The Most Powerful Idea in the World

by William Rosen

  Boulton’s strategic plan was on schedule. He had discerned an opportunity; had exploited it;* and was soon enough unsatisfied by it. By the end of 1782, he had identified the next conquest for the steam engine, an arena whose potential dwarfed that of the mining industry: wheels.

  IT IS NO ACCIDENT that “wheels of industry” is such a cliché description of a manufacturing economy, since the application of force in the form of rotational motion is by far the most important component of useful work. In late eighteenth-century Britain, the wheels that mattered most were the ones turning the mills that ground the nation’s grain, and the ones that spun the nation’s cloth. Most of them used water; some used wind. None used steam. In December 1782, Boulton wrote to his partner announcing his plan to change that: “I think that these mills represent31 a field that is endless, and that will be more permanent than these transient mines.”

  He was scarcely the first to imagine the potential of a factory powered by steam. England’s manufacturers had been pestering steam engine makers on the subject for years, hoping to be freed from the shackles that bound them to the flow of rivers or wind. Unfortunately, the machines on offer were piston drivers, and using a piston to run a mill was as sensible as driving a cart by hitting it in the backside with a sledgehammer.

  One idea for using a steam-driven piston to produce rotation was nearly a century old: converting linear motion into rotation by taking the new machine and using it to run a familiar one, a waterwheel. Sixty years earlier, Thomas Savery had proposed using the water pumped from mines by the steam engine to operate a 36-foot-diameter waterwheel, housed in a mill house along with the pumping engine. In the 1770s, Boulton & Watt borrowed Savery’s idea and recommended using steam engines to pump water not just out of deep mines, but into a reservoir sited well above the engine. Gravity could then pull the water past a waterwheel and so deliver rotational work, and some early steam-powered factories used just such a system, despite its inherent inefficiencies.

  And so the challenge of converting the reciprocating motion of the early atmospheric engines into rotary power occupied a fair chunk of the eighteenth century. The fundamental problem of direct conversion was not ignorance; the crank and cam were well known to Newcomen and his successors. In fact, they were known to his predecessors. The teardrop-shaped cam was in use as far back as first-century Greece, and in 1783, John Wilkinson linked a steam engine to his forging hammer by means of a cam and succeeded in raising, and then dropping, the eight-hundred-pound tool. Work that can be transmitted by a cam, however, is the opposite of regular. The transfer of motion from one plane to another—converting the straight-line motion that was intrinsic to any piston-operated engine into the regular power supplied by a rotating wheel—turned out to be as big a challenge as building the steam engine itself. And as important. In the words of the twentieth-century critic Lewis Mumford, “The technical advance which characterizes32 specifically the modern age is that from reciprocating to rotary motions.”

  To understand this statement requires (forgive the pun) circling back to the branch of mechanics that describes the transformation of motion from one direction to another, otherwise known as kinematics. In fact, understanding the specific transformation of straight-line into rotary motion requires detours into biomechanics, and even developmental psychology.

  IN 1897, THE AUSTRIAN physicist and polymath Ernst Mach (one of the inspirational lights behind the same Gestalt theories so beloved of A. P. Usher) took a vacation from experimenting with the behavior of sound waves to examine the phenomenon of rotary motion. Mach was intrigued by the fact that Neolithic hand querns—really nothing more than two stones between which grain and other substances were ground—used reciprocating, not rotary, motion, with a horizontal handle, for millennia before evolving into rotary querns and finally rotary grindstones, which are really just querns with the upper stone turned ninety degrees. Mach, despite his knowledge of the relatively simple mathematical equations that describe rotary movement, was nonetheless confounded by his observation that infants find rotary motion nearly incomprehensible. In the words of the great medievalist Lynn White, “continuous rotary motion33 is typical of inorganic matter, whereas reciprocating motion is the sole form of movement found in living things … in Europe, at least, crank motion—a kinetic invention more difficult than we can easily conceive—was invented before the crank.”

  Crank motion was tricky enough; but the first true crank-operated machine, the rotary grindstone, has to be, like the steam engine, one of history’s most embarrassingly long-gestating inventions. By the beginning of the third century, cranks were being used in China, typically for winding silk filaments. Arabs were using them34 as surgical drills for trepanning skulls perhaps six centuries later. But it wasn’t until the 1420s that cranks and crankshafts were starting to appear in Europe, in the form of the carpenter’s bit and brace; shortly thereafter, the windlass was being used to wind crossbows, and in 1430, an unknown German machinist applied double compound cranks and connecting rods—effectively an extension, or replacement, of the human arm.

  The significance of the first combination of the crank with a connecting rod, of course, is that it turned rotary into reciprocal motion, and was therefore the basis of early wind and water power. The earliest visual evidence of a crankshaft35 connected to a wheel is a drawing by the fifteenth-century painter Pisanello, on display at the Louvre, that shows a water-driven piston pump in front of a crenellated tower, driving two simple cranks and two connecting rods.

