This paper is based on research by Tony Hadland for a small section of the book The Dancing Chain by Ron Shepherd, Frank Berto and others. The research was carried out in 1996, primarily using British patents.
Introduction to change speed gearing
Early change-speed systems for cycles
Industrial parallels and antecedents for derailleurs
Late Victorian developments in non-derailleur change-speed cycle gearing
Introduction to change-speed gearing
Some wheeled vehicles are propelled by an external force. They may be pulled by a horse, pushed by a human, or towed by a motor vehicle. No vehicle propelled by an external source has a change-speed gearing system, because there would be no benefit.
However, those vehicles where the motive force is mounted on the vehicle often do. These 'self-propelled' vehicles include railway engines, motor cars and, of course, cycles.
But why should you want a change-speed system? If the power source can function adequately under all normal combinations of load and speed (e.g. steam railway engines) then there is no need. But if the power source works best over a relatively restricted range of output rates (e.g. internal combustion engines - and most cyclists) a change-speed system can be extremely beneficial.
As Henry Sturmey put it in 1901:
"After all, the human being is not unlike the internal combustion motor. It has been found absolutely indispensable for the latter that variable speed gears should be provided, simply because the engine cannot develop more than a given amount of power, and if its speed is reduced unduly owing to the heaviness of the drive, its power declines so rapidly that it cannot take the ascent, and although the human motor possesses far more elasticity than the petrol engine, its limits are very soon reached, and a variable speed gear is the correct thing for the rider who desires to combine the maximum of distance and speed with the minimum of exertion."
The purpose of a change-speed system is therefore to provide a mediator between the output source and the driven wheel(s). The purpose of this mediator is to improve the matching of output power to load conditions. Or, as Cycling magazine put it in 1903, to enable the rider "to overcome the increased resistance [of hills or head winds] by a steady flow of energy spread over the longer period caused by the sacrifice of speed."
Whether we are cycling or motoring, to ascend a steep incline we change down to bottom gear, sacrificing speed for power. Having crested the hill, we shift into top, trading power for speed. What enables us to do this is, of course, some form of change-speed gearing.
The concept of change-speed (or variable velocity) gearing goes back a long way. The great pioneer of steam traction, James Watt, referred to a two-speed transmission in his 1784 patent No.1432 relating to steam carriages:
"Motion [from a steam engine] is communicated to the axle-tree of one or more wheels of the carriage by means of the "circulating rotative to machinery" formerly patented by the inventor. Two or more loose wheels of different diameters are placed to be locked on the axle and impart extra power for bad roads or steep ascents."
Watt’s drive system was relatively sophisticated. A more typical early solution involved multiple drive trains.
A typical drive train of the time consisted of a pulley on the output shaft of the power source, another pulley on the driven wheel axle, and a drive belt linking the two. Obviously the relative speed of the output shaft and the driven wheel depended on the relative diameters of the pulleys. If the drive wheel’s pulley was twice the diameter of the output shaft’s pulley, the wheel would turn at half the speed of the shaft.
The simplest way of achieving a choice of speeds was to install a second drive train with a different combination of pulley diameters. Hence you might use one drive train for reasonably fast travel over well-surfaced or level roads, the other for "bad roads or steep ascents".
Obviously only one drive train could be in use at a time or the machine would wreck itself, or at least stall. This is because it would be trying to rotate the wheel axle at two different speeds simultaneously. In the simplest systems, changing gear necessitated stopping the vehicle and its power source, disengaging the drive last used and engaging the other drive. Disengagement typically involved loosening a set-bolt clamping the output pulley to the output shaft: engagement, the converse.
By 1819 this type of transmission had found its way into a human-powered vehicle. The 'British Facilitator or Travelling Car' invented by B. Smythe, a Liverpool surveyor, was an early 19th century equivalent of the current British Brox. It used two drive belts - one for going uphill, the other for going down. The output pulleys were on a crankshaft driven by treadles (see illustration below).
By the middle of the 19th century, references to change-speed gearing where being made in a number of patents related to steam-driven agricultural machinery. For example, Pierre Barratt’s 1847 patent 11,977 for a mechanical digger:
"The travelling wheels have two movements, one for use when digging is going on, and the other for moving the engine from place to place. This is effected by an arrangement of different sized spur wheels, which may be thrown in and out of gear as required."
From Watt’s time until the mid 19th century most change-speed patents dealt with two-speed systems. However, from the 1850s onwards some explicitly mentioned three-speed transmissions. Among these were William Dorman and Charles Cowper’s 1389 of 1859, and Morris Keene’s 1836 of 1865.
Many of the patents relate primarily to agricultural machinery but, not
surprisingly, there were
parallel developments in carriages and other vehicles for "common
roads". These included
patents by John Bischoff (637 of 1866), Charles Appleby (1757 of 1866)
and John Alleyne (962 of 1865).
Early change-speed systems for cycles
The ‘British Facilitator’ referred to above was perhaps the first human powered vehicle design to incorporate change-speed gearing. However, it was not until the commercial exploitation of the velocipede in the 1860s that developments in variable gearing for cycles began in earnest.
In the early 1860s Michaux et Compagnie of Paris started making the first commercially successful bicycles. A few years later, in March 1868, Ferdinand Villepigue of 16 Northumberland Street, London applied for a patent concerning ‘Improvements in the Construction of Velocipedes’ (708 of 1869).
Villepigue’s patent was brief, vague and widely cast. It attempted to cover a dozen key aspects of cycle design, including rubber tyres, adjustable metal spokes and an adjustable saddle. Had it been successful, Villepigue would have become a very rich man. In reality his patent never got beyond provisional acceptance. But it is important because it outlined the concept of change-speed gearing for cycles:
"The application of a differential movement to the fore [driven] wheel, whereby communicating a difference of velocity, or a difference in the number of revolutions made respectively by the crank and the fore wheel; that is to say, that according to the slope or other conditions of the road you are travelling over, or the speed which you wish to attain in relation to the force expended, you can make the wheel describe more or fewer revolutions than the crank. Thus, for example, if the road be good the wheel may perform two revolutions, while the crank revolves once; or contrariwise the wheel may be made to make only half a turn while the crank is making a whole one if the ground be bad and the inclination steep."
Barberon & Meunier (1869)
So, Villepigue had outlined the concept, but not how to put it into practice. This fell to a pair of Parisians, Alphonse Barberon and Joseph Meunier. Six months after Villepigue’s application, they applied via the London patent agent H H Lake, for British patent 2626 of 1869. This also got no further than provisional acceptance, but it did describe two ways of applying two-speed gearing to velocipedes.
