At approximately 10:53 p.m. on January 21, 1910, something peculiar happened in Paris. Clocks throughout the city stopped. Their minute hands froze, their hour hands remained motionless, and thousands of Parisians suddenly found themselves living in a world where it was perpetually a few minutes before eleven.

This was not a coordinated protest by disgruntled clockmakers. Nor had French clocks decided to call a city-wide strike. The clocks stopped because the Seine River was flooding.

There is, admittedly, a missing step in that explanation. Rivers flood without ordinarily bringing time to a standstill, and we could clear up the mystery immediately. But then you would have no reason to continue reading, and we would have gone to all this trouble for nothing.

For the previous three decades, Paris had operated a service that delivered the correct time through underground pipes. Thousands of pneumatic clocks depended on regularly scheduled bursts of compressed air to move their hands forward. The French capital had, in effect, put time on tap.

And in January 1910, the plumbing failed.

When Four O’Clock Lasted Half an Hour

To understand why anyone would attempt to distribute time through plumbing, we must return to an age when knowing the hour required considerably more effort than glancing at the small glowing rectangle currently consuming much of your attention.

Today, our phones, computers, watches, and assorted kitchen appliances generally agree on the time. They may disagree about whether we should install an update, renew a subscription, or purchase an extended warranty, but their clocks are usually cooperative. In nineteenth-century Paris, public clocks displayed a more independent spirit.

A town-hall clock might announce one time, a railway clock another, and a church bell offer a third opinion with the confidence of someone who had never contemplated being wrong. Some of the discrepancy came from individual clocks that simply ran fast or slow. More broadly, rapidly expanding railroads and commerce made the old habit of relying on local time increasingly impractical.

In 1854, astronomer Urbain Le Verrier, director of the Paris Observatory and a man whose calculations had helped identify the planet Neptune, complained about the city’s inconsistent clocks. He recalled an observation that Parisian clocks had once been heard striking midnight across an interval approaching three-quarters of an hour.

Imagine trying to arrange a meeting under those circumstances. The statement “I’ll be there at noon” would require considerably more negotiation than anybody should have to undertake before lunch. By March 1880, the newspaper Le Temps complained that four o’clock could last half an hour in Paris. This was an impressive achievement in leisurely living, but not ideal for keeping trains on schedule.

The problem was international. As we discovered in our examination of the peculiar history of time zones, industrial society was discovering that accurate clocks were only half the battle. The other half involved persuading everyone to use the same time.

Victor Popp’s Excellent Idea, Which Involved a Lot of Air

Parisian astronomers preferred an electrical solution. A master clock could send signals through wires to other clocks, keeping them synchronized with the time established by the observatory. It was scientifically elegant and potentially very precise, which naturally meant that implementing it became a complicated undertaking.

Paris organized a competition for suitable clocks, but the results were underwhelming. Only four instruments were submitted, and just two met the required standards. The ambitious electrical project wasn’t ready for the 1878 Paris World’s Fair, when the city had hoped to show international visitors how enthusiastically it embraced modern technology.

Enter Austrian-born engineer Victor Popp, who approached the question from a different direction. Instead of sending electrical signals to clocks that kept themselves running, why not send compressed air to clocks that would move only when the air told them to?

There was precedent. Pneumatic clocks had appeared in Vienna in 1877, and the idea attracted attention at the 1878 exposition. Paris already had another air-powered curiosity: a pneumatic postal network that used underground tubes to propel containers carrying messages. The city had been moving information around with air pressure for years. In 1897, New York City’s postal service even sent a live cat through its pneumatic mail system, because someone apparently thought ordinary letters weren’t sufficiently challenging. Moving clock hands seemed almost restrained by comparison.

Popp installed a central station at 7 Rue Sainte-Anne and connected it to pipes running beneath nearby streets. On March 15, 1880, the pneumatic clock service began operation. For the first time, Parisian customers could have the correct time delivered directly to their premises, without wires, clock winding, or personal visits from an astronomer.

