Count Rumford's Cannon-Boring Experiment (Heat as Motion)
In the 1790s, the military arsenal in Munich was run by one of the strangest figures in the history of science: an American-born former loyalist officer named Benjamin Thompson, who by then held the title Count Rumford of the Holy Roman Empire. While supervising the boring of cannon, he was struck by how hot the brass and its metal chips became. He turned that workshop curiosity into an experiment. Using a deliberately blunt borer pressed against a brass casting submerged in a box of water, and with the machinery turned by horses, he brought the water to a boil in about two and a half hours — without any fire.
Rumford reported the work to the Royal Society in London in a paper titled "An Inquiry concerning the Source of the Heat which is excited by Friction," published in the Philosophical Transactions in 1798. His conclusion was blunt: whatever an isolated body can keep supplying without limit cannot be a material substance, and the only thing being supplied to the cannon was motion. Heat, he argued, must be a kind of motion.
He was right, but the scientific world did not fully agree for another half century. The story of why is as instructive as the experiment itself.
The caloric theory Rumford was fighting
At the end of the eighteenth century, the leading theory treated heat as a substance. Antoine Lavoisier and his colleagues gave it a name, calorique, in their 1787 reform of chemical nomenclature, and Lavoisier listed caloric among the elements in his 1789 Traité élémentaire de chimie. Our article on Lavoisier's combustion experiments shows how central that framework was to the new chemistry.
In this picture, caloric was a weightless, self-repelling fluid. Its particles repelled one another but were attracted to ordinary matter. A hot body contained a lot of caloric; heating meant adding caloric and cooling meant losing it. Because caloric was conserved, heat flow from hot to cold was simply caloric spreading out.
The theory was not foolish. It explained a great deal:
- Thermal expansion: adding self-repelling caloric pushed particles of matter apart.
- Calorimetry: the Scottish chemist Joseph Black's work on specific heat and latent heat in the 1750s and 1760s fit naturally into a bookkeeping model in which a quantity flowed between bodies and was conserved.
- Heat engines: as late as 1824, Sadi Carnot analysed the steam engine using caloric, comparing heat falling from a hot to a cold reservoir to water falling over a water wheel. His results survived the later collapse of the theory. See our explainer on Carnot and heat engines.
The rival view, that heat is a motion of the tiny parts of matter, was older and had been favoured in the seventeenth century by Francis Bacon, Robert Boyle, Robert Hooke and John Locke, among others. But it was qualitative. It offered no measurements, no bookkeeping, and no clear predictions. By the 1790s, caloric was the working theory of most chemists and physicists.
Friction was a known problem for caloric. Rubbing produced heat, and caloric theorists had an answer: friction and pressure squeeze caloric out of a body, much as squeezing a wet sponge releases water. Rubbing or cutting a metal, on this account, changed its capacity to hold caloric, so some was released as sensible heat. Rumford set out to test that explanation.
Who Rumford was
Benjamin Thompson was born in 1753 in Woburn, Massachusetts. He sided with the British during the American Revolution, left for London in 1776, and later entered the service of the Elector of Bavaria. In Munich he reorganized the army, reformed poor relief, set up workhouses, promoted the potato as a cheap food, and laid out the English Garden park. In 1792 he was made a count of the Holy Roman Empire, choosing the title Rumford after the old name of the New Hampshire town (now Concord) where he had lived and married.
He was also a prolific experimenter and inventor. He worked on fireplace design, cooking ranges, insulation, the photometer, and coffee makers, and he had studied the heat-insulating properties of fabrics and the way heat moves in fluids. His interest in heat was practical as well as theoretical; he wanted to save fuel.
The setup in the Munich arsenal
Cannon at the time were cast as solid or nearly solid metal blocks and then bored out by a cutting tool while the casting rotated. Rumford noticed that the brass became very hot and that the metallic chips thrown off by the borer were hotter still — hot enough, he wrote, to exceed the boiling point of water.
To make a controlled experiment, he took an unfinished brass casting, a piece still attached to the gun as it came from the foundry, and turned part of it into a short cylinder. A steel borer, intentionally blunted so it would grind rather than cut, was pressed hard against the bottom of a hole bored in that cylinder. The cylinder was rotated on the boring machine against the stationary borer at about 32 revolutions per minute, driven by horses.
