How Interurbans Operated

The ICC accident reports provide concrete examples of interurban operating methods


Books on the American electric interurban railways seldom (if ever) mention how the cars operated, aside from the briefest allusions, and the photographs do not show signals or other operating appurtenances, except by chance. We now have a resource, the DOT website that contains historic ICC railroad accident reports, 1911-1966, that includes descriptions of operating methods in the official reports. These, unfortunately, have no photographs and few descriptions of signals, but explain the methods of operation in detail. There was no standard method of operation, but a broad spectrum from the most primitive to careful operation identical to that of steam railways. There were, however, certain widespread features distinct from steam railway practice and peculiar to interurbans. This page will concentrate on these features.

The interurban electric railway typically drew its power from an overhead wire supplying about 600 V DC that was produced by rotary converters fed by high-voltage AC feeders. The equipment consisted of self-propelled cars with two two-axle trucks, where two or four axles were powered by geared, nose-suspendeded, axle-hung series traction motors of 50 to 75 horsepower. Unpowered trailer cars were often used, making a two-car train. Four-motor electric locomotives, or heavy express motors, were used for freight traffic, that became more prominent in later years. Freight trains were usually short, perhaps from four to ten cars, usually without a caboose. Motor cars were equipped with automatic and straight-air brakes, trailers with automatic brakes only. The straight-air brake operated rapidly, making emergency stops in a much shorter distance than those of a steam train, where several seconds elapse before the brakes begin to apply.

Magnetic track brakes were sometimes used. The Public Service Co. of Indiana used 26-ton cars with steel underframes and aluminum bodies equipped with magnetic brakes on its Indianapolis-Louisville line, 117.16 miles, in 1931. These cars could stop from 65-70 mph in 900'. A conventional passenger train would require 4000'. The magnetic brake shoe was pressed on the rail head by magnetic forces, using the traction supply for the exciting coils, and did not depend on the weight of the car and normal adhesion, as wheel brakes do.

Some cars had sanders, which were of most use when braking rapidly on greasy track, though of course sand would aid traction in similar circumstances. Interurban cars were not heavy enough to polish the track as a steam locomotive would, and so were susceptible to leaves on the track and frost. Photographs indicate that many cars did not have sanders.

The motorman's controller cut in starting resistances to limit the current at low speeds, and controlled the series to parallel transition in motor connections, and field shunting. There was only a limited range of running speeds, corresponding to series, parallel and parallel-shunt connections, but this was adequate. For this reason, there were usually no stated speed limits for passenger cars. The control handle had to be pressed down against a spring when the car was running. If it was let up, the power was cut off, and sometimes the brakes were applied. Motormen often taped down the handle for their comfort, negating this safety feature.

A car generally had a two-man crew, motorman and conductor, and there was usually a second conductor for a trailer. Later, single-man cars were common, where the motorman also handled the fares. Freight trains may have added a brakeman, since it was difficult to do switching with only one man on the ground. Larger stations and termini may have had agents, but stops were usually unstaffed and every effort was taken to reduce the number of employees required, since economy was usually necessary.

Also for reasons of economy, most lines were single-track, with frequent turnouts to allow cars to meet. A siding connected at both ends was typically about 500' long, the switches operated manually by the conductor. Single-ended spur tracks used as sidings were quite frequently found, unlike on steam railways where it was much more difficult to back a train. Each stop and siding had a telephone booth for calling the dispatcher, who was located at some central location. This telephone was essential to interurban operation.

Block signals were provided on many lines with relatively fast service between larger towns and cities. These were gradually installed after about 1917, and were quite different from the automatic block signals of steam railways in appearance and operation, since they were almost always electric light signals, not semaphores. In later years, those lines that had become essentially rapid transit operations adopted signalling like that of steam railways, but we will not consider this here, since it was not typical. The use of track circuits was difficult on interurban lines, if for no other reason than stray traction current, which made the use of a return rail and a signal rail very unsatisfactory. AC track circuits could solve this problem very well, of course, but they required such expensive apparatus as impedance bonds, and so were not used on the typical interurban. Switches operated by conductors or the trolley were used instead, or short track circuits at stations.

Interurbans developed from city street railways, where operation was simple due to low speeds, double track and regular schedules. Cars operated visually, at a speed such that they could stop in half the range of vision. Where there was single track with sidings, the meets were regular, and a car would wait for the expected car to arrive. With spring switches, there was no need for the conductor to operate the switch, facilitating one-man operation. Many of the smaller interurbans needed nothing beyond this elementary method.