In both versions the pedal crank axle passed through a hollow front wheel axle. In one version the drive went via a belt from a pulley on the crank axle, over the top of the wheel (via an idler pulley) and down to a pulley fixed to the other side of the wheel. The pulleys were sized to give a gearing up effect. The specification also allowed for the use of a chain and sprockets instead of the belt and pulleys.
Gear changing was controlled by a pawl and ratchet system. When the rider pedalled forwards in the usual way, the cranks drove the front wheel directly. If the rider pedalled backwards the belt or chain came into action, giving a higher gear.
Barberon and Meunier realised that pedalling backwards could be problematic. Their specification therefore included an alternative version whereby either gear could be used whilst pedalling forwards. This had the whole drive system on one side of the front wheel and necessitated a manual gear selector.
Hot on the heels of Barberon and Meunier’s patent application, came a British one from James Starley and William Hillman, No.2236 of 1870.
The front (driven) wheel ran loose on the crank axle. On each end of the hub was a pinion that meshed with a compound (stepped) pinion on a countershaft. The other ‘step’ of the compound pinion meshed with a pinion keyed to the crank axle. The relative sizing of the pinions caused the wheel to rotate faster than the crank, and in the same direction (see patent illustrations below).
The countershafts were in hinged frames, which could be moved via a rod linkage, thus disengaging the compound pinions from the wheel and crank pinions. As the wheel then slowed to match the speed of the crank, a pawl on the inside of the hub engaged with a ratchet wheel on the crank axle, thus giving direct drive.
The countershaft system adopted by Starley and Hillman was similar to that described in John Alleyne’s 1865 patent 962 for steam-driven road carriages referred to above. However, Alleyne’s earlier system was much cruder in that it necessitated:
A plethora of patents (1877-1906)
For the rest of the Queen Victoria’s reign and into the Edwardian era, there was intense interest in change-speed systems, especially for cycles. ‘Logos’, writing in 1908, listed over 750 "patents relating to speed-gears" from the years 1877-1906 - and even this listing is incomplete.
Despite the very wide range of change-speed systems evolved during this period, they essentially fell into two broad classes:
This latter was used in some tricycles of the 1880s. More sophisticated means of selectively clutching the chainwheels were evolved than the earlier approach of stopping the vehicle, getting out the tool kit and adjusting set screws.
Applying a similar approach to rear driven safety bicycles was more difficult, as there was insufficient width easily to accommodate two or more separate chain drives. However, fitting multiple sprockets and chainwheels, and moving the chain from one combination to another was feasible, and relatively lightweight - albeit inconvenient in operation.
Out of the multiple drive systems evolved various derailleur systems.
These applied remote control
to the process and allowed the chain to be shifted whilst the machine
was ridden. But this did not
happen without the influence of another area in which change-speed
systems had been evolving -
industrial machinery.
Industrial parallels and antecedents for derailleurs
Belt drives had long been used in industrial equipment. As with the early road vehicles noted above, the simplest and oldest industrial change-speed belt-drive systems employed two or more parallel drive systems, each giving a different ratio. To select the desired drive train, its pulleys were locked to the drive shaft, the pulleys of the other trains being left loose. Typically this would be done using set screws to lock an otherwise freely running pulley to the shaft.
Shifting the drive belt on and off a pulley (before 1855)
By the mid 19th century the idea of using a fork or similar device to shift a belt on or off a pulley was well established. In most cases the technique was used to engage or disengage drive, rather than to change speed.
Treadle-driven domestic sewing machines typically used the same principle. The leather drive band passed through a metal loop, somewhat reminiscent of a derailleur front changer mechanism. This was mounted on a pivot and had a hand-operated lever. Actuating the lever ‘derailed’ the band, thus acting as a clutch. The late 1850s saw a number of patents concerned with placing on and throwing off drive belts from pulleys. These included L R Bodmer’s 2886 of 1855, S & D Taylor’s 1375 of 1858, M A F Mennon’s 1973 of 1858 and J Brière’s 2101 of 1859.
Shifting the belt from pulley to pulley (by 1855)
A particularly interesting provisional patent from this period was J H Johnson’s 2415 of 1855. This involved moving the belt from one set of pulleys to another. Essentially it concerned an anti-overload device in which a weighted lever was actuated by excess power input. The lever was connected to a belt fork that moved the drive belt from the drive pulley to a loose pulley. Both were of the same diameter on the same axle and flangeless. Hence the system was still a form of clutch rather than a true change-speed system. And although it moved the belt from one pulley to another pulley, the belt was not moving to and from different-sized pulleys.
The technique of shifting a drive belt from a drive pulley to a free pulley of the same diameter mounted alongside was probably already commonplace by this time.
A later patent, L Higginbottom and T Mannock’s 14,102 of 1886 for a travelling crane transmission, used two drive systems of the type just described, mounted side by side. Each system gave a different drive ratio and had its own drive belt. The two belt forks were operated by a single lever. This lever could therefore simultaneously disengage one drive system and engage the other. Thereby the speed could be changed without either belt having to cope with pulleys of differing diameter.
Continuously variable speed change by shifting the belt on opposed matching conical pulleys (by 1857)
A more sophisticated approach using belts running on conical pulleys developed from the late 1850s onwards. The underlying principle is similar to that used in the DAF Variomatic automobile transmission, variants of which are still in production today. The input and output drive shafts are parallel and each carries a conical pulley, but facing in opposite directions. One or more drive belts run over the pair of pulleys, and can be moved laterally by one or more forks.
When the drive belt is at one extremity, the effective diameter of the input pulley will be large and that of the output pulley small, giving a high gear. With the belt at the other extremity, the effective diameter of the input pulley will be small and that of the output pulley large, giving a low gear. Between these two extremes a continuously variable transmission is offered, the ratios being proportional to the movement of the belt - the mid point giving a 1:1 drive.
An early and particularly interesting application of the principle is given in C Barlow’s patent 1045 of 1857. This concerned a fully automatic speed controller for a rotary blower delivering fuel gas and air to a furnace. A rubber diaphragm reacted to the pressure in the system and activated the belt fork of a dual-cone drive system to vary the speed of the blower.
The dual-cone principle was the subject of patents for the rest of the 19th century. So too was the expanding pulley principle, whereby the effective diameter of a single, relatively narrow (and usually flanged) pulley was varied according to a variety of ingenious but often complicated methods. But neither of these need detain us further - the key point is that dual-cone change-speeds established the idea of laterally shifting the drive belt at will, to achieve varying output velocity from a relatively constant speed source.