  Pisanello’s pump, however, was never built (though unlike so many of Leonardo’s inventions, this one would probably have worked). It wasn’t until December 1593 that a Dutch farmer named Cornelis Corneliszoon not only built a wind-powered sawmill, but, because the United Provinces had an even better developed property law than Britain, secured a patent on it lasting fourteen years.* Even better, four years later he extended the patent to embrace the use of a crankshaft that turned the rotational motion of the windmill into the reciprocating motion needed to saw wood; though because it worked in reverse, the linkage used by Corneliszoon’s sawmill (which he named Het Juffertje, or “The Little Miss”), the first of more than one thousand industrial windmills in the Zaan, is not technically a crank, but a pitman.

  Reversing the process—getting a piston to drive a wheel rather than the other way around—is a trivial enough task. Attach the crank (or crankshaft: a rod connected to a wheel by a right-angled arm that pivots with the rotation of the wheel) or cam directly to the piston. But getting that wheel to rotate at a constant speed was a different matter, and any useful mechanical solution required something that smoothed out the variability in the push-pull motion of a piston—something that eliminated the “dead spot” when the piston reversed direction.

  Rotating disks had been used to smooth out those dead spots—to store kinetic energy—since Neolithic times. Several millennia later, the equations of Kepler and Newton explained why: Angular momentum tends to be conserved. This is the same phenomenon that makes an ice skater spin faster when her arms are drawn in to her body, since her angular momentum is the product of her moment of inertia and angular velocity; bringing her arms closer to her axis of rotation reduces her moment, which means her velocity must increase.

  A disk rotating around an axle—better known as a flywheel—does the same trick in reverse, keeping speed constant when power input is intermittent, as it always is with a piston or other reciprocating motion. Potter’s wheels, an early example, take the pumping motion of a hand or foot and convert it into relatively smooth rotation. The first flywheel used as part of a larger machine appears in Theophilus Presbyter’s Di divers artibus; by the fourteenth century, the Parisian philosopher John Buridan was observing that rotary grindstones store power, because they keep turning even when no one is turning them.

  It is baffling, except perhaps in light of some infantile resistance to mixing rotary with reciprocal motion, that combining the crank with the flywhee
l took centuries rather than months. Not until 1740 did an inventor named John Wise suggest using a flywheel—he called it a “double tumbling wheel” in his patent*—to produce rotary motion from a Newcomen-type engine, and that went nowhere until 1778, when the engineer and inventor Matthew Wasbrough, originally from Bristol but working in Birmingham, took the first step toward a useful rotary engine, patenting (in March 1779, number 1213—things were accelerating) a complicated assortment of pulleys, wheel segments, and flywheels (“to render the motion more regular and uniform”36).

  Two years later, Wasbrough incorporated a crank and flywheel into an iron rolling mill, but the circumstances were tainted. Watt, and his assistants at the Soho Manufactory, had been preparing a new design that would produce rotary motion from their existing—and highly successful—piston engine. They had been working on their own crankshaft model for more than a year when a workman at Soho leaked the plans to Wasbrough and his partners, John Steed and James Pickard. In August 1780, the three secured a patent (number 1263) for a beam-operated engine using the crank plus connecting rod (though how connected, the patent does not say).

  Watt had not believed37 the crank plus rod novel enough to be patentable, because of its long history. This belief didn’t do much to calm him. In most aspects of his life, James Watt was a gentle sort, but he was extraordinarily sensitive when it came to his inventions, and he was sputteringly furious at what he saw as Wasbrough’s treachery, writing to Boulton, “I know the contrivance is my own,38 and has been stolen from me by the most infamous means…. Had I esteemed [Wasbrough] a man of Ingenuity and the real inventor of the thing in question, I should not have made any objection, but when I know the contrivance is my own and has been stole [sic] from me by the most infamous means and to add to the provocation a patent surreptitiously obtained for it, I think it would be descending below the Character of a Man to be any ways aiding or assisting to him or to his pretended inventions.”

  Watt’s pride in his inventions may have been the defining characteristic of his personality. It certainly explains the amount of time he would spend in subsequent years defending his own patent claims, and frequently the claims of others, in court. The immediate reaction, however, was to take his revenge not in a courtroom, but in his workshop, and not working with a lawyer, but with another engineer, the remarkable William Murdock.

  MURDOCK—ORIGINALLY MURDOCH; the Scottish spelling was anglicized—was only twenty-two years old in 1777 when he left Ayrshire in Scotland and walked three hundred miles to Birmingham, hoping for employment at Boulton & Watt. In legend, at least, Boulton asked about the young Scot’s strange-looking hat, and when he answered that the hat was made of “timmer [timber] … turned on my little lathey [lathe],”39 he evidently decided that the young Scot’s mechanical skills were sufficient to overcome his unfortunate accent. Two years later, his employers trusted him enough that they assigned him to build, unsupervised, what was only their fourth engine. Before the year was out, he was the firm’s supervisor in Cornwall.