The traffic density on interurban lines was often very heavy. For example, on the Norman Division of the Oklahoma Railroad in 1944, there were about 114 trains a day. On the Texas Electric between Denison and Dallas, there were 37 trains a day in 1948. On the Shore Line Electric Railway between New Haven and Chester, CT there were 44 trains a day in 1917. This density is much higher than on single-track steam roads, and calls for special methods.

What made these high traffic densities possible were chiefly two things: the regularity of the traffic, and its homogeneity. Most trains stopped at all stations and made the same time along the line. They ran according to a printed schedule where meeting points were prescribed, and usually made by the cars. Running 80 cars a day on a double-track line, such as the Sand Springs Railway at Tulsa, OK is really quite easy and trouble-free. This line also had freight trains, which simply took their turns between the regular passenger cars, which is possible on double track. Without the freights, the 80 schedules could be handled fairly well even on single track.

The usual provision for following trains was that cars were to run no closer together than some minimum distance, say 2000', except in closing up at stations. On the Oklahoma Railroad, a flagman (usually the conductor) had to go back 1000' when his train was unusually detained between stations, a much shorter distance than on steam railways, but still probably thought excessive by crews. It is doubtful that flagging ever took place except in fog or other obscurities. In fact, the Oklahoma Railroad gave oral instructions that it was unnecessary between sections of a train (when one might have thought it most necessary!). Presumably sections would run close enough together to be aware of each other.

Two sections did collide at Berry, OK on 17 June 1944, killing three passengers and a motorman. This stop was in a gap in the block signals, which extended from Norman to Harvest, 4.6 miles, then from Berry to Oklahoma City, 6.4 miles, leaving a gap of 7.4 miles in the middle, at the northern end of which the collision took place. As on steam roads, green signals were carried for a following section, green flags by day and green lights by night. On steam roads, the flags must be displayed night and day, but this was not true on some interurbans. The reason the Standard Code requires flags day and night is that night signals might be put up too early, when they were hard to see.

Dispatching Methods and Signal Systems

There was usually no superiority by direction. When two trains met at a siding, which took siding was generally decided by the crews involved on the basis or convenience. If it was a blind siding, the train that could pull into the siding did so, and backed out after the other was gone. It was not unusual to block a siding with a mobile substation, turning it into a blind siding. A train more than, say, 60 minutes late lost its right to proceed on its schedule, a much shorter time than the 12 hours of a steam railway. When a train stopped at a schedule meeting point and the expected train was not in sight, the conductor telephoned the dispatcher. A train delayed by, say, 5 minutes or more, also phoned the dispatcher, so he generally had both trains on the line either simultaneously or at short intervals.

The dispatcher would then issue orders restricting the delayed train and helping the other train, specifying a meet at some other siding. This might be done verbally, with the conductor repeating the instructions but not writing them down, afterwards carrying out the same procedure with his motorman, who generally was listening to the conductor on the phone. Separate instructions were given to the two trains involved, not in the same words as on a steam railway. This was the so-called "single-order" system that contains many elements of danger. The dispatcher himself might have no order book or train sheet. The positions of trains could be shown by something like a cribbage board, with pegs representing trains. In one case, a dictaphone was used to record the dispatcher's conversations with conductors, but only one side of the conversation was recorded, and that badly, considering the dictaphones of the time. This system was much like anyone today would devise, and does not contain the important safety provisions of the Standard Code.

Since written instructions are safer than verbal, the conductors might be required to write down the order in manifold (2 copies) and deliver one copy to the motorman. Union Traction of Indiana used a machine that produced two copies with double-sided carbon paper. After the order was written, repeated and signed, a crank was turned that issued the two copies, the carbon remaining in the machine. When such care was taken, the dispatcher also now had an order book in which he underlined words and numbers as they were repeated, and a train sheet to record the position of trains. The order system, however, was still the faulty "single-order" system. By 1948, Texas Electric used Standard Code methods with the order issued in the same words to the trains involved in the movement.