Stepped multiple pulleys using the derailing principle (1863 & 1897)
In 1863 A Rolfe obtained a provisional patent (2510) for something much more like a derailleur as we know it. According to this:
"Fast and loose pulleys on the axles of self-propelled vehicles are formed with several "concentric" grooves of different diameters, and with inclined grooves to conduct the straps or bands from one groove to another."
This is, of course, remarkably similar to the derailleur’s multiple sprocket concept - and is all the more interesting in that it relates to road vehicles rather than mill machinery.
Thirty-four years later, in 1897, Soc. Diligeon et Cie took the idea further in patent 23,803. This described a four-speed belt drive system "for automotor vehicles and the like" (see patent illustration below).
Essentially this was an opposed cone system but with the cones stepped to give four distinct speeds. The gear change lever moved in two planes:
Late Victorian developments in non-derailleur change-speed cycle gearing
We now revert to change-speed systems for cycles. Before tracing the evolution of derailleur gears, it is useful to provide a brief outline of how cycle gearing in general developed following the early work of Villepigue, Barberon & Meunier and Starley & Hillman.
Epicyclic gears - the Crypto principle (1878 onwards)
Epicyclic gearing was destined to make the biggest impact. Even today, the derailleur’s main opposition comes from epicyclic hub gears. A typical basic epicyclic system comprises:
The planet pinions are held together in a cage that revolves around the sun and within the annulus. If the sun is not free to rotate, rotation of the planet cage causes the annulus also to rotate, but faster. This is because the planet pinions pick up additional motion from their engagement with the static sun, which they then pass on to the annulus.
In theory a basic epicyclic configuration like this can give seven speeds, though in most cases only two or three are of practical use.
In the commonest modern application, a three-speed hub, the sun is locked to, or part of, the rear wheel axle. Via a system of clutches controlled by a cable linked to the selector lever, drive from the rear sprocket can be fed:
From 1878 onwards numerous patents were filed for epicyclic cycle gears. Notable among these was Thomas Shaw & William Sydenham’s of 1882. Their Crypto Dynamic two-speed was so successful that for a quarter of a century the terms "Crypto principle" and "epicyclic" were synonymous.
In 1883 London lawyer Robert Jay filed patent 2957. This covered a three-speed epicyclic gear for the rear wheel of a tricycle.
Tricycle gears and early apathy (1880s & 90s)
Much of the early interest in cycle gears related to the popularisation of the tricycle in the 1880s. In this market the Crypto Dynamic and other epicyclic gears competed with double chain drives. These latter were descendants of the old dual-belt systems and do not seem to have been adopted widely. An example was fitted to a Sparkbrook tricycle shown at the Stanley Show in London in 1887.
In 1883 a bicycle with two-speed epicyclic gear was demonstrated but aroused little interest. During the rest of the decade and well into the next, the development of the safety bicycle, the popularisation of the pneumatic tyre, and improvements in cycle chains and bearings diverted interest from change-speed gearing.
Nonetheless there were developments. For example, the Collier two-speed, a non-epicyclic gear comprising a special chainwheel/crank assembly, and requiring a special bottom bracket, was launched in 1889. Although heavy, it was still in limited use in the early years of the 20th century.
Epicyclics continue to develop (1890s)
Epicyclic gears meanwhile developed on three main fronts. Firstly they were used to gear up the front wheel of front driven bicycles. This enabled the front wheel size to be reduced, as the front drivers fought an unsuccessful rearguard action against the rear driven safeties in the early 1890s.
Secondly, epicyclic bottom bracket or chainwheel gears were introduced. The Coventry Cycle Company introduced one in 1892 and the principle was later adopted by others; for example, in the Sunbeam bracket gears produced by John Marston Ltd.
Thirdly, and most importantly, compact hub gears were introduced. It was realised that the torque in the hub of a rear driven safety bicycle (being geared up for speed rather than down for power) was much lower than in the bottom bracket. Consequently an epicyclic gear in the hub could be substantially smaller and lighter than a similar gear in the chainwheel or bracket.
In 1895 the Indiana machinist Seward Johnson patented the first compact hub gear for a rear driven safety bicycle. Less than a year later William Reilly of Salford filed patent 6062 of 1896. This was for a very similar hub gear and was the basis of the mould shattering Hub two-speed. This led in turn to the first Sturmey-Archer three-speed, launched in 1902 and based on another design by Reilly.
Expanding chainwheel gears (1884 onwards)
When reviewing industrial belt-drive systems above, we noted the interest in expandable pulleys. There was a parallel interest in expanding chainwheels for cycles.
Charles Linley, of whom more later, was involved in developing one in 1884. He was also co-inventor of the original Protean gear, an expanding chainwheel that went on sale in 1895.
In the early 1900s the Paradox expanding chainwheel gear provided no fewer than seven speeds.
The designers of these systems wished to avoid the frictional losses
inherent in spur gearing
(including epicyclics). As we shall see, this objective was to lead
Linley and others to develop
early derailleur systems.
Having seen the parallel developments in industrial change-speed systems, and in cycle change-speed systems, we have established a good background against which to set the early evolution of the derailleur. We can now review the development of those change-speed systems that relied solely on chain, chainwheel and sprockets.
Just off the evolutionary road to derailleur systems was an interesting dead-end based on "chains within chains". These were lengths of modified drive chain that formed the working surface of a chainwheel and which could be used to vary its diameter.
Although these were essentially expanding chainwheel gears rather than derailleurs, they indicate a developing interest in:
There seem to have been only two examples of the "chains within chains" approach. They were patented within 22 months of each other but differed considerably.
One of the inventors of the first system, Charles Linley, went on to design a very early true derailleur gear The second system, by Henry Holden, actually used a system of forks to derail the "chain within a chain". This was not a true derailleur system as we know it, because it did not derail the driving chain. However, it was probably the first cycle change-speed system to employ the chain derailing technique in any form.
The first of the two "chain within a chain" systems described here was the subject of patent 2547 of 1884. The three inventors were all from the Greater London area - a surgeon from Clapham, Surrey, named George Tandy, and the cycle engineers John Biggs of 9 Melior Street, Southwark, Surrey and Charles Montague Linley of 49 Dynevor Road, Stoke Newington, Middlesex. Of these, Linley was the first named and he was to make the greatest contribution of the three to cycle change-speed gearing.
