
1907 Sturtevant Automatic Touring Car / AACA Library via Chuck’s Toyland
What was the first gasoline-powered automobile with fully automatic transmission? It wasn’t built by GM (which hadn’t even been incorporated yet!), but by the Sturtevant Mill Company in Boston, about 120 years ago. Here’s a brief history of the 1904–1907 Sturtevant Automatic Automobile, and how their pioneering self-shifting transmissions worked.
The Sturtevant Mill Company in Harrison Square, Boston, Massachusetts, was originally a manufacturer of fertilizer-processing machinery. It was founded by Thomas L. Sturtevant, a Harvard Law School alumnus who had previously been treasurer of the Bowker Fertilizer Company. As a 1919 item in the journal Farm Chemicals explained:
The difficulties encountered by the Bowker Fertilizer Company in grinding the various fertilizer ingredients with the machinery of those days led Mr. Sturtevant to invent many improvements in both machines and processes of manufacture, and the Sturtevant Mill Company was formed in 1883, to build these devices for the fertilizer industry.
In the 1890s, T.L. Sturtevant was joined by his younger brother, Thomas J. Sturtevant, an MIT alumnus. Together, they accumulated dozens of patents related to industrial machinery and industrial engines.
Around 1902, the Sturtevants decided to branch out into the new and trendy field of automobiles. Their first car (above), shown in the late summer of 1904, was a touring car on a 108-inch wheelbase, with a 314.2 cu. in. horizontally opposed four making about 30 hp at 1,500 rpm. The car had a variety of advanced features, including centrifugal spark advance, but the most significant was its self-shifting change-speed transmission.
The Sturtevant Automatic Transmission
The early Sturtevant car used a two-speed layshaft transmission controlled by a pair of “wet” centrifugal multi-plate clutches located in a hollow flywheel drum full of oil. Here’s an illustration of it from the Sturtevants’ initial 1904 patent:

Fig. 1 of US766551, “Clutch Device Power Transmitting Mechanism,” filed January 5, 1904 and patented August 2, 1904
How did it work? The two clutches were each controlled by a set of spring-loaded centrifugal weights, as shown in the contemporary illustration shown below. (This is from The Horseless Age, but it’s closely adapted from Fig. 2 of the US766551 patent; I used this version because it’s a little clearer.)

Centrifugal weights of the early Sturtevant automatic clutches / The Horseless Age, Aug. 10, 1904
As engine speed increased, centrifugal action would shift those weights outward, which would actuate a lever forcing the clutch plates together. The weights had pressure screws that could be adjusted to vary the speed necessary to engage each clutch. Let’s take another look at the cross-section (again using the version from The Horseless Age), this time adding color to better illustrate which pieces were which. I’ve used red to highlight the low clutch, which was set to engage first. When it engaged, it sent power through the reduction gears. As engine speed increased, centrifugal action would engage the direct clutch, which I’ve highlighted in blue, sending power directly through the output shaft.

