“Tom Thumb” Locomotive: History, Race Legend, Specs & Replica

Last revised: August 28, 2026

Peter Cooper’s little vertical-boiler engine of 1830 is the first American-built steam locomotive to run on a common-carrier railroad—the Baltimore & Ohio. However, it was not the first steam locomotive in the United States (the English Stourbridge Lion ran on Delaware & Hudson track in August 1829), not the first steam machine Americans had tried (Colonel John Stevens’s circular-track “Steam Waggon” at Hoboken, 1825–26), and not the first American locomotive in regular revenue service (South Carolina Canal & Rail Road’s Best Friend of Charleston, Christmas Day 1830). Cooper never meant it for a timetable. He built a demonstrator to prove that steam could take the B&O’s tight curves and modest grades, and on August 28, 1830, he hauled company directors from Baltimore to Ellicott’s Mills and back.

The horse race that made the engine famous is another matter. The story—locomotive leading, blower belt slipping, gallant gray winning—was first printed in 1868, thirty-eight years later. No 1830 newspaper has been found that describes it. The original engine was salvaged in 1834. Every photograph on this page, and every machine you can still stand next to, is the 1927 Fair of the Iron Horse replica: larger, heavier, and built to perform at a centennial pageant, not to copy lost drawings. The replica is historic in its own right. It is not the 1830 locomotive.

Tom Thumb replica under steam at the Baltimore and Ohio Railroad Museum 1976 William Myers

The Tom Thumb replica under steam at the Baltimore & Ohio Railroad Museum, 1976. William Myers photo. American-Rails.com collection.

On this page: At a glance · What it was—and was not · Cooper, Canton, and the shop · The 1830 machine · The B&O to Ellicott’s Mills · August 28, 1830 · The horse-race legend · York, Grasshoppers, salvage · The 1927 replica · Original vs replica specs

At a Glance

Item	Detail
What it is	First American-built steam locomotive to operate on a common-carrier railroad (B&O). Experimental demonstrator, not a revenue engine
What it is not	Not the first steam locomotive in the United States (Stourbridge Lion, 1829). Not the first regular U.S. revenue locomotive (Best Friend of Charleston, Dec. 25, 1830)
Designer	Peter Cooper. Built 1829–30 in George W. Johnson’s Baltimore shop (James Millholland apprenticed there)
Why Cooper built it	Canton, Baltimore, real-estate holdings whose value depended on the B&O succeeding with steam
Documented run	August 28, 1830: directors’ trip Baltimore–Ellicott’s Mills and return, coinciding with dedication of the Oliver Viaduct
Horse race	Folklore. First published in John H. B. Latrobe’s Maryland Institute lecture, March 23, 1868. No contemporary newspaper account found. B&O Museum: never substantiated
Name “Tom Thumb”	Also first appears in that 1868 lecture. Period papers did not use it in 1830
Original fate	Used in tests into 1831; salvaged for parts in 1834. No original drawings survive
What you can see	1927 Fair of the Iron Horse replica at the B&O Railroad Museum (National Museum of Railroad History & Innovation), Baltimore. Larger and heavier than the original; built for pageant performance

What it was—and was not

Steam on rails is an English story first. Richard Trevithick showed a locomotive at Penydarren on February 21, 1804. George Stephenson’s work, and the 1829 Rainhill Trials that made Rocket famous, are the lineage American promoters were reading when they chartered railroads. The United States imported that idea before it built its own.

Order of the American firsts, scoped tightly:

1825–26, Hoboken. Colonel John Stevens ran a steam wagon on a circular track on his estate. Demonstration, not a common carrier.
August 8, 1829, Honesdale. The English Stourbridge Lion was the first steam locomotive to operate in the United States. Too heavy for Delaware & Hudson track; it did not become a regular engine.
Summer 1830, Baltimore. Cooper’s engine was the first American-built steam locomotive to run on a common-carrier railroad. It was a test machine.
December 25, 1830, Charleston. Best Friend of Charleston was the first American-built locomotive placed in regular passenger service.

The Baltimore & Ohio was the first common-carrier railroad in the United States to prove, on its own track, that a homemade steam engine could take directors thirteen miles into the Patapsco valley and bring them home. That was enough. In 1831 the company advertised for working locomotives, and horses left the main line.