  Murdock’s relationship with Watt was complicated. The nearly twenty-year difference in their ages, their mutual affection and loyalty, their status as employee and employer, their shared Scottish heritage, and their technical brilliance in the same disciplines made for one of the more fraught Oedipal conflicts of the era. Given all the potential arenas for conflict, it is actually somewhat surprising that the relationship survived as long as it did, with the two apparently agreeing to annoy one another mercilessly in lieu of more bloody combat. Watt, in particular, regularly complained about Murdock’s tendency to second-guess (and, even more annoyingly, to improve) the master’s ideas. Anyone who has ever supervised a talented subordinate with a tendency to set his own priorities will find Watt’s letters familiar: “I wish William could be brought to do40 as we do, to mind the business in hand, and let such as Symington [William Symington, the builder of the Charlotte Dundas, one of the world’s first steam-engine boats] and Sadler [James Sadler, balloonist and inventor of a table steam engine] throw away their time and money, hunting shadows.”

  Even so, Watt was nothing if not fair-minded. He may have resented the time Murdock spent inventing rather than on the company’s real business, which was installing, and collecting royalties based on the continued performance of, the Watt engines.* He was, however, just as often delighted by the inventions he made on behalf of Boulton & Watt, including a compressed air pump and a cement that would bond two pieces of iron, calling Murdock “the most active man and best engine erector I ever saw.”41 His value was never, however, to be higher than when Watt enlisted his talents on what came to be known as the sun-and-planet gear.

  If the crank and its cousins are methods for turning reciprocation into rotation, then gears, broadly speaking, are methods for turning one form of rotation into another. The primary reason anyone would want to perform such a trick is the phenomenon known as the mechanical advantage, which is the formal term for the fact that when the teeth of a larger gear engage with those of a smaller, it must rotate faster for each shared revolution; since torque is defined as circumferential force multiplied by the radius, the bigger the radius, the greater the torque.

  The theory and practice of gears were well known to the Greeks of the first century BCE, who were not only familiar with several basic forms of gearing (including the familiar version, in which the teeth project radially from the gear’s center, and the worm, in which another Greek invention, the screw, was set at right angles to a traditional gear) but were also capable of incorporating them into geared engines—though the nature of those engines is revealing. The best-known surviving example, the so-called Antikythera mechanism (so called because Greek sponge fishermen discovered it off the eponymous island in 1900), contains at least eighty separate bronze gears in a wooden case not much larger than the laptop computer on which this is being written.

  Antikythera’s elaborately linked gears, probably used to calculate the positions of planets and stars on a particular date, posed an unanswered question to scholars for decades: Why was all this extraordinary precision and technological expertise—no comparable mechanism would appear until the clocks of the mid-eighteenth century—not used to produce anything that was recognizably useful?

  The answer is related to the similar question that might be asked about Hero of Alexandria: Antikythera’s creators were the ancestors of toymakers (or, at best, clockmakers), not engineers.

  Technical skill is artifact-neutral, and is just as likely to be applied to toys as tools. The builders of the Antikythera gears, like Hero, or even the great clockmakers of the Middle Ages, were just as mechanically adept as Watt or Newcomen, but they responded to a different set of incentives. Those incentives were the satisfaction of an elite segment of society with the means to commission astonishingly intricate clocks and calculators or pyramids and cathedrals. By their very nature, devices like Antikythera were built to satisfy a very different set of demands than those of a mill owner shopping for a steam engine.

  Eighteen centuries later, inventors were responding to a new set of incentives—which might be the best way of explaining the Industrial Revolution. The laws of supply and demand created the market for steam-powered mills. The laws of mechanics limited the number of ways in which the back-and-forth movement of a piston could be translated into the smooth rotation of a wheel. And Britain’s property laws excluded the best one: the crankshaft. Prohibited from copying Wasbrough’s crank, Watt and Murdock nonetheless had to compete with it.

  And not just compete—preempt. Whether or not Pickard and Wasbrough were guilty of theft, they had certainly convinced Watt that he was a victim of it, and he was determined not to be one again. So determined, in fact, that the patent application he and Murdock submitted in 1781 included five separate inventions, including a “swashplate” that used the engine to raise a beam through an arc that looks like nothing so much as an amusement park ride, like the pirate ship that swings high enough for its passengers
to go from nearly vertical to horizontal to vertical again; a counterweighted crank wheel, with the wheel divided along a diameter with one half heavily unbalanced; and an “eccentric” wheel inside an external yoke. Most important was Murdock’s sun-and-planet gear, which linked a connecting rod to one gear that made an orbit around another, larger gear like a planet circling the sun. The last one not only smoothed out the power curve, but “gave the additional advantage42 that the output shaft rotated at twice engine speed.”*

  The application, for patent number 1306, read in part, “for certain new methods43 of applying the vibrating or reciprocating action of steam or fire engines, to produce a continued rotative or circular motion round an axis or centre, and thereby to give motion to the wheels or mills or other machines.”† It was submitted on October 25, 1781, but the law allowed four months to create the final specifications. Watt needed every day. Predictably, his perfectionism was once again his greatest asset and biggest liability; he regularly complained about both his health, and the quality of his assistants’ work; to Boulton he wrote, “I wish you could supply me with a draughtsman44 of abilities [as] I tremble at the thought of making a complete set of drawings…. I must drag on a miserable existence the best way I can” even though afflicted with “backache, headache, and lowness of spirits….”

  But he also produced his most important patent specification since the separate condensing grant of 1769, “the neatest drawing I had ever made.”45