Link & Pin: Symbol of Dangerous Yesterdays

Before 1900, joining two railroad cars meant stepping between them while they were still moving, holding an iron loop steady with your hands, and dropping a pin into place at exactly the right moment. The mechanism was called the link-and-pin coupler. Railroaders called it something shorter: the widow-maker.

How It Worked

The link-and-pin coupler was about as simple as a machine can get. Each car carried a drawbar — a heavy iron bar, typically five inches by five inches in cross-section, projecting from the car’s end beam. The business end of the drawbar was a hollow pocket, or drawhead, with a hole bored vertically through the top and bottom.

To couple two cars, a brakeman held an oval iron link, much like an oversized chain link, and guided it into the pocket of the approaching car. When the link seated, he dropped an iron pin through the hole. The pin passed through the link and locked the two cars together. To uncouple, he pulled the pin back out.

That was the whole mechanism. It was cheap to make, easy to repair with a forge and a hammer, and it worked well enough that American railroads used it for roughly seventy years.²

The parts, and what went wrong with each

• The link — Came straight, bent, or offset. An offset link was needed whenever two cars’ drawbars sat at different heights above the rail, which was often.
• The pin — Came solid-head, eye-head-flat, and bent-head. None of these were interchangeable between railroads.
• The drawbar — Standardized at 28 inches long and 33 inches above the top of the rail by the Master Car Builders’ Association in 1879, with a three-inch allowance between loaded and empty cars. That allowance meant a loaded car and an empty car could still present couplers three inches out of line.
• The deadwood buffers — Heavy blocks flanking the coupler, meant to absorb shock. Their spacing was never standardized. An 1879 survey of five railroads found the gap between buffers ranged from 12½ inches on one line to 28 inches on another — meaning the space a brakeman had to put his body into varied wildly from car to car.

The deeper problem was that none of it matched. The U.S. Patent Office issued more than ten thousand patents for link-and-pin variations, and by some accounts over a thousand distinct designs saw actual service.⁴

An hour to couple a single car

Because nothing was standardized, every coupling was a guessing game, and crews solved it the only way they could — by hoarding. A brakeman working a way freight would size up the two drawbars, then signal his partner on the head end for what he thought he needed: one finger raised for a round pin, two fingers side by side for a flat pin, both hands opening and closing in a circle for a link. The front man would dig through the box and bring back the hardware. Often none of it fit. The round pin was too fat for the hole, the flat pin was no better, and the link seated in one drawbar but was too wide for the other.⁴

At that point the brakeman walked back to the caboose for the crew’s private stash — the accumulated collection of links and pins kept aboard for exactly this situation. As the Railroad Magazine account put it, after “maybe an hour’s fuming, by the trial and error method, you get the coupling made.” That is one car. If the two drawbars also sat at different heights, he needed an offset link on top of everything else.⁴

Those stashes were rarely acquired honestly. Crews stripped loose links and pins from any car standing unattended on a siding, and the practice was widespread enough that railroads tried to engineer around it — chaining pins to their drawbars so they could not be carried off. That backfired badly: a tethered pin would drop back into its groove on the first jolt, forcing the brakeman to step between the cars again, sometimes three or four times, to pull it out. The cost was real money. One estimate put replacement couplings at roughly $60,000 a year for an average Class 1 railroad, and the magazine records a general manager on a large eastern road who won his stockholders’ admiration by saving $10,000 annually through the simple method of refusing every requisition for links and pins — which worked only because his night crews quietly raided a connecting railroad’s yard to make up the difference.⁴

What It Cost

The link-and-pin required a man to place his hands, and much of his body, between two pieces of rolling equipment weighing many tons, while they were in motion. The margin for error was measured in inches and fractions of a second.

• 300railroad workers killed in coupler accidents in 1888 alone¹
• 6,700+injured in coupler accidents that same year¹
• ~1 in 3of all railway accidents traced to coupling and uncoupling, per the ICC’s 1890 annual report⁴
• 2ndmost dangerous occupation in 1880s America, behind only coal mining⁵

Those are the numbers that survive. They almost certainly undercount. Reporting was voluntary and inconsistent before the Interstate Commerce Commission began collecting accident statistics in 1889, and a lost finger or a crushed hand frequently never appeared in any ledger at all.