Cooper, Canton, and the shop

Peter Cooper was a New York inventor and industrialist with a Baltimore problem. He held substantial real estate in what became the Canton neighborhood. If the B&O reached the Ohio, those lots were worth more. If the railroad stalled on horse power, or if English engines proved too rigid for American curves, the investment was a loss. Cooper later said as much: he needed the road to “go.”

He did not walk in as a stranger with a finished locomotive. He assembled one in Baltimore from parts he already had and materials he could find. He brought a small brass engine from New York—about one horsepower, he recalled, with a cylinder on the order of 3½ inches by 14 inches. He rolled a vertical boiler from boiler iron, roughly the size of a kitchen wash-boiler. Iron pipe for a multi-tube boiler was not for sale in this country, so he broke two muskets and used the barrels as flues. A coach-maker’s shop supplied the woodwork. The machine-shop work is credited to George W. Johnson of Baltimore. John H. White Jr. notes that the eighteen-year-old James Millholland was apprenticed there; Millholland later became one of the important American locomotive designers of the mid-nineteenth century.

Cooper’s own late-life interviews are useful and slippery. He said he wanted two demonstrations: that steam could take the B&O’s short-radius curves, and that he could get rotary motion without a crank. The crankless idea did not last. After early trials he used a conventional crank and gearing to one axle. He also, in those same late interviews, used the name Tom Thumb. The name is not in the 1830 papers. Treat Cooper’s reminiscence as color, not as a builder’s plate.

The 1830 machine

Forget a later 4-4-0 in miniature. The original was closer to a powered flatcar. A vertical boiler sat on a four-wheel frame. A single vertical cylinder drove through a crank and a spur-and-pinion gear to one axle. Wheels were about 2½ feet (30 inches) in diameter. Fuel was anthracite. A blower in the stack, belt- or cord-driven from the powered axle, forced draft that a boiler that small could not have kept with natural draft alone. When that belt left the pulley, steam died. That mechanical fact is why the race story, whether or not the race happened, is at least mechanically plausible.

Period descriptions of the original do not agree to the inch, because nobody treated a shop experiment as a diagrammed class engine:

Cylinder. One cylinder. Cooper: about 3½ × 14 inches. Ross Winans, via William H. Brown: 3¼ × 14¼. Other nineteenth-century notes: 3¼–3½ bore, 14–14½ stroke. All of those are a long way from the replica’s 5 × 27.
Boiler. Vertical, tubular. Winans remembered it as not more than about 20 inches in diameter and five to six feet high, firebox in the lower section. Wikipedia’s 27 × 66 inches is the replica.
Tubes. Musket barrels. That detail is in Cooper’s own account and in every serious secondary treatment.
Weight. Latrobe: could not have weighed a ton. Cooper, later: about a ton. There is no shop scale ticket. “Less than one ton” is folklore for small, not a measured figure. The replica weighs about five tons.
Power. On the order of 1.4 horsepower in later calculations of the Ellicott’s Mills run. Do not inflate that into a comparison with Rocket. Rainhill and a one-cylinder demonstrator are not the same test.

Wheel arrangement is a later argument. Whyte notation did not exist in 1830. The original had four wheels and a gear drive to one axle. Sources have called it 0-2-2, 0-4-0, and 2-2-0. The 1927 replica is photographed with the geared wheels at the rear and is usually labeled 2-2-0. Calling the 1830 engine a Planet type imports Stephenson’s 1830 English 2-2-0 onto a geared vertical-boiler car it does not resemble.

Illustration of the legendary race between Tom Thumb and a horse-drawn Baltimore and Ohio car

The race that made the engine famous: Cooper’s locomotive against a horse-drawn B&O car. The scene is a later illustration of a story first printed in 1868. American-Rails.com collection.

The B&O to Ellicott’s Mills

The Maryland legislature chartered the Baltimore & Ohio on February 28, 1827. Construction did not begin on February 4, 1828. The ceremonial beginning was July 4, 1828, when Charles Carroll of Carrollton—last surviving signer of the Declaration of Independence—turned the first spadeful at Mount Clare. The Masons laid the First Stone. Carroll called it among the most important acts of his life, “second only to my signing the Declaration of Independence, if even it be second to that.”