What was needed in a signal system was some way of preventing opposing trains entering the single track between two sidings. A very simple way of doing this is shown at the right. It's simply the same circuit as is used for hallway or stair lights that can be turned off and on from two points, here sidings A and B. I do not know if this exact circuit was actually used, since information on exactly how the signals worked is very rare, but it is such a simple and obvious plan that it would not be surprising to find it. A possible example were the box signals used on the Evansville and Ohio Valley in the 1920's. The power could be 110V AC wires, or even the traction power. In the latter case, there would be sufficient voltage for three ordinary lamps at each end, which could be arranged in a horizontal row to give a position-light effect, or at three chosen locations.

The correct way to use this system was for the conductor to observe that the lamps were in the normal, unlighted state, which meant that the block was not in use. He would then throw the switch, whichever way it happened to be standing, and note that the lights instantly became illuminated. When the lights go on at his end of the block, it is certain that they are lighted at the other end, since they are in series. If a lamp is burned out or a wire is broken, the lamps will not go on. Seeing this, his car can proceed through the block. When he reaches the other end, the switch there is operated to turn the lamps off. If two trains try to obtain the block at about the same time, both seeing the lamps off, the lights will only flash briefly, then go dark again. In this case, one of the trains (specified in the time table) will try again, while the other waits. One section can follow another if the preceding car is within view; otherwise, it should wait until the block is clear again.

Some Examples

The report on the Clay, IN accident on the E&OV on 7 September 1927 mentions block signals consisting of red lights and switch handles in boxes that were mounted on the overhead poles. There was a block southward from Clay to Dade Park, 3.5 miles, and another north from Clay to Kentucky Avenue, on the line from Evansville to Henderson that included a ferry across the Ohio River on what had been an Illinois Central line. The signals were operated by a switchtender in times of heavy traffic. There seems to have been a man at Kentucky Avenue to work the signals there that divided the 3.05 miles to the Evansville terminus on 5th Street. This accident occurred when car 117 backed into the siding to meet a northbound car, then proceeded to Dade Park. The switchtender did not close the switch, and on the return journey the car entered the spur at 40 mph and collided with a freight car. Three passengers and the motorman-conductor were killed. These signals were called "automatic" in the accident report, but apparently were not so. The switch was not connected to the signals. In fact, the signal boxes obscured the switchstand lamp, contributing to the accident.

There was a serious accident on the Shore Line Electric Railway at North Branford, CT on 13 August 1917 caused principally by the primitive operating conditions and poor discipline on this line, which ran from New Haven to Chester, 47.33 miles. There were 22 regular trains each day in each direction, on a regular "hourly" schedule. As an economy measure, after 1914 no printed time table was prepared. Instead, a pen-written time table was posted at the principal stations, where the meeting times were written in red ink. The motormen and conductors were supposed to memorize the meets for their runs, which indeed were quite regular. In cases of delay, conductors called the dispatcher and were given verbal orders, which they relayed to their motormen. The dispatcher kept a written record of his orders. Extra trains were run without orders, like regular trains. On this day, motorman Negus and conductor Tryon were running a "half-hour" extra car. They had a meet with a regular hourly car at North Branford, but motorman Negus made a station stop and left, apparently thinking he was running an hourly car, which had no meet here. If you went through the turnout at more than 5 mph, you would lose the pole, so Negus must have been in control at this point. Conductor Tryon was sleep-deprived because of the long hours of duty at the time, and he fell asleep as they left North Branford.

Motorman Negus said he suffered from fits that had lost him other jobs, and was "unconscious" leaving North Branford. The ICC report says "he came to his sewage only an instant before the collision" with the regular car. He had tied down the controller handle for his convenience, and it was not connected to the brakes anyway. Company officials said that when it had been, the motormen used it to make stops, which jarred the passengers, so it was disconnected. The bang woke up Tryon, and the two survived but 19 passengers did not. There was no communication between Negus and Tryon over the meet, as there should have been.

The recommendations made by the ICC investigator were for (a) physical examinations of employees, (b) written orders to coductors and motormen, (c) train sheets for the dispatcher, and (d) printed time tables distributed to motormen and conductors, all excellent ideas.

The South-West Missouri Railroad was a 39-mile electric line running between Carthage, MO and Picher, OK in the Tri-State lead and zinc mining area. This line illustrates what was probably a very common mode of operation, with verbal orders and no train-order book. Instead of a train sheet, a peg board in the dispatcher's office sufficed. The only record was a dictaphone recording of the dispatcher's conversations. In spite of this, the only accident on record happened on 3 May 1924. Extras were designated by the car number. Car 36 departed Webb City at 3.20 with orders to run to Blendville, the end of city double track. There it received a verbal order: "Meet 69 at Tanyard, 61 at Riceville and go to Lowell," according to the dispatcher. Conductor Surber said the order ended by "go to Baxer" instead, a station beyond Lowell. Car 36 departed Blendville at 3.55 pm.