Rumford then ran a sequence of trials:
- Measuring the heat directly. In early runs he insulated the cylinder and measured its temperature rise with a mercury thermometer inserted into a small hole in the metal.
- Checking for air. He arranged the apparatus so that air could be excluded from the rubbing surfaces, to rule out the idea that the heat came from air being compressed or chemically altered.
- Boiling water. In the most famous trial, the cylinder and borer were enclosed in an oblong, watertight wooden box filled with cold water, so the heat generated by friction went into the water. He timed the temperature rise.
The water warmed steadily. After about an hour it was noticeably hot; after about two and a half hours it boiled. Rumford reported the astonishment of the bystanders at seeing a large quantity of cold water heated and boiled without any fire. He also described his own pleasure at the sight, which he admitted was more than a grave philosopher might properly show.
Testing the sponge explanation
The sponge theory said friction released caloric by reducing the metal's capacity to hold it. If that were true, the metal dust and chips produced by boring should have a lower specific heat than the solid metal; they would be caloric-depleted brass.
Rumford tested this. He compared the heat capacity of the metal chips with that of solid metal of the same kind, by placing equal weights at the same temperature into equal quantities of cold water and measuring the result. The specific heats were the same. Nothing about the metal had changed in a way that could release caloric.
He also pointed out the quantity problem. The amount of metal ground away in his trials was small, yet the heat produced was enough to boil a large amount of water. If all that heat had been stored in the dust, the dust would have needed an absurd caloric content. And the heat kept coming as long as the horses kept walking. Nothing showed signs of running out.
From this he drew his central argument. The cylinder was effectively insulated from any source of caloric. The heat produced appeared to be inexhaustible. Anything an isolated body can furnish without limit cannot be a material substance. The only thing being communicated to the cylinder was motion. Therefore heat had to be motion, or something produced by motion. Rumford admitted he could not say exactly what kind of motion was involved.
A rough mechanical equivalent
Rumford did not convert his observations into a formal ratio of work to heat. He did note that the boring machinery could easily have been turned by a single horse, and he compared the heat produced to that of burning candles to give a sense of its magnitude.
Half a century later, James Prescott Joule worked the numbers. In his 1850 paper "On the Mechanical Equivalent of Heat" in the Philosophical Transactions, Joule took Rumford's statement that one horse could drive the apparatus, assumed the conventional horsepower of 33,000 foot-pounds per minute, and multiplied over the two and a half hours of the boiling trial. That gave 4,950,000 foot-pounds. Dividing by the heat delivered, he estimated that Rumford's data implied about 1,034 foot-pounds of work to raise one pound of water by one degree Fahrenheit.
Joule's own careful paddle-wheel measurements gave 772 foot-pounds. He noted that Rumford's figure was not wildly different and that the excess was likely because no account was taken of heat stored in the wooden box or lost to the surroundings. That an improvised arsenal trial came within about a third of the correct value is a tribute to how strong the effect was. Our article on the Joule paddle-wheel experiment covers how Joule achieved his precision.
Rumford's other attack: weighing heat
Rumford did not stop with friction. In a 1799 paper in the Philosophical Transactions, "An Inquiry concerning the Weight Ascribed to Heat," he described careful weighings of sealed flasks of water and other liquids as they were heated, cooled and frozen. If heat were a substance, a body full of it might weigh differently from the same body depleted of it. Within the sensitivity of his balance, he found no change. Caloric theorists already treated caloric as weightless, so this did not refute them, but it removed any hope of detecting heat as ordinary matter.
At about the same time, the young Humphry Davy published an account of rubbing two pieces of ice together and melting them by friction alone, which he presented as further evidence against caloric. Later historians have pointed out that his setup almost certainly allowed heat to leak in from the surroundings, so the experiment was not the clean test he described. Rumford's cannon-boring, with its large, unmistakable effect, stood on firmer ground.
Rumford also helped found the Royal Institution in London in 1799, where Davy and then Michael Faraday would work, and he endowed the Rumford Medal of the Royal Society, first awarded in 1802, and the Rumford Prize of the American Academy of Arts and Sciences.
Why caloric survived
Rumford's argument was strong, but it did not end the caloric theory. There were several reasons.