The first mile and a half, Pratt Street at Mount Clare to the Carrollton Viaduct, opened to public riding on January 7, 1830, horse-drawn, a few cents a trip. The thirteen-mile division to Ellicott’s Mills (today’s Ellicott City) opened for passengers on May 24, 1830. The company advertised three round trips a day from the Pratt Street depot and charged 75 cents for the 26-mile out-and-back. Trains were still horses. Double track as far as the Mills is what later made a parallel race even imaginable. Ellicott City Station, begun that year, is the oldest surviving railroad station in the United States.

That is the railroad Cooper was trying to save: a common carrier already selling tickets, still using animals, watching the Erie Canal and the C&O Canal and wondering whether English locomotives would even take American curvature.

August 28, 1830

Lock this date to what the papers actually covered. Saturday, August 28, 1830, was the dedication of the Oliver Viaduct, the granite span over the Frederick Turnpike at Ellicott’s Mills. Directors and guests left Baltimore in the morning and came home in the evening. Cooper’s engine made the trip, pushing or towing an open passenger car. The B&O Museum’s collection note is the clean version: Cooper carried the directors to Ellicott’s Mills at something like 10–14 mph.

Cooper’s own later numbers are more specific and should be labeled as recollection: six men on the engine, thirty-six on the car; thirteen miles out in one hour and twelve minutes; return in fifty-seven minutes; average grade about eighteen feet to the mile. A widely reprinted postcard compresses that into eighteen passengers and “about two hours.” Latrobe, writing decades later, remembered curves taken at fifteen miles an hour and gentlemen trying to write their names at eighteen. Speeds in the teens are the consistent picture. A timed, instrumented run is not.

The directors were convinced. Bonds could be sold. Steam, not a horse roster, was the way west. That institutional result does not depend on whether a stage-line gray beat the engine on the way home.

Vintage postcard of Peter Cooper Tom Thumb locomotive describing the August 28 1830 trip and horse race

A vintage postcard of Tom Thumb. The caption treats August 28, 1830, and the horse race as one event. The trip is documented; the race is the later legend. American-Rails.com collection.

The horse-race legend

On March 23, 1868, John H. B. Latrobe—longtime B&O counsel, inventor, and the man whose family name is all over early B&O engineering—lectured at the Maryland Institute in Baltimore. The printed lecture is the first known appearance of the engine named Tom Thumb and of the race with a horse. Paul S. Bridge’s account in the Winter 2011 Maryland Historical Magazine is the modern tracing of that origin.

The story Latrobe told, and that every schoolbook and postage stamp repeated, is simple. On the return from the Mills, Stockton & Stokes—the Baltimore stage proprietors—had a gallant gray on the second track. They met the engine at the Relay House. Cooper accepted a race toward Baltimore. The locomotive pulled away, taking curves at fifteen miles an hour. Then the blower belt ran off. Draft collapsed, steam fell, the horse went by, and Cooper could not catch up. Steam had still made its point.

Problems with treating that as an 1830 news event:

No contemporary newspaper. Something that picturesque would have been a gift to the American and every other Baltimore sheet. Guilford-area B&O research, and Bridge’s survey, have not found one.
The dates wander. August 28 is the usual pairing with the directors’ trip. Other printings give August 25 or September 28. Thomas Hughes’s 1886 Life and Times of Peter Cooper put an informal race on September 28. The Federal Highway Administration’s Carl Rakeman painting series used August 28 as if it were settled.
Cooper himself shrugged. A late recollection, as quoted in Rossiter Raymond’s biography: they tried a little race one day; it didn’t amount to anything; it was rather funny; the locomotive got out of gear.
The museum will not underwrite it. The B&O Railroad Museum’s collection text calls the race a staple of American folklore that was never substantiated through documentation.

None of that proves a race did not happen. Double track existed. Stockton & Stokes had reason to dislike steam. A blower belt is a real single point of failure. An informal dash on a return trip could have left no reporter on the car.