As the history of Paris’s competing time-distribution schemes makes clear, the pneumatic system was not necessarily the astronomers’ favorite. It was, however, practical enough to get built. When you’re trying to impress the public, an operating clock has certain advantages over a theoretically superior clock that isn’t there.

How to Tell Time with Pipes, Bellows, and a Straight Face

The principle behind Popp’s invention was surprisingly simple. At the central station, steam-powered compressors forced air into storage reservoirs. A master clock operated a valve, sending a pulse of compressed air into the network every minute. The pulse traveled through pipes beneath the streets and through smaller branches leading into buildings.

Inside each receiving clock was a small bellows or flexible diaphragm. Air pressure pushed against it, moving a lever connected to a toothed wheel. Each pulse advanced the wheel one notch, shifting the minute hand forward exactly one minute. Ordinary gearing moved the hour hand along with it.

That’s all. No motor in each clock, no spring needing regular winding, and no miniature French engineer stationed behind the dial waiting for instructions. The clocks simply waited for their next puff of air, took one step forward, and settled down again.

Sixty pulses made an hour; 1,440 made a day. According to an 1882 description of the apparatus, the main distribution pipes were roughly an inch in diameter, and each air pulse lasted about twenty seconds. The central machinery compressed air to a much higher pressure before reducing it for distribution to the clocks.

The arrangement produced an unusual visual effect. Unlike a conventional clock whose hands might move continuously, the pneumatic clock’s minute hand remained still, then jumped to the next mark when its air pulse arrived. Anyone watching closely could enjoy sixty tiny demonstrations of municipal efficiency each hour. Entertainment was simpler before streaming services.

There was one technical complication: air takes time to travel. Clocks near the station received their pulses slightly sooner than those farther away. An early assessment found differences of several seconds between clocks at opposite ends of the network. Astronomers were unimpressed by such imperfections, but the average Parisian probably considered a clock that was only two seconds wrong a considerable improvement over the earlier system in which four o’clock was practically an afternoon activity.

Five Centimes a Day, and You Can Keep the Minutes

Of course, a brilliant technology is only truly brilliant once somebody figures out how to charge for it.

Popp’s company offered clocks by subscription. The first cost five centimes per day, the second four, and each additional clock three. The company provided the necessary equipment and installed the pipes. A customer who rented one clock paid about 18.25 francs a year in exchange for the convenience of a clock that required no regular winding and received its marching orders from a central station.

The service attracted banks, hotels, shops, theaters, and private households. By the end of 1880, approximately 1,475 private clock faces had been installed. By 1884, there were roughly 7,800 private clock faces serving about 3,500 subscribers.

That is a substantial number of people paying a recurring fee to learn whether they were running late. Modern software companies would recognize the business model immediately, although Popp had the decency not to introduce separate subscription tiers for clocks that displayed the hour and clocks that also displayed the minute.

The clocks were only the beginning. Popp’s company expanded its compressed-air network to provide mechanical power for workshops, elevators, freight lifts, and industrial machinery. Instead of operating their own steam engines, businesses could purchase power through underground pipes. The enterprise developed into the Compagnie Parisienne de l’Air Comprimé, which eventually built a major generating station beside the Seine at Quai de la Gare.

This was an impressively modern idea: one centralized system supplying essential services to many customers. There was, however, a drawback nobody could afford to overlook. When the central system failed, its ability to disappoint large numbers of people was just as impressive as its ability to serve them.

January 1910: The Seine Files a Complaint

The winter of 1909–1910 was exceptionally wet. Heavy rainfall saturated the countryside, rivers swelled, and the Seine began rising beyond the limits of what Paris’s streets, sewers, and underground tunnels were designed to accommodate. Soon water was appearing in basements and neighborhoods where residents had previously encountered the Seine only by making a deliberate trip to the riverbank.