Caloric theorists had responses. Some argued that caloric could be drawn in from the surroundings through the metal parts of the machinery, so the cylinder was not truly isolated. Others suggested that the large amounts of caloric involved could come from an effectively limitless reservoir in the environment. These replies were weak, but not absurd by the standards of the day.
The motion theory lacked a quantitative alternative. Rumford could say heat was motion, but he could not say which motion, how it was measured, or how much work equalled how much heat. A theory that is merely qualitative struggles to displace one that supports calculation.
Caloric kept working. In the early nineteenth century caloric theory achieved notable successes. Pierre-Simon Laplace used caloric reasoning to correct Newton's calculation of the speed of sound by including the heating of air during compression. Joseph Fourier's mathematical theory of heat conduction (1822) did not depend on the nature of heat at all. Carnot's 1824 analysis used caloric language.
Radiant heat seemed to support caloric. The discovery that heat can travel through a vacuum as radiation, and William Herschel's 1800 discovery of infrared rays beyond the red end of the spectrum, seemed easier to explain if heat were a substance that could stream across space.
The decisive change came in the 1840s. Julius Robert Mayer in 1842, James Joule in a series of experiments from 1843 onward, and Hermann von Helmholtz in his 1847 essay on the conservation of force argued, each in his own way, that heat and work are interconvertible in a fixed ratio. Joule's measurements were the quantitative backbone. In 1850 Rudolf Clausius reworked Carnot's theory without caloric, stating what became the first and second laws of thermodynamics. William Thomson (Lord Kelvin), who had initially resisted, accepted the dynamical theory in the early 1850s. Only then did Rumford's 1798 argument look like the turning point it later became.
What heat as motion means today
Rumford's guess, that heat is a kind of motion of the parts of matter, is broadly what modern physics says, with an important refinement in vocabulary.
Physicists now distinguish internal energy (the total energy of the random motions and interactions of a body's atoms and molecules) from heat (energy transferred between bodies because of a temperature difference). In Rumford's experiment the horses did mechanical work on the cylinder. Friction converted that work into internal energy of the brass, which then flowed as heat into the cooler water. The first law of thermodynamics records the bookkeeping: the change in a system's internal energy equals the heat added to it plus the work done on it.
The kinetic theory of gases, developed by Clausius, James Clerk Maxwell and Ludwig Boltzmann in the second half of the nineteenth century, made the "motion" concrete for gases, linking temperature to the average kinetic energy of molecules. In solids, the motion is largely vibration of atoms about their lattice positions, the picture Rumford's language anticipated. Statistical mechanics and later quantum theory, including the explanation of how heat capacities fall at low temperatures, completed the picture.
Our overview of thermodynamics and entropy follows that story forward.
Common misconceptions
"Rumford measured the mechanical equivalent of heat." He did not calculate one. The figure of about 1,034 foot-pounds per pound-degree is Joule's 1850 reconstruction from Rumford's published description.
"The blunt borer produced no metal dust." It produced some. The borer was blunted so it would grind rather than cut efficiently, but metal was still worn away. Rumford's point was that the amount was far too small to account for the heat, and that the dust had the same specific heat as solid brass.
"Rumford's experiment immediately killed caloric." It did not. The theory stayed dominant for decades and was abandoned only after the energy-conservation work of the 1840s.
"Friction creates energy." Friction converts work into internal energy; it does not create energy. The energy in Rumford's boiling water came from the horses, and ultimately from their food.
Rumford's later life
Rumford left Bavaria and lived mainly in London and Paris after 1798. In 1805 he married Marie-Anne Paulze Lavoisier, the widow of the chemist who had named caloric and had been guillotined in 1794 during the French Revolution. She had been Lavoisier's scientific collaborator. The marriage was unhappy and the two separated within a few years. Rumford died at Auteuil, near Paris, in 1814. Our profile of Antoine Lavoisier covers the other side of that story.
Why the experiment still matters
The cannon-boring experiment is a textbook case of how a single, well-chosen observation can undermine a theory without immediately replacing it. Rumford identified a real anomaly, designed a test that targeted the main rescue explanation (the sponge model), and drew a clear logical conclusion. What he lacked was a quantitative framework, and that gap let a useful but wrong theory persist.
It also illustrates the value of paying attention to industrial processes. The arsenal was not a laboratory, but the large scale of the machinery made the effect impossible to miss. The same pattern repeats in the history of thermodynamics, which grew up around steam engines; see our piece on the steam engine.