Section view of the earliest two-speed Sturtevant automatic transmission / The Horseless Age, Aug. 10, 1904
The low clutch remained engaged in high, but a roller-type one-way clutch (highlighted in green) would let the low gears overrun idly. Reverse was engaged with either a pedal or a side lever, which actuated a lever (purple) to move a sliding gear into mesh with the reverse gears. Interestingly, the transmission would still shift automatically in reverse.
If engine speed dropped enough, either due to decreased road speed or increased load, the high clutch would automatically disengage, eventually followed by the low clutch, dropping the transmission out of gear and allowing the engine to idle. The Sturtevants emphasized that this arrangement made it impossible to stall the engine even when climbing hills at very low speeds, although you could end up in neutral when you didn’t necessarily want to be.
By late 1904, the Sturtevants had come up with a new clutch arrangement, allowing multiple clutches to share the same set of centrifugal weights. Each clutch was carried on a separate spider and was engaged or disengaged sequentially. (As before, I’ve used red to indicate the low clutch and blue to indicate the high clutch.)
This functioned basically the same way as the original arrangement, but it made it easier to incorporate additional clutches for additional forward speeds. In 1906, the Sturtevants installed a three-speed version of this transmission on a two-cylinder Maxwell touring car and a four-cylinder Franklin. The three-speed looked like this in cross-section:
I haven’t gone to the trouble of color-coding the clutches in the above illustration, but as you can see, it incorporated three clutches rather than two and had an additional set of reduction gears. Also, the output was now through the layshaft (the bottom shaft, labeled “Q”) rather than the main shaft, so high gear was indirect rather than direct. This change facilitated a new feature, which The Horseless Age explained like this:
Sliding gear W [highlighted in green above] also has a neutral position in which the jaw clutch is disengaged, but the reverse combination not meshed. When the elements are in this position the direct drive is cut out entirely, as the high clutch acting through spider F revolves idly the free member of the jaw clutch in shaft L”. In traveling very hilly country it might sometimes happen that on account of the high road resistance the car could run only occasionally on the high gear, and then, owing to the great tractive resistance, would not make so good speed as it would were it running on the middle speed with full engine power. By thus disconnecting the high gear, the car becomes practically a two speed vehicle and the full engine speed can be utilized to drive on the intermediate speed.
Otherwise, the Sturtevant transmissions remained completely automatic, with no manual clutching or hand-shifting required. Sturtevant proclaimed, “Self-control tames the driver, and prevents many mischiefs,” and advised, “Ladies, also, please take note, that if you drive the family horse, you can easily manage an Automatic Automobile.”
The Sturtevant Automatic Automobile
Sturtevant Mill Company offered several passenger car models. The original 1904 four-passenger touring car on its 108-inch wheelbase, with the 314.2 cu. in. H-4 engine, was followed in 1905 by a touring car with a 477.1 cu. in. H-6 and about 45 hp, mated to the three-speed automatic. The Automobile noted in August 1905, “Only a few of these machines were built this year, but the Sturtevant Mills [sic] Company is at work on some important designs at its factory in Dorchester.”

1905 Sturtevant chassis — note the right cylinder bank below the dashboard / The Horseless Age, March 15, 1905
Those new models arrived for 1906. The six-cylinder engine was dropped in favor of two new inline-4 engines, a 283.7 cu. in. version making 30–35 hp and a bigger 475.2 cu. in. engine making 45–50 hp. The smaller engine was installed in the new Sturtevant Automatic Flying Roadster, a three-passenger rumble seat roadster on a 98-inch wheelbase, which Sturtevant claimed was “a fine combination for nimbleness, speed, and hill climbing.” It started at $3,500, increased to $4,000 for 1907.

1907 Sturtevant Automatic Flying Roadster / AACA Library via Chuck’s Toyland
The bigger engine was installed in a new seven-passenger tourer, the Sturtevant Automatic Touring Car, with a 120-inch wheelbase and a base price of $5,000, raised in 1907 to $6,000. In 1907, there was also a four-seater tonneau touring car with the smaller engine, priced at $5,000.
The seven-passenger Touring Car used the three-speed automatic, while the Flying Roadster had the two-speed, but the brochure advised Roadster buyers that “four speeds can be easily arranged (at a slightly increased price) when wanted.” This curious arrangement, which was apparently standard on the four-passenger Touring Car, gave four gear changes, but there were two direct drives. I couldn’t find any pictures of this setup, but I assume it mounted two two-speed transmissions in series. The brochure implied that the same could be done with the three-speed transmission, giving a total of six gear changes, again with two direct drive gears.

Sturtevant Touring Car engine, 475.2 cubic inches (7,818 cc) / AACA Library via Chuck’s Toyland
Curiously, late Sturtevant cars now had a clutch pedal in addition to the centrifugal mechanism, allowing the driver to manually control the engagement of the low-speed clutch. As the brochure explained rather owlishly:
the clutch can also be manipulated by the foot just as in other cars, to allow the motor to run free at any time, or to develop its full power on any set of gears. But automatic pedal control cannot be improved for any purpose; for it always does the right thing at the right time, and never forgets or mistakes — and requires no skill. But manual control is at hand for those accustomed to like it. They will soon see the difference.
Here’s an illustration of the clutch, from The Horseless Age:

Late Sturtevant clutch with manual control (the ghosted lever, marked “L”) / The Horseless Age, Jan. 8, 1908, adapted from Fig. 8 of US865449, “Clutch Device Power Transmitting Mechanism,” filed Jan. 21, 1907, patented Sep. 10, 1907
Here’s a brochure illustration of the complete system — I think this is the two-speed:

1907 Sturtevant automatic with manually controllable clutch / AACA Library via Chuck’s Toyland
The final iteration of the three-speed automatic, shown in the illustration below, could now be shifted manually between low and intermediate. The Horseless Age explained, “The direct drive is always under control of its special automatic clutch, but either the intermediate or the low gear may be made operative under the influence of the other clutch” by moving the jaw clutch H (which I’ve highlighted in purple below). It was also still possible to lock out high gear by shifting its jaw clutch J (highlighted in green below) to the neutral position.
Sturtevant cars were very easy to drive for gasoline-powered cars of the 1900s, which attracted considerable attention, but not many sales. Modern sources often claim there were reliability problems with the centrifugal clutches, which is plausible, but I think the bigger problem was price. MeasuringWorth estimates that the $4,000 price tag of a 1907 Sturtevant Flying Roadster represents a relative worth of about $318,000 in 2025, while the $6,000 price of a seven-passenger Automatic Touring Car $6,000 had a relative worth of around $477,000. That was Packard money at the time, and since Sturtevant cars were built only to order, with a 50 percent nonrefundable deposit up front, they were a very expensive proposition even in an age when most cars were still rich man’s toys. I haven’t found any production figures, but the total was undoubtedly very small.

1907 Sturtevant brochure cover / AACA Library via Chuck’s Toyland
The Sturtevant brothers had considered spinning off the automotive business as a separate subsidiary company, but it appears that by the end of 1907, they had realized that it wasn’t going to be financially viable. It was a sideline anyway, and they still hoped they could interest outside licensees. As The Horseless Age remarked while describing the latest Sturtevant automatic transmissions in their January 8, 1908 issue:
It is not the intention of the Sturtevant Company to manufacture these clutches, transmissions or the complete cars embodying them, but to devote themselves to developing to a commercial point the principles of automatic control, leaving to subcompanies or licensees the task of producing them and placing them on the market.
As far as I know, the Sturtevants found no takers. (The Franklin and Maxwell cars with Sturtevant transmissions appear to have been test mules rather than factory options.) However, there were various later efforts at automatic and semiautomatic transmissions that also used some combination of centrifugal and overrunning clutches to change gears, including Oscar Banker’s Mono-Drive transmission, developed over 20 years after the Sturtevants’.
Reliability questions aside, the downfall of transmissions like this was that they had limited ability to respond to changes in load — the transmission wouldn’t delay an upshift unless the engine simply couldn’t turn fast enough to engage the next centrifugal clutch, and it couldn’t downshift automatically unless engine speed actually dropped due to increased load. That might have been fine for underpowered early cars, or for heavy trucks or buses, but as automotive engines became more powerful, it meant the transmission was largely limited to always shifting at the same speeds regardless of accelerator position. The great advance of Hydra-Matic was that its hydraulic controls measured both road speed and throttle opening and balanced both in determining upshift and downshift points.
Although the Sturtevants’ automotive venture didn’t pan out, it didn’t do them any great harm either. The Sturtevant Mill Company continued very successfully in its original fields. Thomas L. Sturtevant died in 1925 at the age of 84. Thomas J. Sturtevant took over as president of the company until his death in 1944. The Sturtevant company still exists today, and one of their modern business lines is processing equipment for the lithium carbonate used in electric car batteries.