The patent covered an expanding chainwheel gear in a casing comprising two discs. These were mounted face-to-face, each face having a matching spiral groove. In these grooves ran a short length of chain that formed the working surface of the chainwheel, and which we will call the inner chain. When this was as deep into the spiral as possible, it formed a chainwheel with a small diameter: when at the outer extremities of the spiral, one with a larger diameter (see patent illustrations below).
To enable the inner chain to move smoothly along the spiral grooves, its connecting pins were extended and carried rollers that ran in the grooves. This left the main width of the inner chain clear to engage with the drive chain. The latter had with arched (semi-circular) links to enable it to engage with the pins of the inner chain. Hence the drive chain somewhat resembled a toothed belt.
The outer end of the inner chain was fixed to a U-shaped bar spanning one of the discs and normally locked to it. This bar could be released by a mechanical linkage to let the chain run along the spiral grooves, thus varying the effective diameter of the chainwheel.
The patent allowed for both the chainwheel and the driven sprocket to be formed in this way, thus allowing a wider range of gearing. However, the single expanding chainwheel was preferred by the inventors, as being easier to control. Chain slack was taken up by an idler pulley or other unspecified means.
Two elements of this invention were to be used later by Charles Linley in one of the earliest true derailleurs, the New Protean. These were:
Revd. Henry Holden (1885)
The second change-speed system to be based on a "chain within a chain" was created by a Yorkshire-based Anglican priest, Henry Maguire Holden, who lived at St Bartholomew’s Vicarage, Bowling, Bradford. In the words of the official abridgment, his patent 15,228 of 1885 described how:
"The relative diameters, and hence the relative speeds of chain-wheels, are altered by means of a special chain [see patent illustration below left] which may be wrapped around either, fitting upon their toothed surfaces, and itself provided with similar teeth for engagement with the driving-chain."
The "enlarging chain" (as Holden called it) had a break in it, but was kept on the first chainwheel by guide rods, wires or channels. To change gear, a fork was activated which derailed the enlarging chain from the first chainwheel. This driving chain then transported the enlarging chain to the second chainwheel. The enlarging chain then wrapped itself round the second chainwheel, where it was again held in place by guide rods, wires or channels. The whole assemblage could conveniently be housed in a chaincase (see patent illustration above right).
So the Reverend Holden was probably the first to use a remotely activated fork to manipulate a length of bicycle chain, albeit not the driving chain as such.
The first true derailleur for cycles (1892:
John Wilson’s double chainwheel, with hinged segmented larger ring and complex chain slack compensation
More than six years were to pass before the next significant development in the evolution of the derailleur in the UK. During this time:
In February 1892 a Lancashire builder, John Wilson of Brothers Street, Livesey near Blackburn, applied for patent No. 3843. This covered a front derailleur system comprising a special double chainwheel.
The teeth of the larger chainwheel were on a series of segments, typically three teeth per segment. Each segment was mounted on a thin arm that passed between the teeth of the smaller chainwheel. The base of each arm pivoted in a groove on the smaller chainwheel (see patent illustration below left).
There was only one chainline, into or out of which the pivoted segments of the larger chainwheel could be swung by a cam actuated via a control pedal or handle. Spring catches held the segments firmly in either position.
The patent outlined how, if the rider selects lower gear, "the segments of the wheel are forced outwards moving on their rounded ends which serve as hinges, and away from the driving chain so that the latter is then driven by the smaller driving [i.e. chain] wheel." Conversely, shifting from low to high gear, "a reverse action takes place and the segments of the wheel are brought under, and in contact with the driving chain so that the latter is driven by a wheel of larger diameter."
The system was designed so that the cam moved the segments one-by-one when "not in contact with the driving chain but just previous to their coming into contact therewith". By this means the larger chainwheel "is segment by segment removed from the line of the driving chain and the latter is transferred to the smaller wheel" or vice versa.
Unfortunately this elegant gear change systemwas marred by a comparatively inelegant method of taking up the chain slack in low gear. Instead of using a jockey wheel on a spring-loaded arm, Wilson proposed a system whereby the distance between the rear wheel axle and the bottom bracket could be varied automatically to take up the slack (see patent illustration above right).
This entailed mounting the rear wheel axle on short downward-facing pivoted links. These links were attached to a fork below the chainstays. The forward end of the fork was attached by a coil spring to the bottom bracket. Rearward movement of the wheel to take up slack in low gear was provided by the coil spring: forward movement in high gear, by the drive chain. A system of automatic catches held the fork in either the forward or rearward position, thus stopping "hunting" of the wheelbase and consequent damage to the system components.
Although Wilson’s segmented chainwheel could have been produced as a component suitable for a wide range of bicycles, both as original and after-market equipment, his chain-tensioning system implied a purpose-built frame. It also introduced many points of wear, reduced rigidity of the frame’s rear triangle and a fluctuating wheelbase. It is, perhaps, little wonder that it failed to make much impact.
However, it does appear to have been the first true derailleur for a cycle, in that it shifted the drive chain from one chainwheel to another of different diameter by remote control.
Multiple chainwheel and sprocket systems without remote control
Bernard Cadot’s three/four-speed (1893)
Eighteen months after Wilson’s patent came No. 15,838 of 1893 - a sign of Gallic interest, as it was in the name of Bernard Cadot, a retired civil engineer of 27 Rue Caumartin, Paris.
Cadot’s system involved a configuration familiar enough (at first sight) to modern cyclists - a multiple chainwheel with multiple sprockets, typically three of each. A nine-speed derailleur, you might think.
But no, Cadot was merely echoing the old principle of two or more parallel drive systems, each offering a different speed - and without remote control. To change speed, Cadot proposed a chain "capable of being instantaneously opened by the employment of suitable links". In other words, the rider had to dismount, unclip the chain, move it to another chainline (and a different combination of chainwheel and sprocket) and clip the chain together again.
The system thus typically offered three speeds. As there was no device to take up chain slack, the sum of the number of teeth of each combination of chainwheel and sprocket had to be the same. Interestingly, showing perhaps the influence of industrial continuously variable opposed cone belt-drive systems and stepped pulleys, Cadot referred to his multiple chainwheel and sprocket clusters as "toothed cones" or "stepped wheels" (see patent illustration below).
More than two years later, a patent by the Friedlein Company (No.21,104 of 1895) also employed a special "readily detachable" link. This design had only two sprockets and a double chainwheel (see patent illustration below).