The flooding became one of the worst disasters in modern Parisian history. Streets turned into canals, residents crossed submerged areas on temporary wooden walkways, and boats appeared in places where the municipal planners had expected horses and pedestrians. On January 28, the Seine reached roughly 28 feet at the Austerlitz gauge, measured against its reference level.

Unfortunately, the compressed-air generating plant at Quai de la Gare sat beside the river. This had seemed an excellent location while the river was behaving itself. It became considerably less appealing when the Seine decided to investigate the facility by conducting an unauthorized tour of the premises.

The company’s engineers initially believed their generating station was high enough to withstand the flooding. But water forced its way through an underground opening and reached essential equipment. Workers tried to contain the damage, but machinery associated with the compressors was submerged. Without the equipment operating, the company could no longer maintain its regular air supply.

On Friday evening, January 21, the network failed. According to reporting published in Le Petit Parisien on January 23, 1910, pneumatic clocks stopped between 10:48 and 10:53 p.m. The exact time varied, so the familiar claim that all Paris’s clocks froze at precisely 10:53 is a slight exaggeration. Ordinary clocks still worked, and not every timepiece in the city was connected to the pneumatic system.

But thousands of connected clocks had abruptly stopped receiving their instructions. At eleven o’clock, they insisted that it was still ten fifty-something. At midnight, they remained unconvinced. By breakfast, they were offering a particularly unhelpful assessment of the previous evening.

The city had not merely lost track of time. It had discovered how to distribute the same mistake to thousands of customers at once.

In that respect, Paris was stunningly ahead of its time. More than a century later, in November 2026, a massive Microsoft outage disrupted workplaces around the world, and the very next day, an unrelated communications failure at the Federal Aviation Administration brought departing flights across the United States to a standstill. Apparently, more than a century of technological progress has not eliminated our ability to turn a single malfunction into everybody’s problem. In 1910, at least the clocks had the decency to tell everyone exactly when they’d stopped working.

Unfortunately, the Elevators Also Needed Air

The stopped clocks were only the most conspicuous evidence of the outage. Compressed air also powered elevators, freight lifts, tools, and other machinery.

Disappointingly, the compressed-air elevators did not operate like enormous potato guns, launching the elevator and its passengers upward with bone-shattering G-forces. Had they worked that way, you can be certain this article would be devoted exclusively to that magnificent achievement in vertical transportation. Instead, some of these elevators used compressed air to pressurize water, which then operated the hydraulic lifting mechanism. A perfectly sensible arrangement, but a dreadful waste of comedic potential.

When the station failed, businesses that relied on those services found themselves facing considerably more serious problems than a clock permanently announcing that bedtime was approaching. Customers descended on the company’s offices at 54 Rue Étienne-Marcel seeking an explanation.

Officials could tell them that floodwater had stopped the generating equipment, but they could not promise exactly when normal service would resume. Engineers deployed pumps and tried to drain the affected machinery while the river continued to make itself unwelcome.

One can imagine the conversations. “My elevator isn’t working.” “I’m afraid we’ve run out of air.” “In Paris?” “Yes, monsieur. We are hoping to have more air by Monday.” Customer service was evidently no easier before outsourced customer service call centers.

The flood also revealed the system’s fundamental weakness. Unlike ordinary clocks that continued running between adjustments, Popp’s clocks depended entirely on the next air pulse. A broken signal didn’t merely make the clock inaccurate. It made the clock stop.

It’s the same uncomfortable tradeoff found in many modern technologies. Centralization makes individual devices simpler and services convenient, right up until the central provider experiences technical difficulties. Paris invented a networked service that worked beautifully until its server room flooded. All it lacked was an apologetic message urging users to try again later.

1911: The Government Decides to Stop Time, Too

Surely, after the flood, Paris had exhausted its supply of memorable incidents involving stopped clocks. Not quite. A little over a year later, the clocks stopped again, this time because the French government wanted them to.