Car 65 departed Picher at 3.55 pm with orders to go to Military, the second station west of Lowell. At Military, it was told: "Meet No. 8, meet two cars at Baxter and go to Lowell." Apparently, car 65 was to wait at Military for No. 8, then go to Baxter. It seems to have met the two cars at Baxter, and was proceeding to Lowell when it encountered car 36 unexpectedly at 4.30 pm. Two passengers were killed, there was a fire, and car 36 was consumed.

The motorman and conductor of Car 36 supported each other that they had been given instructions to go to Baxter. The motorman probably had, by the conductor, who had misinterpreted the order. The motorman was supposed to listen in when an order was being received, but motorman Crump took the opportunity to visit a nearby drug store for refreshments. The dictaphone was unintelligible. The lack of written orders was a serious deficiency, but this example shows that it was still permitted as late as 1924. The elementary method of dispatching exhibited here has only the benefits of centralized control and direct communication between dispatcher and crews, with none of the safeguards of the Standard Code.

The Union Traction Company of Indiana

The Union Traction Company of Indiana suffered from accidents on 22 March 1912 at Fortville, on 13 June 1919 at Carmel, on 2 February 1924 at Ingalls, and on 16 November 1926 at Bell. These accidents killed 0, 2, about 16 and 0 passengers, respectively, and all were head-end collisions due to motormen's passing meeting points. The investigations of this series of accidents furnishes a good view of operations on a typical midwestern interurban line.

Fortville was on the Indianapolis-Muncie (Anderson) Division. In 1912 there were no block signals. The dispatcher could control signals at certain locations to stop trains for orders. Cars carried a portable telephone that could be plugged in at sidings, and forms for writing orders in duplicate. Both the motorman and the conductor repeated the orders to the dispatcher, who gave them OK. The conductor took the original, the motorman the duplicate, which was turned in at the terminal. At certain stations there was a permanent telephone booth and an order box where orders were copied in triplicate, the lowest copy remaining in the box.

Train No. 61, combine car 279, left Indianapolis on time at 7.05 pm. Train No. 60, combine car 379, left Muncie at 6.35 pm. The trains met by schedule at Siding 19, 20 miles from Indianapolis and 39-1/2 miles from Muncie, at 7.52 pm. Train No. 61 arrived at Siding 19 at 7.56 pm, but Train No. 60 was not in sight, so the dispatcher was phoned. He issued order No. 89: "Train No 61 car No 279 and Train No 60 car No 290 will meet at Pendleton." [Car numbers are scrambled in the digitized report, as well at Pendleton/Pendiston, so I will simply be consistent.] The dispatcher's record showed order No. 89 to be: "Train No. 61 car No 279 and Train No Extra West car No 290 will meet at Pendleton." This order was made complete at 6.58 pm, and destroyed in the fire following the collison. Pendleton is 10 miles east of Siding 19. The Extra West received this order at Siding 13, 1-1/2 miles east of Pendleton, at 8.09 pm.

When Train No. 61 received order No. 89, it was already at its meeting point with Train No. 60, which was closely approaching. By some confusion, the order relating to the Extra West was taken as concerning No. 60 by the motorman and conductor of Train No. 61, and they left Siding 19 immediately after the order was complete, and two minutes later collided with Train No. 60, 2700' east of Siding 19 and 1-1/2 mile west of Fortville. The cars telescoped each other by ten feet and caught fire, though not derailed. The motorman of No. 61 died in the fire.

The dispatcher and the conductor of Train No. 61 disputed the contents of order No. 89, but the dispatcher was evidently correct, and the conductor trying to explain away his responsibility by throwing the blame elsewhere, a very common thing after such accidents. We notice the use of "single orders" and the helping of trains before opposing trains are restricted. A train needing help gets it immediately, while the restriction comes later when convenient. Under the Standard Code, the order is given in the same words to both trains, and the superior train is restricted before the inferior train is helped.