It is worth setting the coupler against the whole picture. By 1907, American railroads had become the single largest cause of violent death in the United States, killing nearly twelve thousand passengers, workers, and bystanders in that year alone.⁶ Coupling was one hazard among many in an industry that was, by modern standards, extraordinarily dangerous to work for and to live near.

Mangled hands and missing fingers were the badge of the experienced worker. Standing between cars as they were being pushed together and holding the iron link up to guide it into place, working around the “dead wood” buffers that kept the cars from telescoping together, the switchman would only need a moment of lapsed attentiveness to lose a finger, a hand, or his life.
— Linda Hall Library, The Transcontinental Railroad: Couplers & Brakes²

The injury was common enough to be a form of professional identification. A veteran railroader could be recognized by his hands. A. F. Whitney, who rose to become president of the Brotherhood of Railroad Trainmen, carried the mark of it — he had lost the ends of two fingers coupling cars as a young man, and spoke about it publicly for the rest of his career.⁴

The coupling stick that nobody used

Some railroads issued brakemen a coupling stick: a wooden paddle with a notch cut into it, designed to guide the link into the pocket from a safe distance, carried in a leather sheath on the belt. Rule books mandated its use. A Wisconsin Central rulebook from 1878 devoted a full paragraph to it, instructing that “coupling by hand is strictly prohibited in all cases where a stick can be used.”⁴

It was widely ignored. Using the stick was slower, and on a railroad that paid by the trip and measured men by how fast they worked, slower was its own kind of hazard. Veteran crews mocked the men who used them. The prevailing attitude, as one account put it, was that fingers were made before coupling sticks.

The Long Road to Something Better

The link-and-pin’s failings were understood almost from the beginning. Fixing them took roughly forty years, and the delay had less to do with engineering than with the difficulty of getting hundreds of competing companies to agree on anything.

The interchange problem

By 1890, between 17 and 45 percent of a major railroad’s cars were “off-line” at any given moment — running on somebody else’s track.¹ A railroad that installed a better coupler on its own fleet gained little, because its cars still had to couple to everyone else’s. As one railway economist wrote in 1887, “owing to the continuous interchange of cars, no real benefit would be derived from such a coupler until it had come into almost universal use.”¹ Everyone had a reason to wait for someone else to go first.

Passenger service leads

Progress started where interchange didn’t apply. Passenger cars stayed on their home railroad, so any company could modernize them freely — and passengers, unlike freight, complained and sued. Ezra Miller patented his “Miller hook” in 1863; by 1875 it was in passenger service on 574 railroads.¹ Paired with the Miller platform and buffer, it pulled car ends tight enough that passengers could walk safely between cars, and it reduced the telescoping that made 19th-century wrecks so lethal. It was judged too expensive for freight cars, which is where most of the dying was happening.

Eli Janney’s knuckle

Eli Hamilton Janney was a Confederate veteran working as a dry-goods clerk in Alexandria, Virginia after the war. He was also a skilled whittler, and he carved his solution out of wood: a coupler shaped like a human hand with the fingers curled, which would grasp an identical hand reaching back. Two cars pushed together and the knuckles closed on their own. No one had to stand between them.

Janney patented it on April 29, 1873 (U.S. Patent No. 138,405), with strength improvements following in 1879 and 1882. His patent claimed the design would “couple readily under all circumstances if one of the hooks is open, but will not couple if both are closed,” and that it was “adapted for use on cars of different heights” — solving the drawbar-height problem that had forced crews to hunt for offset links.³

Forty-two couplers in a Buffalo rail yard

In September 1885, the Master Car Builders’ Association staged what one historian called the Olympics of car couplers: a three-day trial in an Erie Railroad freight yard in Buffalo, New York, testing 42 competing designs.¹ Each entry had to couple on straight track and on twenty-degree curves, couple with existing link-and-pin drawbars, and refuse to couple when told not to. Couplers snapped in switching. One broke a draft bolt. One uncoupled itself repeatedly on a curve. One coupled so well it took a pinch bar to get it apart again.

