Note: This article was written in 2001, and revised in 2003 by addition of the information cited in notes 4 and 5. The conclusions were limited to those that could be drawn using materials then available. The article now appears somewhat dated, although additional information from digitized newspapers has largely verified its conclusions.
It is generally believed that the first roller coaster in the United States was built by L. A. Thompson and opened at Coney Island, New York, in June, 1884. This ride is described by Thompson's patent #310,966, Roller Coasting Structure, filed April, 1884, and issued January, 1885. Although this roller coaster is the first that is known to have been a commercial success, I will argue here that it was probably not the first roller coaster to be built in the United States. Thompson's firm appears to have dominated the roller coaster market in its first decade, building at least twenty Switchback Railways in the U.S. from 1884 to 1888, and a far greater number of Scenic Railways beginning in 1887. As the winners generally get to write the histories, I believe that Thompson's success has colored our view of the roller coaster's beginnings.
I will ignore here the development of roller-coaster type rides in France in the first quarter of 19th century. This is documented both by contemporary descriptions and by French patents, and these rides are discussed in the books by Mangels (1), Cartmell (2), and Adams (3). Since I have no access to the primary sources, and cannot read French, I have nothing to add to the accounts of those authors. These rides apparently died out in France, and were largely ignored or forgotten in the U.S.
I will also ignore the Mauch Chunk Switchback Railway, which was not initially built for amusement purposes, but for hauling coal. Coal-hauling operations terminated and full-time operation as an amusement railway began about 1872, while coal operations and excursions coexisted at least as early as 1865. The Mauch Chunk became a major tourist destination, and it has been suggested that the commercial success of this operation may have been a stimulus for Thompson's inventions. In any case, the technology used for the Mauch Chunk railway is not particularly relevant to the development of the roller coaster. If we did consider this to be a roller coaster, it would appear to hold the all-time records for length and vertical extent, as well being the first in the U.S.
We can assume that some of the earlier steps in the development of the roller coaster were too obvious to patent. Using the rough equivalent of a sledding hill or a playground slide, a small cart could be hauled to the top by the passenger and then used for the trip down. Even after the development of full-scale roller coasters, innovations in 'coasters' or 'playground coasters' continued to be patented. The problem with these simple machines is the lack of a convenient way to make the return trip. The simplest improvement is to use a winch to haul the car backwards up the same incline used for the descent, which limits the ride to one that is essentially linear in plan and profile. Although this system does not appear to have been patented in the context of the earliest roller-coaster rides, it does appear later in the first shoot-the-chutes patent (US #357,790, 1887, Marine Boat Slide, granted to H. H. Schaefer).
I have attempted to use the U.S. patent literature to deduce something about the early history of roller coasters. Patents consist (in brief) of two parts: the specification, a disclosure of the invention sufficient to enable someone practiced in the art to duplicate it; and the claims, a specific list of novel claims that are to be protected by the patent. In reading the patents, I have used the claims as a guide to the innovations in the patent. Of course, not all claims of novelty are actually valid. (Some recent software and business methods patents come to mind.) Those aspects of the descriptions that fail to make it into the claims are generally a combination of standard practice and details too obvious to patent. The inventor generally attempts to avoid disclosing anything potentially patentable without including it in the claims, because that would decrease the value of the patent. Sometimes the patent descriptions are quite schematic, incomplete, or difficult to interpret, since they assume that the reader is familiar with the then-current state of the art.
The earliest patent relevant to roller coasters is that of John G. Taylor of Baltimore (US #128,674, 1872, Improvement in Inclined Railways). The description is of a ride with obvious similarities to Thompson's first ride of more than a decade later. The layout is out-and-back, with a pair of parallel tracks, linear in plan, connected by a transfer switch at the far end. There is a single claim which covers both the general arrangement and the use of transfer switches at the ends.
Although primitive, this ride is clearly recognizable as a roller coaster. The tracks exhibit undulations, so that the car ascends some segments, unpowered, under its own momentum. There is no lift mechanism: the passengers exit at the end of the ride, and the car is pushed by hand up a short incline, then transferred to the starting track again. The rails are shown schematically in the patent, and the wheels are depicted as flanged. The vehicle is shown as a single car having two longitudinal benches facing across the center, with openings at each end of the car.
While described by its title as an improvement in inclined railways, that seems to have been the standard for patent language, and it appears likely that Taylor's invention was completely novel. It is clear that the patent covers the concept rather than being based on an actual ride. A mechanism for stopping the car at the end of run is not shown or described. Passenger platforms for ingress and egress are not depicted, and the arrangement shown schematically would have been awkward to implement as a working ride. [Note: It seems likely that the patent drawing is a reasonably accurate depiction of the small-scale model that was a required part of the patent application process at that time.] A ride corresponding to this patent may have been built at Savin Rock in the 1870s (4).