Carmel was on the Indianapolis-Logansport (Tipton) Division, where the Monon was crossed at grade. Train No. 331 was bound for Indianapolis from Logansport. It was 23 minutes late at Tipton, leaving at 5.23 pm. At Miller, an order to meet Train No. 338 at Hawk was received. At Hawk, order No. 47 was received at 7.05 pm: "Train No 331 car No 265 and Train No 338 car No 269 will meet at Grove instead of Hawk." Grove was the schedule meeting point for Trains 331 and 338.

Train No. 338 left Indianapolis for Tipton at 6 pm. and arrived at Ripple, 7.7 miles north, at 6.43 pm. It received order No. 46, at 6.43 pm, as follows: "Train No 338 car No 269 and Train No I.S. [sic, an error in digitizing] car No 297 will meet at Akers. Train No 338 car No 269 and train No 335 car unknown will meet at Nobles. Disregard block at red and bad order Ripple to Tipton." There was no mention of Train No. 331 in this order, since the time table meeting point governed. The dispatcher had issued orders to No. 331 that he could use to help No. 338, but this turned out not to be necessary, and so such orders were not issued. Here we see the "single-order" system in all its glory. Under the Standard Code, both trains would have received order No. 47 and the meet would have been emphasized. Note that order No. 47 is later than order No. 46, and the dispatcher simply moved Train 331 forward in the later order.

Train 338 made the meet at Akers, then simply forgot about Train 331 and passed Grove without stopping. The cars collided 2558' beyond the Monon crossing at St. Clair Stop, at 7.10 pm. Car 269 was telescoped about 20' as car 265 overshot the end of car 269. The relative speeds were 5 and 15 mph.

Motorman Smith and conductor Orr thought fast about how to cover their error. The "335" on order No. 46 was altered to "331" on their copies, and they claimed that the dispatcher had changed the number as the order was being written, and each backed up the other's story. The dispatcher had a different view of things, and said that he had never transmitted anything but "335" and it had been repeated by Orr. What showed Orr and Smith to be villains were the altered orders they produced as evidence. The "1" was indeed scratched on the front, but the back, where the two-sided carbon had made its impression, there was only a "5" so the order had to have been altered after removal from the order machine. It is a rule that an order with erasures, interlineations or corrections must not be acted upon, but a new order obtained.

This line had been equipped with block signals (we are, unfortunately, not given any details) in 1917, but they were so poorly maintained as to be often out of service. A photograph of a signal from the Ingalls ICC accident report is shown at the right. Orders were issued to disregard signals at red, and those that had been taken out of service ("bad order"), as in order No. 46. These bad signals should have been passed with caution, but this was obviously ignored. On this day, an electrical storm had damaged the signals, the dispatcher said. There were 59 orders to disregard signals that day. The ICC inspectors made trips on 17, 18, 25, 27 and 30 June, and on all these days orders were given to disregard signals, although there were no storms. The single signal maintainer was responsible for 122 signals on 99 miles of track, in three sections.

The next accident was at Ingalls on the Anderson Division, between Indianapolis and Muncie, 56.52 miles. It was the most costly of these accidents in passenger fatalities, and is the only one mentioned in Shaw. The exact number of fatalities is not known, because of the subsequent fire. The block signals were not in operation, since they were taken out of service for overhaul between 27 December 1923 and 6 February 1924. As at Fortville, the cause was confusion over the wording of an order. The trains involved were eastbound No. 24, with motor car 286 and trailer 305, and westbound No. 21, with motor car 276 and trailer 306. The stations involved are, from west to east, Lawrence, Mur, Dent, Ingalls, Goul and Pendleton. "Mur" is also "Min" in the report, and neither sounds much like an Indiana name.

The meet between No. 24 and No. 21 was scheduled at Mur, while Train No. 23 was following Train No. 21. No. 24 left Indianapolis at 3.17 pm, 17 minutes late, and arrived at Lawrence, where it met local train No. 917. The crew of No. 917 had copied, repeated and obtained an OK for an order for delivery to No. 24. This order, No. 45, stated: "Train No 24 car No 286 will meet Train No. 23 car No 275 at Goul." It was understood improperly by the conductor and motorman of No. 24 to refer to No. 21 instead of No. 23, either because the order was written down incorrectly or it was misread. No. 24 left Lawrence at 3.50 pm. At 4.03 pm, No. 21 received order No. 46 at Pendleton, saying: "Train No. 21 car 276 will meet Train No. 24 car No. 286 at Dent." The dispatcher intended to help No. 24 with this order at Mur, when it telephoned in on not seeing No. 21 there. However, No. 24 did not stop at Mur on its way to Goul, and passed Dent about 4.08 pm. The cars came together a short distance west of Ingalls in cloudy weather at 4.12 pm, on a curve with poor visibility. The wreckage caught fire, and both motor cars were consumed. The order in question was not found in the conductor's grip.

