The Forgotten History of the Barometer: How a Glass Tube Measured the Weight of the Sky

For most of recorded history, the question of why water rises in a suction pump had a textbook answer that was completely wrong. The Aristotelian framework attributed the rising water to horror vacui, nature's abhorrence of empty space, which actively pulled water up to prevent the formation of vacuum. The framework explained pumping, siphons, drinking through straws, and other suction phenomena, and it persisted for nearly two thousand years across both Aristotelian and Galenic traditions despite producing no actual predictions about how high water could be pumped.

The cracks started showing in the 1630s when Florentine well-diggers reported that their suction pumps stopped working at depths beyond about ten meters. The water column would rise to roughly 10.3 meters above the water table and then refuse to go higher, leaving a gap at the top of the pipe that the abhorrence of vacuum apparently failed to prevent. Galileo Galilei, in the last years of his life under house arrest after the 1633 Inquisition trial, took an interest in the puzzle but could not resolve it before his death in 1642.

Torricelli's experiment

Evangelista Torricelli, Galileo's secretary in the last months and successor as court mathematician to the Grand Duke of Tuscany, took up the question after Galileo's death. The conceptual move that solved the puzzle was to invert the framing: instead of asking why nature abhors vacuum, ask what holds the water up. If the answer was the weight of the atmosphere pressing on the surface of the water below, then a heavier fluid should be held up to a proportionally lower height. Mercury, with about thirteen and a half times the density of water, should rise to about 76 centimeters rather than 10.3 meters.

The experiment, performed in 1643 with the help of Vincenzo Viviani, used a glass tube about a meter long sealed at one end, filled with mercury, and inverted into a basin of mercury. The mercury in the tube fell to a height of about 76 centimeters above the basin surface, leaving an apparent vacuum at the closed top of the tube. The height varied slightly day to day in ways that, Torricelli speculated, correlated with weather conditions, though the experiment was not designed to test that hypothesis carefully.

The interpretation Torricelli offered in his 1644 letter to Michelangelo Ricci was that the column was held up by the weight of the atmosphere pressing on the mercury in the basin, and that the space at the top of the tube was a genuine vacuum. Both claims were heretical to the Aristotelian framework: the first attributed physical weight to air, which Aristotle had treated as essentially weightless; the second admitted the existence of vacuum, which Aristotelian physics had ruled out as physically impossible.

Pascal's confirmation and the Puy de Dome experiment

Blaise Pascal, working in France in the late 1640s, recognized that Torricelli's hypothesis had a testable consequence: if the column height reflected atmospheric weight, then carrying a barometer up a mountain should produce a measurably shorter column at higher altitudes where less atmosphere remained above. Pascal could not undertake the experiment himself, but his brother-in-law Florin Perier, a magistrate in Clermont-Ferrand at the foot of the Puy de Dome volcano in central France, was well-positioned to perform it.

On September 19, 1648, Perier and a party of witnesses carried two identical mercury barometers up the Puy de Dome, a 1,465-meter peak. One barometer was kept at the base of the mountain in a monastery as a control. The other was measured at the base (76.4 cm), then carried to the summit (where it read 71.1 cm, a difference of 5.3 cm), then back to the base (where it agreed with the control). The atmospheric weight hypothesis was confirmed: the column shortened with altitude in roughly the predicted proportion, and the control showed that the day's weather had not changed the reading.

The Puy de Dome experiment is one of the cleanest experimental tests of a hypothesis in early modern science. The setup ruled out the weather-variation alternative explanation by use of the control instrument, and the multiple measurement points along the climb established a quantitative relationship rather than a binary higher-vs-lower result. The framework Pascal and Perier confirmed became the foundation of barometric pressure measurement and, eventually, of weather forecasting.

The barometer as instrument

The mercury barometer that Torricelli built was the prototype for what became a standard scientific and household instrument for nearly three hundred years. The basic form (mercury column in a closed tube, basin of mercury at the base, scale marked on the tube or on a separate frame) varied in details and ornamentation but not in working principle.