The name is part of the same package. “Tom Thumb” fit a tiny engine, and there was already a famous trotting horse by that name, plus 1830 jokes about a “Tom Thumb of steam engines.” Latrobe’s lecture stuck the nickname to Cooper’s machine. After 1868, every retelling used it. Cooper’s later interviews used it too. The 1830 B&O did not stencil it on a tank.

York, Grasshoppers, salvage

Cooper’s engine kept making trips toward Ellicott’s Mills into the winter and was still in the picture as late as March 1831. It was never diagrammed for a freight pool or a passenger assignment. The B&O needed a real locomotive, not a one-horsepower demonstrator.

In January 1831 the company advertised for engines that could pull fifteen tons at fifteen miles an hour. The 1831 trials produced Phineas Davis’s York, then the vertical-boiler “Grasshoppers” that replaced the horses in everyday work. Grasshoppers stayed in B&O yard and transfer service for decades. That fleet, not Cooper’s demonstrator, is the first American production answer to the 1830 showing.

The original engine had no constituency as a relic. Winans remembered the boiler lying around his shops for years. In 1834 it was salvaged for parts and disappeared. A few musket-barrel tubes were said to survive in private hands into later decades. There is no original Tom Thumb to restore.

The 1927 replica

No original drawings survive. What exists is a chain of memory:

c. 1875. Cooper, finally getting public credit, described the engine and made rough sketches. The museum dates his recognition and drawing to 1875.
1892. Major Joseph Pangborn, newspaperman and railroad publicist, built a wooden model as part of a series of early-locomotive models. The 1927 replica followed Pangborn more than any lost 1830 shop drawing.
1926–27. The B&O built an operating replica at Mount Clare for the 1927 Fair of the Iron Horse, the railroad’s centennial pageant at Halethorpe. The museum is blunt: it follows Cooper’s general description but was built for performance at the fair, not historical accuracy.

The differences are not cosmetic. The replica is larger, heavier, and considerably taller. Length is 13 feet 2¾ inches; height to the stack is 12 feet 9 inches. Those are replica figures. Listing 13 feet 2¾ inches as the original’s width is a catalog error. Cylinder is a single 5-inch bore by 27-inch stroke, 90 psi, calculated tractive effort 820 pounds. Engine weight is about 10,000 pounds in the museum’s figures (Staufer’s diagram, via Locobase, 10,800), with 5,800 pounds on the geared driver. Heating surface is a real boiler: 38 two-inch tubes, about 40 square feet total with the firebox. The original’s musket flues and wash-boiler shell were not that machine. Draft on the replica comes from a large blower on the platform, not a belt to a fan in the stack.

The replica is itself approaching a century old. The museum lists it as operational and brings it out for special appearances; it is not a daily road engine. William Myers photographed it under steam at the museum in 1976. A.C. Kalmbach photographed demonstrations at Ellicott City in the 1950s. A working half-scale version has been shown at the Hesston Steam Museum in Indiana. In 1952 the engine—race and all—appeared on a U.S. postage stamp. The machine in Baltimore is the Fair of the Iron Horse locomotive.

Baltimore and Ohio Tom Thumb replica demonstrating at Ellicott City depot 1950s A C Kalmbach

The B&O’s Tom Thumb replica demonstrating at the historic Ellicott City depot, Maryland, during the 1950s. A.C. Kalmbach photo. American-Rails.com collection.

Original vs replica specifications

Put the two machines in one table so the replica is not silently substituted for 1830. Original figures are the range in Cooper, Winans/Brown, Wright, and the 1929 New York Times summary. Replica figures follow the B&O Museum collection sheet and the Staufer/Sagle as compiled in Locobase 1015, note that those measures are the replica’s.

