, the New York Central System was the scene of 8, nearly half, with 259 fatalities. The Pennsylvania System was the scene of one, which killed no passengers, but also of the Walton accident of 1947 mentioned in Watt with 4 fatalities. The disparity is far greater than could be the result of chance, and suggests some fundamental differences in management. In fact, the record of the Pennsylvania is remarkable.
There is only one head-on collision from the early days, in 1864, between Lafayette and Indianapolis on what became the Big Four, in which some 30 soldiers died. We don't know why this collision occurred, but some possibilities are, in order of probability: (1) confusion over the rights of delayed trains, (2) defective watches, (3) misinterpretation of orders or messages (this was before the days of proper train orders), and (4) lack of observance of clearance times. Most railways in Indiana were single-track, and the absence of head-on collisions is notable.
In later days, the several bad head-on collisions teach excellent lessons. In 1906, on the B&O Chicago line at Woodville, 44 passengers died when a freight train collided with theirs. This accident is fairly well described by Shaw. The passenger train was running in two sections, the first carrying green signals to show a train was following. Normally, a superior train need not concern itself with any inferior train. The inferior train must keep clear of the superior train, and not proceed until it has passed. The trouble with sections is that the train is in two parts, and has not passed until both have passed. The engine of the first section carries green flags and lamps to show that there is more train coming, an old American practice that began with red flags instead of green. On this stormy night, both the green lamps had gone out. That both could have gone out is extraordinary, as is the fact that the engineman had not noticed the fact. However, the lamps are not depended upon alone. On seeing the freight in the siding, the engineman should give the whistle signal -oo to call attention to his flags, and the freight engineman should acknowledge by two short, oo. If the signals are not acknowledged, the superior train was obliged to stop to make sure the second section was protected. On this night, this did not happen. One suspects that the passenger engineman simply did not notice the freight at all, standing with its headlight extinguished in the storm. This suggests some of the hazards of running sections, even close together. The only good way to protect sections is by train order (by giving all opposing trains a copy of the order creating the sections), but this amounted to the extra work that the running of sections was to eliminate.
In 1912, a passenger train of the Cincinnati, Hamilton and Dayton collided head-on with a freight train near Indianapolis with 15 fatalities. This was not due to improper rights, but cannot have been other than gross negligence of the employees involved. The passenger train went through an open switch to collide with the freight train on the siding. It is said, "an employee neglected to close the switch." Well, perhaps so or perhaps not. It was the duty of the freight train to be on the siding with switch lined and locked ten minutes or so before the passenger train was due. That it was not, and that no flag had been sent out, shows the area where the blame must lie. At any rate, the freight engineman should have kept his headlight on bright until the switch was lined, and until he had a signal from the rear end that the train was in the clear. One wonders whether he had received the signal from the rear and extinguished his headlight before the switch was lined. The engineman is responsible for seeing that his end of the train is protected. In this case, the position of the switch may have been neglected, and the brakeman who should have lined it greatly at fault, but it was the engineman who was responsible.
In September 1944, the C&EI Dixie Flyer collided head-on with a mail train near Terre Haute, with 29 fatalities. At that time, this was single track with automatic block signals, operated by time table and train order. It is usually said to have been caused by "neglect of signals" since this obviously occurred. However, the signals are only a back-up system; two opposing trains are never authorized to occupy the same track, except in error. That such error is not mentioned implies that one of the trains proceeded without authority. An order could have been overlooked, or more likely, a train register misread. There is, of course, an accident report that gives the reasons in this particular case, which I have not read. Train registers are established to give evidence that certain trains have arrived and departed. A train may not proceed unless all superior trains have left, and the register is the evidence for this. Careless register checking, which is performed by the conductor alone, has been the cause of many accidents.
A collision within yard limits occurred on the Big Four at Indianapolis in 1903. A passenger extra collided with a switching movement with the loss of 17 lives, some of them Purdue football players. Yard limits, like sections, are another peculiarly American operational characteristic. Lengths of the main track, usually passing through switching yards, are marked off in which trains may proceed only as the way is seen to be clear. Only scheduled passenger trains are excepted from this, and all other trains must look out for yard movements, and are responsible if they collide with them. In this case, it seemed to be the practice for the yardmaster to inform yard engines when passenger extras were expected, and the extras moved through yard limits at speed. On this day, he failed to notify. Tolerance of such arrangements, which work most days but sometimes fail, may explain the Big Four's shameful safety record.
The conveniences of running sections and establishing yard limits were often unduly stretched to sections hours apart, and yard limits where there was no yard, simply a rural station. No examples of accidents resulting from these practices appears in the 18, but it is worth remarking that the real cause of such incidents is not often realized. There are also hazards when sections and yard limits are combined. How is the engineman of the first section to notify all yard engines, wherever they may be, that a section follows? The yardmasters were usually notified in such cases, but this is as patched-up a solution as in the Big Four arrangement above.