William Weatherill’s four-speed (1894)
In March 1894, little more than six months after Cadot’s patent, William Weatherill, a coach body maker of Cheetham, Manchester applied for patent No.4770. This again used a multiple chainwheel and multiple sprocket, and also required moving the chain from one pair to another.
It differed from Cadot’s in that:
But, as with Cadot’s design, it was still necessary to dismount and tinker with the machine to change gear (see patent illustration below).
Various patents concerning taking up chain slack
The idea of multiple sprockets and multiple chainwheels to give two, three or four speeds with a true chainline featured in a number of patents of the mid 1890s. It was obviously attractive. It offered no loss of efficiency compared with a single-speed transmission and relatively light weight.
The need to dismount and manually reposition the chain seems to have been accepted by inventors. What occupied their minds most was how to take up chain slack.
Hence we find the Austrian railway employee Hugo Exner’s patent 6685 of 1895. This offered an alternative way of realigning the chain in a three-speed system otherwise on the lines set out by Cadot. Exner’s main idea was a quick-release eccentric mounting for the multiple chainwheel. Operating the device slackened the chain, obviating both the need to use a special chain link and the need to reposition the rear wheel in its dropouts (see patent illustration below).
W M Woodland’s patent 9230 of 1895 dealt with a transmission in which "Two or more chain wheels of different sizes are fixed upon the pedal crank axle and driving wheel respectively, being geared by a single chain." The essence of this patent was a long slotted rear wheel dropout. This, via a quick-release adjustment, enabled the chain to be quickly slackened, placed on a different (but still in line) combination of sprocket and chainwheel, then retightened.
Woodland’s patent also offered an alternative to the long slotted dropout. In this the wheel axle was "carried by a pair of bent pivoted levers which enable the chain to be slackened" (see patent illustration below).
The sprung arm solution - Linley, Biggs & Archer (1894)
But the most elegant solution to the chain slack problem was incorporated into a patent filed the previous autumn - 17,908 of 1894. It was the work of engineers Charles Linley, John Biggs and John Archer, then based at the 57 High Street, Clapham, Surrey (now part of London).
Ten years earlier Linley and Biggs had patented the expanding chainwheel gear that used a "chain within a chain" (see above). This new patent also dealt with an expanding chainwheel, but this time it comprised two half chainwheels, connected by a hinge at the periphery. For lowest gear the chainwheel was fully contracted and circular: it became "somewhat oval" as higher gearing was selected (see patent illustration below left).
Inside the chainwheel assembly were two eccentrics, one connected to each half of the chainwheel. These were controlled by a fulcrum disc that, in different interpretations of the invention, could be operated by back-pedalling, or by hand lever or by stopping the machine and using a locking screw.
But the aspect of most relevance to the development of the derailleur was the means of automatically taking up chain slack (see catalogue illustration above right):
"As the wheel expands and contracts the chain will become tight and loose and we therefore provide pulleys or rollers mounted on spring levers to take up the slack of the chain. The particular arrangement we prefer is a lever fulcrated near its centre with a roller at each end and the chain passing over one roller and under the other. We use a spring with this arrangement so as to keep a constant tension on the chain."
This was a more specific refinement of Linley and Biggs’ solution a decade earlier to the same problem, cited above in respect of patent 2547 of 1884. In that earlier case, chain slack was "taken up by any convenient means, such, for instance, as a tension spring or a weighted lever pivotted (sic) to the frame and carrying a pulley or roller round which the chain passes."
The new gear was marketed as the Protean four-speed and was still being advertised as late as September 1899. Users included the Dukes of Connaught and Portland, the Marquis of Winchester, half-a-dozen earls and a number of other titled folk.
Multiple chainwheel/sprocket systems with remote control - designs from overseas
From France - a true derailleur with out-of-line chain drive- Jean Loubeyre’s Polycelere (1895)
The Polycelere (i.e. Multispeed) was an important development in the evolution of the derailleur. Although no British patent seems to have been granted for it, the gear was patented in France. Its inventor was Jean Loubeyre of Paris and the device was advertised in the 1896 catalogue of the Compagnie Generale des Cycles. However, no surviving example has been traced.
The Polycelere had two sprockets with 8 and 10 teeth. (Chains at this time were typically 1" (25.4 mm) pitch. The equivalent sprocket sizes for later 1/2" (12.7mm) pitch chain would be 16 and 20 teeth.). A rod-operated fork pushed the chain sideways to induce it to move from one sprocket to the other.
It is most significant that there was no provision for truing the chainline. The Polycelere relied on the ability of the chain to function satisfactorily whilst running out of line. This is a key characteristic of most later derailleurs and all modern ones; but was not exploited in the early British designs, which went to great lengths to ensure that the chainline always remained true.
The Polycelere did not incorporate a freewheel mechanism. Therefore, because the sprockets were fixed, sprung arms to maintain chain tension were needed on both the upper and lower runs of the chain. The arms were pivoted from just behind the chainwheel.
From the USA - almost a derailleur - Weed & Gridley’s two-speed (1896)
In August 1896, Harry de Lyne Weed and Francis Webster Gridley, both of Syracuse, State of New York, applied for patent 17,779. This declared:
"Our invention relates to chain-driven cycles and particularly to mechanisms for changing the speed thereof, from high to low speed or vice-a-versa all being done by the rider without stopping the machine or removing the feet from the pedals."
This system had a single chainwheel and two sprockets, one significantly larger than the other. The smaller sprocket was fixed to the rear wheel hub. The larger sprocket was mounted alongside it but on double eccentrics. These were controlled by the rider via a rod linkage, and enabled the larger sprocket to be positioned either concentric with, or eccentric to, the hub.
The chain was sufficiently wide that it ran over both sprockets at the same time.
For high gear, the larger sprocket was placed in the eccentric position. It ran free on the eccentrics, merely acting to take up any chain slack that would have resulted from using the small sprocket alone.
For low gear, the larger sprocket was made concentric with the hub. In moving to this position, it also moved sideways slightly - though not enough to significantly alter the chainline. This movement locked the small sprocket into recesses in the larger and provided the driving connection (see patent illustration below).
Hence, Weed and Gridley had overcome (at least in theory) the problems of:
However, as the chain never left either sprocket, we cannot class this gear as a derailleur.
From Germany - Salomon Frank’s derailleur with movable guide pulleys and the option of out-of-line chain drive(1897)
In June 1897, the German mechanical engineer Salomon Frank of Frankfurt am Main filed patent 13,533. This described in outline a number of derailleur systems in which chain tension was maintained by one or more guide pulleys or jockey wheels on a frame ahead of the chainwheel or behind the sprockets (see patent illustration below).