In 1891, France had established a single national time standard based on the Paris meridian, replacing the bewildering assortment of local times previously used throughout the country. It was a considerable improvement, but international transportation and communications increasingly demanded something more ambitious: a common reference point for the entire world. Unfortunately for French national pride, the International Meridian Conference of 1884 had recommended Greenwich, England, as the world’s prime meridian.

France was not in an overwhelming hurry to accept a system associated with Britain. The French possessed an impressive history of astronomy, a perfectly serviceable meridian of their own, and centuries of experience in disagreeing with the English. Nevertheless, practicality eventually triumphed over national pride.

In March 1911, France adopted a new official time standard, moving its clocks backward by nine minutes and twenty-one seconds to align with Greenwich Mean Time. The change took effect at midnight between March 10 and 11.

This presented a problem for the pneumatic clocks. Their mechanisms were designed to move forward one minute at a time whenever air arrived, not gracefully reverse themselves because politicians had revised the legal definition of noon.

The solution was delightfully straightforward: stop sending pulses for nine minutes and twenty-one seconds, then resume the service. As accounts of the change describe, Paris’s pneumatic clocks simply waited for the new official time to catch up with them.

In 1910, a river had stopped the clocks accidentally. In 1911, the government did it deliberately. For anyone interested in the relationship between politicians and timekeeping, the distinction is worth savoring.

Benjamin Franklin might have appreciated the episode. During his stay in Paris, he had jokingly suggested waking Parisians earlier to save candles, with church bells and even cannon fire as possible incentives. By 1911, France had developed the opposite approach. Instead of waking people up, it simply gave the clocks a nine-minute rest.

Eventually, Time Runs Out

The pneumatic clock network survived the flood, the government time adjustment, and several more decades of technological change. But by the 1920s, its fortunes were declining. Electrical timekeeping had become increasingly practical, and maintaining a specialized underground network merely to advance clock hands was difficult to justify.

The company still supplied compressed air for industrial purposes, but its original clock business was shrinking. From approximately 3,500 subscribers in 1884, it had fallen to only 919 in 1926. There were still thousands of clock faces connected, but fewer customers to help pay for the pipes, machinery, repairs, and municipal charges.

When the concession expired in 1927, the company chose not to renew it. According to its corporate history, it transferred the remaining clock business and equipment to Magneta. Paris’s extraordinary experiment in delivering time by compressed air was over.

The larger compressed-air utility endured for decades, finally closing in the 1990s. A former generating station beside the Seine survives as part of the École nationale supérieure d’architecture de Paris-Val de Seine. And an 1880 depiction of the pneumatic clocks preserved by the Musée Carnavalet offers a glimpse of the devices that once told Parisians when another minute had arrived.

Time Was on Tap. Unfortunately, So Was the Seine.

It’s tempting to dismiss the pneumatic clock network as one of those eccentric nineteenth-century inventions that looked promising until someone developed a less complicated alternative. But Popp’s system addressed a problem that remains essential today: getting enormous numbers of devices to agree about the time.

Our computers, phones, financial networks, and navigation systems depend on synchronized clocks. They use electronics and satellites instead of pipes and bellows, but the underlying ambition is familiar. Establish the correct time, distribute it reliably, and hope nobody disconnects whatever makes the whole arrangement work.

For nearly half a century, Paris made that system function with steam engines, underground pipes, and precisely timed puffs of air. It was clever, ambitious, and surprisingly successful. It also produced a citywide demonstration of what happens when an essential service depends on machinery beside a flooding river.

And perhaps that is the most enduring lesson. Engineers spent years figuring out how to deliver time through pipes. The government later demonstrated it could make time stand still by decree. But when the Seine decided to get involved, neither ingenuity nor legislation made much difference.

Paris had figured out how to put time on tap. Unfortunately, nobody had thought to install a shutoff valve for the Seine.

---