Rear-end collisions have different causes. The worst in Indiana was the grievous collision when a passenger train ran into a circus train stopped to fix a hotbox at Ivanhoe, Ind. in 1918 on the Michigan Central. The circus train was protected by a flagman, and by automatic block signals. The passenger engineman was asleep and not awakened by the torpedoes. He survived, but 60 or 68 others did not. Circus train accidents were pitiable, resulting from the type of equipment and its contents, so their motions became very strictly regulated, with severe speed restrictions and other measures.
Previous to this, the Big Four had suffered a bad rear-end collision in 1907 killing 29. Ignoring signals and bad flagging are mentioned, but no details. It is a little early for automatic block signals, especially on the Big Four, but perhaps it had already begun using the manual block that it had long spurned. Whatever the case, any signals are, in American operation, only supplementary. Primary responsibility is placed on flagging to prevent rear-end collisions. There are cases when flagging cannot be effective, as with short headways, bad weather, and unexpected circumstances. It requires that the engineman be alert, and the rather common circumstance of sleeping because of insufficient rest means this requirement is often absent. Torpedos are essential flagging equipment, to wake up enginemen. There is no evidence that four eyes are any better than two. The fireman, not responsible for the running of the train, was usually busy with the fire or, of course, asleep as well.
One of the two Pennsylvania accidents occurred at Fort Wayne in 1911, when No. 28, the best train on the line, entered a temporary crossover at excessive speed and derailed, killing four crewmen but no passengers. This probably was a case of forgetting a slow order and not seeing the temporary signals protecting the area. The engineman was used to proceeding at speed at the point, and habit won out. When speed must be temporarily limited at certain points, it is usual to issue a train order to the effect, until the restriction appears in bulletin form if it is to last for a while. In addition, fixed signals are placed by the track to mark the location of the restriction. Unless men were actually working at the time, it was not usual to station a flagman.
There were many crossings at grade in Indiana, sometimes protected only by stop boards, and in some cases by gates that could move across one line or the other. The more important crossings sometimes had a "crossing target," a semaphore signal consisting of an arm pivoted at the center that was horizontal to give precedence to one line, and inclined to give precedence to the other, operated by the operator in the nearby station. Lamps were hung at the ends of the arm for the night aspects. This very useful signal was apparently invented on the P. Ft. W. & C. before 1874, and lasted in places (such as Vincennes) until after 1950. Trains still had to approach the crossing at restricted speed, but could usually proceed without stopping if they had precedence. No accidents at such crossings appear in the list of 18.
The one crossing accident that does appear in the list occurred on the Michigan Central at Porter, Ind. in 1921. By this time, the crossing of the two double-track main lines was fully protected by modern interlocking and signals, as well as derails on both lines. The MC train ignored the signals for some reason, was derailed, rerailed again at a trailing switch, and stopped fouling the crossing, where it was immediately run into by a Lake Shore (both lines were actually NYC at the time) train, and 37 were killed. This represents a type of accident that has become more common, in which the actions of the engineman in ignoring signals is simply inexplicable. The testimony is usually that the signal was seen at clear, but this is only a grasp at justification, since usually more than one adverse signal is passed. In the Porter accident, the signal in rear was seen at Approach, but obviously the proper action was not taken. The CSS&SB accident on the gauntlet track in Gary in 1993 was similar in that respect. Simons and Parker say that clearing the signals for the Lake Shore train "set the signals at Stop for the MC." Actually, the signals always stand at Stop, and the derails at derail. Clearing the Lake Shore signals only locked the MC signals and derails in their existing positions.
The final accident for discussion occurred at Walton, Ind., just south of Logansport, in 1947, when the Pennsylvania's Train 207 encountered a large coil of wire on the track that had been put there by country boys having a little fun, and killing 4 passengers. As we have noted, this is the only fatal accident on the Pennsylvania mentioned in the information available. Sabotage and vandalism have caused a number of serious accidents in the United States throughout the years.
We have now briefly discussed the accidents mentioned in the references that occurred in Indiana, which form a representative sample of accidents involving passenger trains. Much more could be said on certain topics, such as flagging, which is by no means as simple as it appears. There seem also to have been few train order accidents. Such accidents were, indeed, rare after standard rules were introduced in 1887 and rapidly spread. Most commentators, especially modern ones, have little knowledge or understanding of train order operation. Signals and block systems were always predominantly backups to timetable and train order operation. Only in recent times, with the spread of CTC, have trains been operated generally by signal indication. An account of the methods of operation and signaling of the Pennsylvania Railroad is located elsewhere on this website, and may provide an explanation for the lack of accidents that we have noticed.
Composed by J. B. Calvert
Created 8 November 2000