Shortlists that couldn’t couple

The hard part was never inventing a better coupler. It was getting everyone to pick the same better coupler — and for two decades, nearly every attempt to narrow the field produced a shortlist whose own members would not connect to each other.

State legislatures tried first, and made things worse. Connecticut passed an automatic-coupler law in 1881. Massachusetts followed and approved five varieties; railroads operating in the state tried all five, only to find that not one of the five would couple with the other four. Michigan approved seven, and none of those seven would couple with each other either. By 1889, thirty-nine different coupler types were in revenue service across 8,510 cars, and each of the thirty-nine refused to connect with any of the remaining thirty-eight.⁴

The Master Car Builders’ Association, which was actually trying to set a national standard, moved more carefully but ran into the same problem. After the Buffalo trials, its executive committee deliberated for several hours and advanced twelve designs for further service testing — six vertical-plane couplers and six of the link type. Only in 1887 did the field finally narrow to three: the Janney, the Dowling, and the Thurmond, all vertical-plane designs. The committee leaned toward Janney because the contact between the curved faces of its vertical cylinders held up best on both curves and straight track, and because it tolerated variation in coupler height between cars.⁴

One obstacle remained. The MCB favored Janney’s design but had a rule against endorsing patented technology. The deadlock broke in 1888, when Janney’s patent holders agreed to waive rights to the contour of the knuckle — the profile that lets couplers from different makers mate. That single concession is what made a national standard possible.¹

Even then, adoption crawled. In 1890, MCB couplers were installed on less than 14 percent of the national freight fleet.¹

Lorenzo Coffin

What broke the logjam was a Congregationalist minister named Lorenzo S. Coffin. Serving as an Iowa railroad commissioner, Coffin rode in a freight caboose one day and watched a brakeman he had befriended slip while making a coupling and lose his hand. He spent the rest of his life on the problem — riding freights, writing to newspapers, and lobbying Congress for six years. Many dismissed him as a crank.⁵

On March 2, 1893, President Benjamin Harrison signed the Railroad Safety Appliance Act. It required that every car in interstate service be equipped with couplers “coupling automatically by impact, and which can be uncoupled without the necessity of men going between the ends of the cars.” The carriers asked for more time, and Congress extended the compliance deadline to 1900.⁵

Between 1890 and 1909, the rate of coupling accidents on American railroads fell by half.¹

Seeing One Up Close

The pieces in this exhibit are the actual hardware — links, pins, and a drawhead of the type that was standard equipment on American railroads for the better part of a century. They are heavier than most visitors expect. That weight is the point: it is what a brakeman was lifting and steadying by hand, in the dark, in the rain, between two moving cars.

Plan Your Visit

Sources & Further Reading

1. Bethanne Knudson and the ASME History and Heritage Committee, “The Janney Coupler” (ASME Historic Mechanical Engineering Landmark No. 267). Casualty figures for 1888 and the 1890–1909 accident-rate decline are drawn from Mark Aldrich, Death Rode the Rails, pp. 111 and 114, as cited in this document.
2. Linda Hall Library, “Couplers & Brakes,” The Transcontinental Railroad digital exhibition.
3. Eli H. Janney, “Improvement in Car-Couplings,” U.S. Patent No. 138,405, April 29, 1873. Full text and drawings reproduced at the Central Pacific Railroad Photographic History Museum.
4. “Link and Pin: Symbol of Dangerous Yesterdays,” Railroad Magazine, May 1941, pp. 8–21. Original issue in the ORHF collection.
5. Charles McDonald, The Federal Railroad Safety Program: 100 Years of Safer Railroads (U.S. Federal Railroad Administration, 1993); and Railroad Safety Appliance Act, 27 Stat. 531, recodified at 49 U.S.C. § 20302.
6. Mark Aldrich, Death Rode the Rails: American Railroad Accidents and Safety, 1828–1965 (Johns Hopkins University Press, 2006). The standard scholarly treatment, and the source behind most reliable figures in this field; its Appendix Two compiles ICC casualty statistics from 1888 onward. Available via Project MUSE.
7. John H. White, Jr., The American Railroad Freight Car: From the Wood-Car Era to the Coming of Steel (Johns Hopkins University Press, 1993). The definitive source on coupler design proliferation.