The investigators again found the block signals poorly maintained, and, of course, out of service when this accident occurred. Rules for the block signals were not even in the rule book. When they were installed, a pamphlet with these orders was published that was supposed to be pasted in the back of the rule book. The issuance of orders to disregard stop aspects was still common. The time table even showed two meeting points between one pair of trains. The practices of taking orders for another crew, registering trains for another crew, and signing the names of crew members to which the orders were to be delivered, were also criticized. The Standard Code is very strict about orders delivered by another crew, and provides that they cannot be assumed to have been delivered until the signatures of the receiving crew have been transmitted to the dispatcher.

The accident at Bell, a siding 1/2 mile south of Poneto, was on the line from Muncie northward to Bluffton, 42 miles. This track was used jointly by Union Traction and Indiana Public Service Co. trains. The accident, a collision between two locomotive-hauled freight trains, each with three cars, occurred early in the morning in a light rain. The siding at Bell was a through siding, but on this day a portable substation had been left at the north end of the siding, converting it into a blind siding. Train order No. 2 established a meet between southbound UT Extra 604 and northbound IPS Extra 49 at Bell. Extra 604 received the order at Bluffton, 7 miles north of Bell, and was prepared for the meet. Because of the substation, it was assumed that Extra 49 would head in.

Extra 49 received order No. 2 at Hartford City. This might indicate that the "single order" system had been abandoned by this time. The conductor copied and repeated the order, then read it to the motorman (not the best practice) and gave him a copy. The motorman said that he saw Extra 604's headlight as he approached Bell, and then fell asleep. Conductor Hudson said he thought about the meet order, then went to shuffle his waybills. Their brakeman had not been shown the order. The collision occurred at 5.12 am, 346' north of the north switch, with Extra 49 still working power. Fire broke out afterwards. The motorman said Hudson frequently failed to remind about meets. Hudson said this was the first time he had not done so.

Key System

The Key System operated a network of streecar lines in the East Bay region that converged on a double-track trunk line that extended 3.85 miles west of San Pablo Avenue on a fill to the Key System Pier Terminal, from where a ferry connected to Market Street, San Francisco. This line also carried the cars of the San Francisco-Sacramento Railroad from 40th and Shafter to the Pier Terminal. Street cars were coupled to form trains of from 1 to 8 cars for operation over this line. Poles supporting the catenary were spaced at 105' between the tracks, and every fourth pole had a semaphore block signal for each direction. One caution and two stop aspects were provided behind a train. This system was designed for a 45-second headway with cars running 36 mph that could stop within 300' to 350'.

One of the signal pairs is shown in the photograph at the right. This is the signal near which the rear-end collision of 4 December 1924 occurred. You will notice two rods, one with each signal and below it. One rod is extended over the track, while the other is retracted, in this photo. When the signal displays the Stop aspect, the rod is extended and can come into contact with a valve on the roof of a car that passes it, applying the brakes in emergency.

On the day in question, Key System train No. 729 from the 12th Street line had four center-entrance cars. It was followed by SF-S train No. 15 with car No. 1014, which ran from Concord to Oakland. The rear car had been dropped at 40th Street, when the train entered Key System tracks. Train No. 729 stopped with its rear just past Signal 104, with its front in the block approaching Signal 100. The motorman did not want to release his brakes to stop more closely in case he could not stop short of Signal 100. He whistled out a flag, but before anyone could go back, the train was run into by car 1014 at about 20 mph. Car 1014 rose up and telescoped the lighter Key System equipment, killing 8 passengers and 2 employees off duty. 38 people were injured.

It was established that the automatic train stop had operated properly at signals 104 and 108, both of which were at stop. Signal 112 was at caution when No. 15 passed it. Train No. 729 had stopped at the most unfavorable location, so the available stopping distance was 420'. The SF-S cars had a maximum speed of 55 mph, instead of the 36 mph of the Key System cars. Their stopping distance from 55 mph can be estimated to be about 756'. The motorman of Train No. 15 apparently took no action at the Approach aspect, so the brakes were applied in emergency by the automatic apparatus at Signal 108, but there was insufficient stopping distance. There was no prescribed speed limit on the line, and the investigator recommended that one be established.