In a typical configuration, the jockey wheel frame had two widely spaced pulleys ahead of a multiple chainwheel. The configuration of the jockey wheels allowed the chain merely to engage with the top of the chainwheel. By a rod linkage or other suitable means, the jockey wheel frame could be moved in an arc, the focus of which was the rear wheel axle.
Upward movement of the frame lifted the chain clear of the selected chainwheel, and allowed the chainwheel cluster to be moved laterally. The frame would then be lowered, bringing the chain into contact with a different chainwheel. Gear changing could be effected either stationary or with the machine in motion.
Significantly, Frank noted the possibility of leaving the multiple chainwheel (or sprocket) cluster laterally static, and moving the jockey wheel frame sideways to reposition the chain on a different chainwheel (or sprocket). Hence, like Loubeyre, he accepted, at least in principle, the idea of the chain running out-of-line.
Multiple chainwheel/sprocket systems with remote control - the two key British designs
There is little or no evidence that any of the early derailleurs referred to above, apart from the Polycelere, were actually marketed or even promoted actively. However, two British designs received a good deal of attention about the turn of the century.
Charles Linley’s New Protean or Whippet two-speed gear was advertised in the cycling press. At least a few were sold. Edmund Hodgkinson’s Gradient does not seem to have been advertised but was ridden and reviewed.
Edmund Hodgkinson’s Gradient three-speed derailleur (1896)
In January 1896, soon after Loubeyre’s Polycelere patent but before either Weed & Gridley or Frank had filed their specifications, Edmund Hugh Hodgkinson, "of 8 Lancaster Gate, in the County of Middlesex [west London], Gentleman" filed patent 1570 of 1896. This formed the basis of the Gradient three-speed derailleur, which was launched nearly four years later.
The patent described a gear in which "For the purpose of altering the speed of the gearing two or more pinions [sprockets] of different sizes are provided and arrangements made for shifting the driving chain from pinion to pinion."
The sprockets were mounted on a tubular spindle, which drove the rear hub via sprung pawls, thereby allowing free-wheeling and gear changing whilst in motion. (The patent also allowed for provision of a clamping screw so that the gear could be locked into a selected fixed gear.) The sprockets could be "slidden along the spindle", the driving chain having first been lifted off to put them out of gear. In the patent drawings the sprockets were arranged in the reverse order to that which is normal today; the smallest being nearest the centre of the hub and the largest nearest the chainstay. This allowed the smallest sprocket to slide into the hub shell when the higher gears were selected - thus eliminating the need for an excessively long axle and unusually wide "rear jaw" for the frame. (The right-hand drop-out had nonetheless to be set out 1 5/8" (41mm) from the standard width.)
The chain used was of conventional design and maintained a true line, regardless of gear selected. A jockey pulley on a sprung arm was used to take up chain slack, as in Charles Linley’s original Protean gear.
The idea of the rider being able to change gear whilst riding was still sufficiently novel for the patent to state that "sometimes" the mechanism was "so arranged that the change of gear can be made whilst riding ..." This was done by "the movement of a lever and a slider thereon". The lever moved the sprocket cluster laterally: the slider actuated the chain lifter (see patent illustration below).
There were two chain lifters, which pivoted around the rear wheel axle, outboard of the sprocket cluster but inboard of the drop-out. The chain-lifters comprised "short-armed inturned sectors, having each outwardly-flanged segments". In their rest position, they both faced straight ahead, between the upper and lower chain runs. If the rider pulled up the slider control, the chain-lifters rotated backwards (anti-clockwise viewed from the chain side) from their rest position. When the flanged segments made "gentle contact" with underside of the chain, the rider back-pedalled. The chain then carried the segments with it (one through a total travel of about 110%, the other about 220%) thus lifting the chain clear of the sprockets.
"Thus by means of the slider first the chain lifter is made to set the pinions free, and then by shifting the lever the pinions are placed as required bringing either a larger or smaller one into position for use, and finally the slider is placed in another notch and returned to the lower end of the lever; the chain then regains its working position."
Over the next five years Edmund Hodgkinson continued to develop the design of his gear. In July 1897 he filed patent 16,715 which refined the chain lifting mechanism. Further refinements to the chain lifter were covered by patent 9192 of May 1901 (see patent illustration below).
In March 1899 he applied for patent 5317 which improved the automatic chain tensioning system. The jockey wheel arm now pivoted not from at or near the rear wheel axle, but from below the bottom bracket (see patent ilustration below). With this revised geometry:
"... the greatest pressure can be applied to the jockey wheel when there is the greatest amount of slack in the chain; that is, when the smallest pinion is in use. This arrangement tends to prevent the chain jumping cogs: as it is most likely to occur when the chain is on the smallest pinion."
Hodgkinson’s Gradient derailleur was launched at the annual Stanley Show in London, late in 1899. The CTC Gazette described it as "the most thaumaturgic [wonder working] feature of the Show". But otherwise it received little coverage until the summer of 1901.
Reaction to Hodgkinson’s Gradient and comments from users
C W Brown, writing in The CTC Gazette, for July 1901, stated that "In practice the gear works very well, and changes beautifully .... He added the important point that "it has the great advantage that, save for the idle wheel, no more parts are at work with any one of the gearings than would be the case of the average [single-speed] machine."
Brown pointed out that the rider could have gears of 88, 68 and 51". This equates to an increase of 29.4% over middle gear, and a decrease of 25%
In August 1901 The Cyclist was "able fully to describe and detail Hodgkinson’s patent gradient three-speed gear as it has never been shown before." The magazine revealed that Harry J Swindley had been riding a bicycle fitted with the Gradient for some time, and that he thought highly of it. However, the writer added:
"No one, not even the inventor, will claim that it improves the appearance of the bicycle. It does not. The chain-lifting and striking gear give a suggestion of rigging about the frame."
The sprocket cluster of the gear described in The Cyclist had the smallest sprocket inboard, the medium-sized sprocket outboard, and the largest sprocket in between.
Despite the apparent complexity of the gear changing action (involving back-pedalling and co-ordinating movements of the slider controlling the chain-lifters and the lever that moved the sprockets laterally), the magazine stated that "the movements are quickly and sweetly performed after a very little practice." Indeed, the reviewer had "a considerable opinion" of the Gradient:
"It will undoubtedly prove most useful on motor bicycles, as well as on pedal-propelled machines, for with the former it is scarcely possible to have the low gear too low, or the high combination too high."
