The Cannon-Boring Cauldron

Count Rumford and the Death of the Caloric Fluid in Munich

Inside the Electoral Military Arsenal in Munich in the spring of 1798, American-born physicist Benjamin Thompson, Count Rumford, submerged a brass cannon cylinder and a purposely blunted steel borer inside an oak box filled with 18.77 pounds of cold water and harnessed two horses to turn the drive shaft. At the close of the eighteenth century, chemical consensus led by Antoine Lavoisier in Paris held that heat was a conserved, weightless material substance called caloric (calorique). In Lavoisier's 1789 table of the chemical elements, caloric stood at the top of the list as an indestructible elastic fluid whose self-repulsive particles filled the microscopic pores between atoms. Heating a body meant adding caloric to its pores, and cooling meant draining caloric out.

To explain why hammering an iron bar on an anvil or boring out a cannon barrel produced intense heat without any fire nearby, caloric theorists relied on the concept of differential heat capacity, first measured by William Irvine and Adair Crawford in Scotland. They argued that solid block metal possessed a larger internal sponge-like capacity for storing latent caloric than compressed or finely shredded metal. When a sharp lathe bit shaved thin ribbons of bronze out of a cannon barrel, the metallic turnings supposedly suffered a permanent reduction in their specific heat capacity, squeezing their stored caloric fluid out as sensible heat.

While supervising the boring of six-pounder brass field guns at the Munich foundry, Rumford noticed two facts that clashed with the caloric-squeezing explanation. First, the metallic chips dropping from the cannon muzzle were far hotter than boiling water. Second, even when a dull cutting tool rubbed against the bronze for hours while detaching almost no shavings, the entire mass of the cannon barrel grew scorching hot. Determined to test whether frictional heat came from a change in heat capacity or from a finite internal stock of caloric, Rumford cast a cylindrical brass head onto the end of a cannon casting. He turned it down to a cylinder 7.75 inches in diameter and 9.8 inches long linked to the gun by a narrow neck, bored a central cavity into it, and pressed a flat, blunt steel borer against its base with an axial force of ten thousand pounds.

Fitting an oak water box tightly around the brass cylinder via oiled leather collars, Rumford filled the box with cold water at $60^\circ\text{F}$ ($15.6^\circ\text{C}$) and set two horses walking around the capstan to rotate the cannon barrel at thirty-two revolutions per minute against the stationary blunt steel bit. After one hour of continuous friction, a mercury thermometer inside the water box read $107^\circ\text{F}$. At ninety minutes it reached $142^\circ\text{F}$, at two hours it hit $178^\circ\text{F}$, and at two hours and thirty minutes, the entire 18.77 pounds (8.5 kilograms) of water erupted into a vigorous rolling boil without a spark of fire anywhere in the building.

Rumford then executed three quantitative checks that systematically closed every escape hatch of the caloric theory:

1. 1. Identical Specific Heat Capacity: Collecting the 4,145 grains (268.6 grams) of metallic powder ground off by the blunt borer during the trial, Rumford heated equal weights of that abraded dust and thin strips of unworked solid cannon bronze to the boiling point of water and plunged each into separate cold-water calorimeters. Both samples raised the calorimeter water by the exact same temperature increment, proving that mechanical abrasion had not altered the metal's heat capacity.
2. 2. Exclusion of Atmospheric Oxidation and Latent Fluids: Because the brass cylinder, the blunt steel bit, and the boring interface remained completely submerged beneath water inside sealed leather collars throughout the run, the heat could not have entered from chemical combustion with atmospheric oxygen or decomposition of surrounding air.
3. 3. Inexhaustible Linear Generation: The thermal output of the apparatus—which Rumford calculated as equal to nine large wax candles burning simultaneously—never slowed down or showed the slightest sign of depletion across repeated trials, continuing at a constant rate for as long as the horses supplied mechanical work.

Because no finite block of metal can contain an inexhaustible reservoir of a conserved material fluid, Rumford reported to the Royal Society of London that heat could only be microscopic mechanical motion. His Munich trials established the first experimental ratio between horsepower and thermal output forty-five years before James Prescott Joule fixed the mechanical equivalent of heat at 4.184 joules per calorie.