This example illustrates short-headway operation on double track, which is characteristic of urban mass transit systems, if not of the typical interurban. These signals were installed in 1911, and at the time of the accident, had experienced only 4 false clear failures, tbe last in 1916. There were 18 million 45° movements in a typical year.

The Key System consisted of the Key Terminal Railway, Ltd. (9 miles), the Key System Ltd. (57 miles) and the East Bay Street Railways, Ltd. (180 miles), three interlocking companies. The reason for the "Ltd." designation I do not know.

Pacific Electric

Los Angeles also had a heavily-used electric interurban railway, the Pacific Electric, which advertised itself as "The Largest Electric Interurban System in the World." Its lines were double and single track, generally without block signals, on private right-of-way, extending as far as, for example, San Bernardino, 58 miles, and Santa Ana. Accidents reportable to the ICC occurred in 1913, 1925, 1945, 1950 and 1953. They were rear-end collisions, except for the 1945 Athens, CA head-end collision on the San Bernardino line involving a freight motor that had stalled on a grade when its sanders became plugged.

None of the accidents had a significant casualty toll, except for the first one, on the Venice Short Line in the evening of 15 July 1913, which killed 14 and injured 200. The Venice Short Line was a double-track line extending from the junction at Vineyard westward 8.5 miles to Venice. It was 5.5 miles from Vineyard to Hill Street Station in the center of Los Angeles, in a generally southwestern direction. Cars from North Beach, Santa Monica, Venice Beach and Redondo Beach, joining at Ivy Junction, 3.6 miles west of Vineyard all used this line on their way to Hill Street. Traffic to the beaches could be very heavy in the summer. On Sunday, 13 July, there were 123 trains between 9 am and 9.20 pm, so the average headway was only 6 minutes. Typical trains were composed of three cars running in MU. In 1931, 54 trains daily in each direction were advertised, running from 3.45 am to 2.10 am.

This line was operated without signals or train orders. Starters at Venice and Hill Street spaced the trains, with extra trains running between the scheduled trains. When a train stopped unexpectedly, the motorman whistled out a flagman, who went back 600' to 700' with stop signals. In a test, a typical train could stop in 400' from a speed of 40 mph. It seemed that cars could run at 50 mph, but there was a restriction of 10 mph at junctions. Speeds of 30-35 mph were probably typical.

On the evening in question, the trolley wire had broken 50' east of Vineyard, and this had caused a backup. The wire had been picked up and secured so that it was not dangerous, but cars had to get up speed and coast across the gap. Since this was on a 1.7% rising gradient, it took a bit of a run for it. Extra 532 had left Venice at 8.55 pm and had come to a stop at the end of the queue at Vineyard, which was on a curve. They had encountered a flagman from the preceding train and picked him up, whistling out a flag of their own. Extra 874 had left Santa Monica at 8.53 pm, required 12 minutes to reach Venice, and then proceeded. Extra 874 went by a sign saying "Slow" at 30 to 35 mph instead of the prescribed 10 mph approaching the junction, and by the surprised flagman of Extra 532, with its brakes in emergency, and rammed the last car of Extra 532.

Extra 874 was certainly going too fast for the location; perhaps the flagman was expecting the train to approach the junction more slowly. If this had not been at a junction, however, the usual line speed would be appropriate, and it would be concluded that the flagging was insufficient. Since this is the only accident recorded for this line, we must conclude that the methods were generally adequate and only failed in these unusual circumstances. It is interesting to compare this with the Key System line, with equally heavy traffic and the same speeds. However, the Key System line was plagued with fogs, which the Venice Short Line was not.