A couple of months later, in October 1901, Cycling mentioned the Gradient in passing:
"We have had no practical experience with it, and it has been little pushed. Two riders of our acquaintance who have used the gear have spoken well of it."
In January 1902’s issue of The CTC Gazette, Archibald Sharp reviewing the recent 1901 Stanley Show, reported that the Gradient’s changing mechanism had been much improved.
By October 1903, nearly four years after its launch, the Gradient was illustrated in a major feature in Cycling on variable speed gears, but was not described in the text. It seems never to have been advertised to consumers, and the following year Hodgkinson sold the design to Terot et Compagnie of Dijon. They went on to produce a similar gear, shown at the Paris Motor and Cycle Show in 1907.
Linley & Biggs and The Whippet Cycle Syndicate Ltd.
In 1899, at the same Stanley Show that saw the launch of the Gradient, another British derailleur was revealed to the world. This was the New Protean or Whippet gear, designed by C M Linley. Before describing the gear itself, a little background on Linley and his company.
The inventive Charles Montague Linley was involved in patenting not only the two expanding chainwheel gears referred to previously, but also a number of other inventions. These included two freewheel mechanisms (19,280 of 1897 & 18,711 of 1898), a back-pedal rim brake (6302 of 1898) and, with Charles Sangster, a variable speed gear for motor cars and industrial machinery (13,705 of 1898).
Linley & Biggs launched the Whippet solid-tyred sprung frame rear driving safety bicycle about 1886. It was made at their works at 29 Clerkenwell Road, London, along with a twin front wheel ("double steerer") tricycle version and a "convertible" single front wheel ("direct steerer") tricycle version, which could be transformed into a bicycle.
From time to time the company issued free booklets, promoting cycling in general and their products in particular. Their 1888 booklet, Practical Hints on Cycling, dealt with riding position, pedalling, hill climbing, tips for beginners, machine care, touring, food and drink, advice for lady riders and on clothing. Needless to say it also included advertisements for the Whippet range of cycles.
By the late 1890s Whippet cycles were mostly pneumatic-tyred, unsprung diamond-frame and loop-frame machines. The company operated as the Whippet Cycle Syndicate Ltd, from showrooms in London near Oxford Circus at 281 Oxford Street. Manufacture was at the Whippet Works, Falcon Terrace, Clapham Junction in south west London.
W Chater Lea, founder of the cycle fittings company that bore his name, started his engineering career at the Whippet Works. Riding a Whippet, he won the Catford hill-climb in 1888, 1889 and 1891.
The 1899 advertising focused on the Whippet freewheel, the back-pedal rim brake and the Protean four-speed expanding chainwheel. That is, until December, when the syndicate launched the New Protean two-speed derailleur.
The New Protean or Whippet two-speed derailleur (1899)
The patent for the New Protean had been filed three months earlier (18,240 of 1899) and was in Linley’s name alone. To describe the gear one can do no better than to quote from the advertisement that appeared in The CTC Gazette of December 1899. For clarity, the originally continuous text has been split into paragraphs. The author’s comments are in squared brackets:
"Fig. 1 shows how the alteration of gear is obtained. On the back hub are placed two chain wheels [rear sprockets] fixed together (fig. 2) and mounted on a free wheel. The chain which is of special construction [having arched links and somewhat resembling a toothed belt] is capable of running on either of these wheels, and is shifted from one to the other by means of a lever (not shown in the accompanying illustration).
"The small idle pulley is merely for the purpose of keeping the slack chain that occurs when the high gear is in use from rattling about.
"Fig.2 shows the essential feature of the invention. The side of the larger chain wheel at the point marked A has a projection or step.
"A fork [not shown in fig. 1] embraces the chain to determine its position, without touching it except when the alteration is being made. This fork is capable of sliding sideways; that is to say, it can guide the chain on to either the large or the small wheel, and is operated by a lever in a convenient position to the hand.
"When the chain is on the smaller wheel, giving the high gear, and the rider desires the low gear, he moves the lever whilst he is riding, which guides the chain towards the large wheel. When the step A comes into the opening of the chain, it engages the part of the chain that is coming on to the wheel, and lifts it up without the slightest jerk or sound, on to the large diameter, and there it remains until the rider wishes to change again to the high gear.
"When the change is made from the low to the high gear, the fork is moved in the opposite direction, guiding the chain towards the small chain wheel. The chain is of such construction that it will drop down and immediately engage with the teeth on the small wheel, also without the slightest jerk or sound.
"No one who has not seen it work would credit the ease and smoothness with which these operations can be performed, at any speed or under any conditions.
"The front chain wheel is of double width so as to allow the chain to find its own position and maintain an absolutely true alignment."
So, in this invention, Charles Linley brought together:
In so doing he produced a change-speed gear that embodied most of the key principles of today’s derailleurs. That is, apart from the idea of running the chain out of line, which Loubeyre in France and Frank in Germany had proposed.
The gear appears typically to have been supplied with sprockets either of:
Reaction to the New Protean/Whippet and comments from users - 1899-1903
Archibald Sharp, author of the definitive work on bicycles and tricycles of the period, mentioned the New Protean gear in his report "Novelties from the Shows" in The CTC Gazette of December 1899. He observed:
"The chain is moved laterally by a fork, somewhat in the same manner as the shifting of a belt from the loose to the driving pulley of a lathe or machine tool is effected."
He also noted:
"The ratio of the two speeds, depending on the numbers of teeth on the two hub chainwheels, can be chosen over a considerable range, whereas in the old Protean gear with four speeds, the extreme ratio was 18:21, i.e. 100:117. Personally, I think I am best suited for the roads to the south and west of London by a ratio between 100:120 and 100:125. Faster riders will probably be best suited with a smaller ratio, ladies with a larger."
Six months after its launch, Sharp reviewed the gear in greater detail. He commented:
"The chain is, of course, of special design, and must be heavier than block or roller chains of equal strength. The [New] Protean gear introduces no additional rubbing surfaces to waste power in friction; there is little or no pressure on the jockey pulley, so its frictional resistance is a negligible quantity. On the other hand the chain shown at the recent Stanley Show is scarcely up to the prevailing high standard of block and roller chains; but no doubt improvements can be made in its manufacture."
Sharp had not at this time used the gear. In the autumn of the following year, 1901, Cycling described the New Protean as "a practical article". Their writer’s only quibble was with the width of the ratios typically offered. He would have liked the difference to be less than half that normally offered.