The next relevant patent is that of Richard Knudsen of Brooklyn (US # 198,888, 1878, Improvement in Inclined-Plane Railways). The arrangement is similar to Taylor's, but consists in this case of a symmetrical pair of downtracks with similar lifting mechanisms at both ends. The cars are elevated vertically at each end using a pulley arrangement. The layout and the lifting apparatus are not mentioned in the claims, which are for a fixed skid brake used to stop the cars at each end. Although we have no direct evidence on this point, I conjecture from the restricted nature of the claims that rides of this sort may have already been in use. The inventor notes that the device was designed "to be used in public gardens and other places of amusement."
The 1882 patent of Alexander (US #269,554) describes an artificial sledding hill. This is clearly not a roller coaster, although there are some technological similarities. In particular, Alexander's patent includes a steam-operated cable lift to return sleds to the top of the hill. A lift system appears to have been absent from the earliest roller coasters, including Thompson's.
A flurry of patents issued in 1884 coincides approximately with Thompson's ride at Coney Island. The patents of Wood (US #291,261 Circular Gravity-Railway) and Stevens (US #298,710 Roller Coasting Device) describe roller coasters which are extremely similar. In both cases, the layout is a full circle, with adjacent passenger loading and unloading platforms. The platforms are within the circle and accessible by stairs. In both cases, the passengers are seated facing sideways. The former patent relies on precisely determined distances and heights such that the car stops on a level stretch of track, while the latter patent allows the car to stop on the final upgrade, where it is held by an anti-rollback pawl. The inclusion of exact dimensions in Wood's patent suggests that an experimental prototype had been built, although it may not have been full-size. Although it has been suggested that Wood never built a ride that was operated for the public, construction of a ride at Ponce de Leon Springs (Georgia) was reported in the Augusta Chronicle and the New York Times (5). Wood's obituary (6) indicates that he licensed his patent for as much as $17,000 in a single year, strongly suggesting that multiple rides based on his invention were built by others. A possible example may be a ride at Oakland Beach, Conneaut Lake, Pennsylvania. An 1885 advertisement for this ride is reproduced in Bush & Hershey's book (7) about Conneaut Lake Park. The drawing in the ad appears to match the Wood and Stevens patents quite closely, and the description of the ride as a "circular gravity railway or roller coaster" is remarkably similar to the titles of both patents.
The Wood and Stevens patents were issued before Thompson's ride opened, and their patent applications had been filed the previous year, in 1883. The patent dates indicate that these developments in roller coaster technology preceded the exhibition of Thompson's ride. As the Stevens patent was assigned to the Roller Coaster Company of America, even the name that would eventually become the generic term for the device had already been coined. As further evidence that rides of this sort had been built by the 1884 season, it should be noted that Hinkle's patent (US #307,942), filed in August, and granted in November, 1884, elaborates this arrangement into the form of a double circle. The Stoddard & Terwilliger patent (US #314,626), which was filed in June, 1884, (probably before Thompson's ride opened) and granted in March, 1885, states that "The operations of a roller coasting-car are so well known that a description in detail of its movements we deem unnecessary." And the Pusey patent (US #318,025), filed in January, 1884, notes "I am also aware of the fact that a coasting-hill has been used consisting of a circular railway descending on each side from an elevated plaform, upon which railway the cars descend by gravity from the highest to the lowest point and are carried around by their momentum nearly up to the starting-point." I think we can conclude that Thompson's Switchback Railway of 1884 was not the first roller coaster to be publicly operated in the United States. Unfortunately, it will likely be a tedious job for coaster historians to locate references to the construction or operation of these earlier rides (8).
As a concluding note, it appears from the patent literature that linear and circular roller coasters initially followed parallel but separate paths of development. Several patents describe improvements to linear roller coasters similar to Thompson's. Stoddard & Terwilliger's patent uses flangeless wheels and side-friction rollers instead of the then-standard flanged wheels, but limits its claims for this arrangement specifically to circular roller coasters. The first patent that clearly describes a roller coaster having a layout with curves of differing curvatures is that of Alcoke (US #317,273, 1885). This has a generally oval-shaped path, with a linear ending ramp which terminates very near the starting point. The ends (which meet essentially at right angles) are joined by a turntable, switch, or a track of very sharp curvature. This patent was preceded, however, by Rankin's 1884 patent for a closed-circuit ice railway
Copyright © 2001 Victor Canfield
1. William F. Mangels, The Outdoor Amusement Industry, from the Earliest Times to the Present. (NY: Vantage Press, 1952). [return]
2. Robert Cartmell, The Incredible Scream Machine: A History of the Roller Coaster. (Bowling Green: Bowling Green State University Popular Press, 1987). [return]
3. Judith A. Adams, The American Amusement Park Industry: A History of Technology and Thrills. (Boston: Twayne Publishers, 1991). [return]
4. Bennett W. Dorman, Savin Rock, An Illustrated History. (Photo Restoration & Design, 1998). Photograph, p. 21. [return]
5. "Sliding Up Hill," New York Times, June 27, 1884. [return]
6. Toledo Blade, May 4, 1909. [return]
7. Lee O. Bush & Richard F. Hershey, Conneaut Lake Park, The First 100 Years of Fun. (Amusement Park Books, 1992). [return]
8. This process may be simplified by automated conversion of historic newspapers into searchable digital formats. An initial effort in this direction can be found at my Early Accounts of Roller Coasters page. [return]