Denver and Intermountain

The only reportable accident on the Denver Tramway system occurred 1 March 1943 on the Denver and Northwestern line to Leyden, CO when a southbound coal train with two motors collided with a northbound passenger car on Route 82 at Swadley station, killing the motorman and a passenger in the passenger car. This line left Denver at its northwest corner, at Lakeside, past Clear Creek Junction, where the line to Golden diverged, and Arvada Junction, where there was interchange with the Colorado and Southern, and dual-gauge track began that ran just south of the Denver and Salt Lake Railway (as it was then) and straight ahead to the mines between Leyden and Lindsay, now gone. The gauges were the 3' 6" of the Denver Tramway system and standard gauge. The Denver and Intermountain (D&IM) operated the freight trains, with narrow gauge motors and cars, while standard gauge cars could also be handled for the C&S. The coal mines at Leyden were the source of much heating coal for Denver, and the fuel source for the Denver Tramway's power plant west of the South Platte in central Denver. The mines and railway operated until 1950. Pictures of the equipment mentioned here, specifications, history and maps can be found in Jones, Wagner and McKeever, Mile-High Trolleys (Golden, CO: NRHS, 1965).

The D&IM, which also operated the standard-gauge line to Golden that served the Federal Center as well, claimed to operate under the Interurban Rule Book, which would be very interesting to study. The ICC report quotes two rules. Rule 107 is "(a) No car or train will leave any initial station, a siding, or enter the main line without orders from the Dispatcher. (b) When any train reaches a meeting point and finds that the train or trains to be met have not arrived, the motorman shall immediately call the Dispatcher for orders." Orders were apparently written down by the motorman and repeated.

Train 125 consisted of motor car .07 and left Denver at 12.30 pm for Leyden. At Lakeside, the motorman copied order No. 53, reading in part: "Number 125 Motor 07 schedule meet train No. Exa-East 1110-1111 at Swadley," and left at 12.54 pm. Swadley was 3.52 miles west of Lakeside, and there was a 910' siding on the south side. Train 125 was due to leave Swadley at 1.09 pm, but did not stop there but sailed through to collide at 20 mph with the coal train west of the station on a 3° curve and ascending 1.5% grade. There is no explanation for the motorman's failure to stop and wait for Extra 1110-1111.

Engines 1110-1111, which were steeple-cab narrow-gauge electric locomotives with 4 70 hp motors each and weighing 30 tons, had 14 cars and caboose. The conductor copied order No. 52 at Leyden, 6.22 miles west of Swadley, at 12.52 pm, to the effect: "Extra Frt Engine 1110-1111 Exa East Ley to Lake Report Arvada Meet train No 125 Motor 07 at Swadley." We note that single-order dispatching is in force, with each train only told what it needs to know. This train left at once, and collided with No. 125 0.43 mile west of the west switch at Swadley at 1.08 pm, in falling snow, while moving at 10 mph. Car .07 was demolished (Mile-High Trolleys says 1942, but it was 1943).

A time table for the standard-gauge Golden line that went into effect the day of the Swadley accident shows hourly trains in each direction, taking 48 minutes for the 13.1 miles and 15 stops, always meeting at Devinney, 5 miles from Denver, where there was a spring switch at the west end. Eastward trains were superior to westward, and there were 8 telephones, including those at the termini. The speed limit for extra trains was 20 mph.


The ICC reports have given us concrete examples of interurban operation practice that come together to form a general picture that is probably not far from correct. It is remarkable that there were so few serious accidents due to all causes. From some of the operating practices, we should certainly expect more, but they did not happen. The motor cars that replaced interurbans are many, many times more dangerous, but the grievous toll seems to excite very little public outcry, and continues unabated.

A tentative conclusion that can be drawn from the accident reports is that operation with one man in the cab, as on interurbans, was not significantly more hazardous than operation with two men, as on steam railways. There does, however, seem to be a greater hazard when there are not two independent checks on the interpretation of written orders, as between engineman and conductor at meeting points. When operation is by signal indication, this advantage disappears.

It is also clear that a possible excellent, environmentally friendly personal transportation system has been consigned to oblivion. In 1931, one could travel between Indianapolis and Louisville in steel and aluminum electric cars at 70 mph. There is an Amtrak train now that lurches and rocks along the weedy track of a parallel line at 30 mph a few nights a week.


The ICC accident reports can be found at DOT Website. We are very grateful for these documents, but the digitization creates some odd statements. For example, in one report "Muncie" appears as "Humeis." Also, the year 1960 appears to be absent.

R. B. Shaw, Down Brakes (London: P. R. MacMillan, 1961). Considers only the accidents with numerous passenger casualties. The 1913 accident at Vallejo, CA on the San Francisco, Napa and Calistoga is not found in the ICC reports.

Return to Railway Index

Composed by J. B. Calvert
Created 4 March 2002
Last revised 27 March 2002