1907 Sturtevant Automatic Flying Roadster / AACA Library via Chuck’s Toyland
I don’t know if any Sturtevant cars survive intact today, but if they do, they’re significant historical artifacts. Trying to identify the first of anything is always tricky, but the Sturtevant Automatic Automobile was (probably!) the first production car with true self-shifting transmission.
Related Reading
Driving Impressions: Ford Model T – These Are Not The Controls You Are Used To (by David Saunders)
1904-1909 Fiat 60 HP: The Tallest Production Passenger Car Ever (by Paul N)
Museum Classic/Automotive History: 1905 Premier – The First OHC Hemi Head Automobile Engine And The Search For The Hemi’s True Father (by Paul N)
Car Show Capsule: 1907 Ford Model K Roadster – Watch The Malcomsons Go By (by Aaron65)
I got fascinated by the Sturtevant several years ago and made an animation of the transmission shifting gears. The centrifugal clutch is a subtle movement.
http://polistrasmill.blogspot.com/2014/09/coulda-hada-automatic-in-1905.html
Very nice! I was content just to add some color shading.
Is this company related to the Sturtevant that made the first residential air conditioning systems? They were also based out of Boston.
It seems the Sturtevants at the time were better at devising innovative products than figuring out how to make them inexpensive enough to be able to sell them (with or without an attached car) in decent numbers. The big price increases in the first few years suggest they didn’t realize how much their machinery would cost them to make and sell profitably.
The optional “manual control is at hand for those accustomed to like it” clutch reminds me of the steering wheel upshift/downshift controls on several modern cars with automatic transmissions. There are still those who don’t always want to hand over gear selection to a machine.
Nope, that was the B.F. Sturtevant Company, no relation.
I think the big problem for Sturtevant was that making cars was really just a sideline. They were still making farm equipment, grinding machinery, all that sort of thing, and they didn’t really even have a separate plant to build cars. It was like if a fabric story decided to start making parachutes in their storeroom while still also selling fabric.
It appears that it was kind of an exploratory venture: If there appeared to be a big enough market, they planned to spin off the automobile business as a subsidiary, which could have its own factory and perhaps brought production costs down to a more manageable level. Even failing that, I think they figured they’d have better luck getting licensing deals for their technology if they had actual cars to show rather than just a stack of patent drawings. Ultimately, my assumption is that Thomas L. was too sensible a businessman to risk their profitable core business if the numbers didn’t work. I didn’t find much about him personally, but he was in his 60s by then, and he’d been running his company with considerable success for going on 25 years, so while he was obviously willing to take some risks, I suspect there were limits.
Thanks for this. I knew a little about the existence of the Sturtevant but not the details of the transmission. For the times, this was a pretty impressive piece of engineering. It is a bit surprising that it all ended so quickly. It was expensive, but I could readily see some of the high end brands at least offering it as an option although that kind of thing didn’t really exist. And the luxury brands were all about their own technical solutions, so I suppose that was a major hurdle too.
VERY interesting!\
The “successful Ford 2 speed planetary of 1908 was NOT a delight. It was a dog, noted for it’s constant grinding, clanking, and growling noise. (Still made fun of constantly by Model T drivers today!)
A far more successful 2 speed planetary was used by the Brush Runabout Co., which was virtually silent, could never jerk, grind, or clank, because of oil holes that only permitted gradual engagement of the clutch plates, no matter WHAT the driver did.
Henry Ford didn’t care about noise, because he figured the naive early car owners would never know any different.
None of these were planetary transmissions — they had layshaft gearboxes with sliding gears that mostly remained in constant mesh.
Wonderful article! I have always wondered about this transmission. In Oscar Banker’s autobiography, he says that “I am the father of the automatic transmission” and not the inventor. He writes that the Sturtevant Brothers made the first automatic. Oscar describes some of the complications of the Sturtevant transmission. I would have to find the book in my library to tell you exactly what, but one of them seems to be the engagement and disengagement actuators. Oscar Banker was born Asadour Sarafian in the village of Moonjoosoon in present day Kayseri, Turkey. He is Armenian. Moonjoosoon is a combined word from “Monch Heesoon” meaning fifty young men who founded the village about 600 A.D. Again, MANY THANKS for this article.
Banker was NOT in any sense the father of the automatic transmission, just a blowhard with a shaky grasp of facts and a remarkable capacity for holding grudges. (And yes, I have read his memoir.) He’ll be covered at more length in an upcoming post.
In any event, his memoir contains one brief mention of the Sturtevants, whose name he misspells. It does not contain any insights into the transmission, which was designed and marketed when Banker was a child.
Aaron, thank you so much for this article – it’s fascinating to read about the transmission and the company.