Specification	Original (1830)	Replica (1926–27)
Builder	Peter Cooper, George W. Johnson shop, Baltimore	B&O, Mount Clare, for the Fair of the Iron Horse
Purpose	Demonstrator; not revenue service	Centennial pageant engine, built to pull and to be seen
Arrangement	Four wheels, gear drive to one axle. Whyte labels (0-2-2 / 0-4-0 / 2-2-0) are later	Photographed as 2-2-0 (geared wheels at the rear). Vertical boiler
Cylinder	One, vertical; about 3¼–3½ in × 14–14½ in	One, vertical; 5 in × 27 in
Boiler	Vertical; ~20 in diameter × ~5–6 ft high; musket-barrel tubes; anthracite	27 in diameter × 66 in high; 38 × 2 in tubes, 27 in long; anthracite
Heating surface	Not reliably recorded	~28.4 sq ft tubes + 11.5 sq ft firebox ≈ 39.9 sq ft
Blower	Fan in the stack, belt/cord from the powered axle	Large platform-mounted blower
Wheel diameter	~30 in (2½ ft)	Geared driver as diagrammed with the replica (do not back-apply to 1830)
Boiler pressure	Not recorded	90 psi
Tractive effort	Not recorded	820 lbf (calculated)
Weight	On the order of a ton or less (Latrobe / Cooper; no scale ticket)	10,000 lb (museum); 10,800 lb (Staufer diagram). 5,800 lb on the geared driver
Length / height	Not recorded. Do not use replica numbers	Length 13 ft 2¾ in; height 12 ft 9 in (to the stack)
Power	~1.4 hp in later calculations of the Mills run; Cooper said about 1 hp	Sized to haul exhibition cars; not a stand-in for 1830 output
Documented speed	Directors’ trip: about 10–14 mph; later memories up to ~15–18 mph on curves	Museum lists about 13 mph
Disposition	Salvaged 1834	B&O Railroad Museum, Baltimore; operational for special appearances

Sources

B&O Railroad Museum / National Museum of Railroad History & Innovation, collection notes on Tom Thumb (Aug. 28, 1830 directors’ trip; race as unsubstantiated legend; 1875 Cooper sketches; 1927 replica built for fair performance)
Paul S. Bridge, “The Real Story of the ‘Tom Thumb’,” Maryland Historical Magazine 106, no. 4 (Winter 2011): Latrobe’s March 23, 1868 Maryland Institute lecture as first appearance of the name and the race; Oliver Viaduct dedication August 28, 1830
John H. B. Latrobe, Maryland Institute lecture, March 23, 1868 (race narrative; engine “could not have weighed a ton”)
Peter Cooper, late-life accounts as quoted in Rossiter W. Raymond, Peter Cooper (Riverside Biographical Series), and in Thomas Hughes, Life and Times of Peter Cooper (1886): musket-barrel tubes, ~3½ × 14 cylinder, Canton motive, shrug at the race
William H. Brown, The History of the First Locomotives in America (1871/1877), quoting Ross Winans: ~20 in × 5–6 ft boiler, 3¼ × 14¼ cylinder, 2½ ft wheels, gear to the forward axle
John H. White Jr., A History of the American Locomotive: Its Development, 1830–1880 (1968): Johnson shop; Millholland as apprentice
Lawrence W. Sagle, B&O Power (Staufer, 1964), and Lawrence W. Sagle, “Tom Thumb: Little Known Facts About First Locomotive Built in America Slowly Come to Light,” Railway and Locomotive Historical Society Bulletin 73 (May 1948)
Steve Llanso, Locobase 1015 (original notes plus Staufer replica diagram: 5 × 27 cylinder, 90 psi, 820 lbf TE, 5,800 lb adhesion, 10,800 lb engine)
“Our First Locomotive Tested a Century Ago,” New York Times, December 15, 1929 (3½ × 14½ cylinder, 20 in × 5 ft boiler, musket tubes)
John F. Stover, History of the Baltimore and Ohio Railroad (Purdue, 1987)
Edward Hungerford, The Story of the Baltimore & Ohio Railroad, 1827–1927 (1928): First Stone July 4, 1828; January 7, 1830 Mount Clare–Carrollton Viaduct; May 24, 1830 Ellicott’s Mills opening
Guilford Area History (Howard County), “The Legend of the Tom Thumb Race”: no contemporary race notice; August 28 as Oliver Viaduct day

The original was scrapped in 1834. The locomotive in Baltimore is the 1927 replica. Confirm hours, steam demonstrations, and whether the replica is under fire with the museum before you plan a visit around it. American-Rails.com does not operate the exhibit.

Facebook X Pinterest Tumblr Reddit WhatsApp Messenger