The takeaway
Rumford's 1798 cannon-boring experiment showed that friction from a blunt borer could bring a box of water to the boil in about two and a half hours, that the metal dust had the same specific heat as solid brass, and that the supply of heat showed no sign of running out. He concluded that heat is motion, not a material fluid. The argument was correct but qualitative, and caloric theory kept its grip until Mayer, Joule, Helmholtz and Clausius built a quantitative science of energy in the 1840s and 1850s. When Joule reworked Rumford's numbers, they implied a mechanical equivalent within about a third of the true value — strong evidence that the Munich arsenal had seen the conversion of work into heat half a century before physics was ready to name it.
Frequently Asked Questions
Q: How long did it take Rumford to boil the water? A: In his water-box trial, the water boiled after about two and a half hours of continuous boring. Bystanders were reportedly astonished to see cold water boil without any fire.
Q: Why did Rumford use a blunt borer? A: A blunt tool grinds instead of cutting efficiently, so most of the horses' work turned into friction and little metal was removed. That made it easy to show that the heat produced far exceeded anything that could be stored in the small amount of dust.
Q: How did Rumford rule out the idea that friction squeezed caloric out of the metal? A: He compared the specific heat of the metal chips with that of solid brass and found them the same. If the chips had lost caloric-holding capacity, their specific heat should have been lower.
Q: Did Rumford calculate the mechanical equivalent of heat? A: No. Joule did so from Rumford's description in 1850, assuming one horse at 33,000 foot-pounds per minute for two and a half hours. He obtained about 1,034 foot-pounds per pound of water per degree Fahrenheit, compared with his own value of 772.
Q: Why did caloric theory survive after 1798? A: Caloric theorists offered alternative explanations, Rumford had no quantitative theory of his own, and caloric reasoning continued to give useful results, from Laplace's speed-of-sound correction to Carnot's analysis of heat engines. The energy-conservation work of the 1840s finally displaced it.
Q: Was Rumford the first to say heat is motion? A: No. Bacon, Boyle, Hooke and Locke had proposed motion theories of heat in the seventeenth century. Rumford's contribution was a decisive experiment that made the material theory hard to defend.
Q: Where was the experiment done? A: At the military arsenal in Munich, in the service of the Elector of Bavaria, where Rumford oversaw the boring of brass cannon.
Q: What happened to the heat Rumford produced? A: The horses' work became internal energy of the brass through friction, and that energy flowed as heat into the cooler water in the box. Some escaped into the wooden box and the air, which is one reason Joule's reconstruction from Rumford's numbers came out higher than the true mechanical equivalent.
Sources
- Thompson, B. (Count Rumford) (1798). "An Inquiry concerning the Source of the Heat which is excited by Friction." Philosophical Transactions of the Royal Society of London, 88, 80–102.
- Thompson, B. (Count Rumford) (1799). "An Inquiry concerning the Weight Ascribed to Heat." Philosophical Transactions of the Royal Society of London, 89.
- Joule, J. P. (1850). "On the Mechanical Equivalent of Heat." Philosophical Transactions of the Royal Society of London, 140, 61–82.
- Lavoisier, A. (1789). Traité élémentaire de chimie. Paris.
- Carnot, S. (1824). Réflexions sur la puissance motrice du feu. Paris.
- Brown, S. C. (1979). Benjamin Thompson, Count Rumford. MIT Press.
- Encyclopedia.com, "Thompson, Benjamin (Count Rumford)," Complete Dictionary of Scientific Biography.
Read next:
Follow new articles
Add SciHub101 to your feed reader to see new science and technology stories.
About the Author
SciHub101 is an independent science and technology publication founded and edited by Buğra Sözeri, a software engineer and app developer based in Türkiye. AI tools assist with drafting and research. Source documentation in the existing archive is uneven, and articles are being reviewed and corrected. The Editorial Team byline identifies editorial responsibility; it does not certify specialist or independent human review. Where we compare products, anything built by our own team is labelled as such — see the editorial guidelines and About page for how corrections and disclosures work.
Related Articles
Hess's Balloon Flights: How Cosmic Rays Were Found (1912)
In 1912 Victor Hess carried electroscopes to 5,350 m in a balloon and found radiation rising with height. Here is how the flights worked and what they proved.