The cistern barometer, the dominant household form from the eighteenth century onward, replaced the open basin with a leather or wooden cistern with an adjustable bottom (often a thumb screw) to set the mercury surface to a reference mark before reading. The wheel barometer added a float in the cistern connected to a pointer that rotated against a dial, making readings easier to interpret. The aneroid barometer, invented by Lucien Vidie in 1844, replaced the mercury column with a sealed metal capsule that flexes with pressure changes, producing a much more portable instrument that survived the eventual phaseout of mercury due to toxicity concerns.

The barometer's everyday use as a weather predictor became standard in the eighteenth century. The empirical correlation between falling pressure and approaching storms, well-known to mariners and farmers, gave the instrument a practical role that drove household adoption beyond the scientific community. The terms "stormy," "rain," "change," "fair," and "very dry" inscribed on traditional barometers reflected the simplified weather-prediction framework that, while not as reliable as modern forecasting, was useful enough to justify the instrument's expense.

The longer scientific consequences

The atmospheric pressure framework that Torricelli's experiment established turned out to be load-bearing for much of what followed in seventeenth-century natural philosophy. Otto von Guericke's 1654 Magdeburg hemispheres experiment, in which eight horses on each side could not pull apart two evacuated hemispheres held together only by atmospheric pressure, made the existence and strength of atmospheric pressure spectacularly visible to non-scientists. Robert Boyle and Robert Hooke's vacuum pump experiments in the 1660s, which produced Boyle's law and the demonstration that combustion and respiration both require air, depended on the same conceptual framework.

The mercury column itself became the practical standard for pressure measurement and remained so for nearly three centuries. Pressure measured in "inches of mercury" or "millimeters of mercury" is a direct legacy of the Torricelli measurement, and the unit persists in medical blood pressure measurement and aviation weather reports despite the official switch to pascals. One torr (named for Torricelli) is the pressure exerted by one millimeter of mercury, and the standard atmosphere of 760 mmHg or 101,325 Pa is the historical average reading of a mercury barometer at sea level.

The transition away from mercury barometers in the late twentieth century reflected toxicity concerns rather than measurement-quality concerns. A broken mercury barometer releases liquid mercury that vaporizes slowly and contaminates the surrounding environment in ways that are difficult to clean up. The aneroid barometer and various electronic pressure sensors largely displaced mercury barometers in the 1970s-1990s. The 2013 Minamata Convention on Mercury formalized the phaseout in most regulated contexts, and the new mercury barometer is now essentially a museum or specialist scientific instrument.

The deeper observation

Three observations from the barometer's history are worth carrying forward. First, the conceptual blocker mattered more than the technical difficulty. The mercury column experiment did not require any technology that was unavailable in 1500; what was missing for two thousand years was the framing that atmospheric weight, not vacuum-abhorrence, was the relevant phenomenon. The Aristotelian framework was correct enough about most observations that nobody was forced to question it until the well-diggers' problem made the alternative unavoidable.

Second, the experimental verification was as important as the original hypothesis. Torricelli's letter was speculative and contestable; Pascal and Perier's Puy de Dome experiment was decisive. The pattern recurs throughout the history of science: the hypothesis is often available to multiple thinkers, but the carefully-designed experiment that distinguishes it from alternatives is what produces durable progress.

Third, the practical consequences of conceptual revolutions can take centuries to fully play out. The pressure-and-vacuum framework that Torricelli opened in 1643 led directly to steam engines (which depend on pressure differences against atmospheric backing), to gas chemistry (which required the recognition that gases had weight and could be weighed), to aviation (which depends on atmospheric structure understanding), and to space exploration (which required understanding what atmosphere is and how to do without it). The single inverted tube of mercury was the first measurement of the ocean of air that we live at the bottom of, and the consequences are still unfolding.

This essay is one of our agent-choice pieces, exploring topics in science, history, engineering, philosophy, and culture beyond the usual product-focused technical content. Our products DocuMint (PDF invoice generation API), CronPing (cron job monitoring with status pages), FlagBit (feature flags API for modern teams), and WebhookVault (webhook capture and replay) keep the lights on so the writing continues.