About the same time several riders went into print in The Cyclist> magazine, expressing praise for the New Protean.
A F C Phillips rode a Whippet cycle fitted with the gear (sometimes known as the Whippet gear). In reply to a critic of two-speed gears, Phillips wrote:
"Possibly, Mr Walker has never had an opportunity of trying a two-speed gear in which the friction of the bicycle is normal at both gears; that is to say, where no extra friction (such as that caused by internal cog wheels) exists. The extra friction in most of the forms of two-speed gears of the past was particularly undesirable. ... There are two points in two-speed gears which I consider a sine quo non - first, no extra friction whatever at either gear; second, ability to change gear while pedalling without loss of tie and consequent loss of momentum."
J M Wellington, a mechanical engineer, wrote that:
"The ‘New Protean’ gear seems to me to leave nothing to be desired in a two-speed gear. The friction is normal at both gears, there are no extra parts at work except the jockey (which is quite negligible), there is no limit to the range, the low gear is produced by throwing the chain into a larger sprocket on the hub (in the correct way it should be, and not by reducing the crank wheel), and, most important of all, the changes can be made while pedalling, and without the slightest loss of momentum."
Chas T Cowley reported covering "upwards of 6,000 miles" on his Whippet cycle with New Protean gear and had recently taken delivery of a new machine with the same gear. He had found "when riding with friends that the ability to change to the lower gear whilst ascending a hill gave me a great advantage, and enabled me to reach the top an easy winner." He praised the lack of friction compared with a Crypto epicyclic gear he had used with a tricycle, and added:
"... the ability to change to the higher gear whilst riding, so as to take advantage of favourable winds and level roads, is so great that after thirty years of cycling I find that I am now able to travel greater distances and at a faster rate than in my younger days."
But late in November 1902, "The Magpie", writing in Cycling, had a dig at the New Protean and Charles Linley:
"The Protean is admittedly a bold device, but it has not taken the fancy of the public to a tithe of the extent which the Hub two-speed gear has done, and, after all, an inventor should seek to please the public rather than to gratify the dictates of his own genius."
Interest in change-speed gears for cycles grew significantly in 1903 following the launch late in the previous year of the first Sturmey-Archer three-speed hub. Production difficulties dogged the new gear, but by the summer it was becoming available. The press followed it with great interest.
In June 1903,"Spiral", writing in Cycling produced a matrix comparing five gears already on the market with the soon to arrive Sturmey-Archer. He awarded points on a scale of 0 to 10 (10 being "perfection") for adaptability, gearing range, neatness and weight. Highest scoring was the Hub two-speed, the first commercially successful compact hub gear, which gained 35 points. Second was the Sturmey-Archer three-speed hub (designed by William Reilly who also designed the Hub) with 32 points.
By contrast the Whippet came last in this race, scoring just 19. Because of its special chain, chainwheel and sprockets it got zero for adaptability. "Spiral" added:
"A dangling jockey chain has an insecure look about it. It looks as though it would easily slip off. Looks count. Riders never like to have anything uncommon about them, to attract the small boy’s attention."
In October 1903 Cycling ran an in-depth feature on variable speed gears. This dealt first with the arguments in favour of change-speed gears, including the physiological and psychological aspects. The writer next covered the history of change-speed gear development, then described in detail the main systems then available.
He was unable to quote a price for the New Protean and made no comment on its merits other than to note its relative lightness.
The National Cyclists’ Union Review for 1903 also included a comprehensive survey of gears then on the market. Walter Phillips wrote:
"It must be remembered that there is only one tooth of either chain or hub wheel in an ordinary cycle that is really doing all the work. This fact has been recognised in the Whippet gear, and advantage taken of it to do the required shunting and lifting up of the chain from the small to the large ring."
He commented on the "sweetness" of the shifting action and added (in a particularly long sentence but nonetheless lucid sentence):
"At first sight the gear looks rather clumsy and cumbersome; on further investigation it is only the shape of the chain which gives it this appearance, but it is necessary to adopt this form of chain to carry out what is required of it, viz. the shifting sideways on and off the teeth: it is in reality a toothed belt instead of an orthodox chain, and it should be considered as such."
However, just as with Hodgkinson’s Gradient, after 1903 references in the cycling press to the New Protean ceased abruptly.
It is often said that history is written by the winners, reflecting their prejudices. Occasionally this is true of cycle history - especially where a commercial interest or personal pride is concerned.
In this context, the losers were the early British pioneers of derailleur gears, and the winners, the protagonists of compact epicyclic hub gears. The demise of the early British derailleurs resulted from the huge and rapid success of the three-speed hub gear, and in particular, the Sturmey-Archer.
Therefore two reviews of variable gearing are noteworthy, both written by Henry Sturmey but 22 years apart.
In May 1902 Henry Sturmey, who had recently invented a three-speed hub gear, gave a paper on Variable Gearing at a meeting of the Cycle Engineers’ Institute in Coventry. Apart from demonstrating his own prototype, he reviewed in detail all the major change-speed systems then on the British market. These were the Collier two-speed bracket gear, the Garrard two-speed screw-on hub gear, the Hub two-speed hub, the Paradox seven-speed expanding chainwheel, and the New Protean and Gradient derailleurs. He commented on the lightness of the New Protean and even demonstrated a bicycle fitted with the Gradient gear, which Edmund Hodgkinson himself had lent him.
Twenty-two years later, in 1924, Sturmey (long since having parted company with Sturmey-Archer) was again promoting a hub of his own design. This time it was a single-cable five-speed, which sadly never went into production. In an article in Cycling he reviewed in outline the "origin, development and future" of variable gearing. In this he made not a single mention of derailleur gears, other than this:
"In France, especially in the Alpine districts, quite a number of somewhat quaint and crude devices are in use for the purpose of giving the rider more gears - in one case as many as seven - and an annual reliability trial is held over a mountainous route, mainly for the purpose of testing these gears."
He made this reference only to support his case that there could be a demand for more than three speeds. His own view was clear:
"The only mechanically correct place for a gear in a pedal cycle is the hub ..."
Today the derailleur dominates in many markets. The pioneering work of Linley, Hodgkinson and the other British pioneers of the derailleur was prophetic. And, as we all know, prophets are rarely acknowledged in their own lands. But in France .....
References
Excluding British patents cited in the text.
Acknowledgments
The author records his thanks for the considerable assistance provided by:
©Tony Hadland
Researched 1996, published February 1999.
