The Drop That Refused to Be Seen

In 1930, a physicist cut the sealed stem of a glass funnel in Brisbane and walked away. The material inside looked solid. It could be struck with a hammer and shattered. Then it began to fall.

By 2026, it has produced nine drops.

The funnel belongs to the University of Queensland’s Pitch Drop experiment, a piece of apparatus so simple that it almost seems like a joke: a little pitch in a funnel, a beaker underneath, and enough time to make the joke very patient. The university’s current account still records only nine drops, the last in April 2014. It is a scientific experiment whose most famous result is not a measurement but an absence: nobody has ever watched a drop fall.

A liquid with no interest in looking like one

Thomas Parnell, the university’s first professor of physics, set up the experiment in 1927. He heated pitch, a derivative of tar once used to waterproof boats, and poured it into a glass funnel whose narrow stem was sealed. The heat made its fluidity obvious. Parnell then left it alone for three years, allowing the pitch to cool and settle before cutting the stem in October 1930.

At room temperature pitch behaves like a solid in the brief span available to a person. It feels hard. Hit it and it breaks. But “solid” is not a complete description of what matter does; it is partly a description of how quickly the matter is being asked to change. The pitch is moving, only at a speed that makes ordinary human patience a poor instrument for detecting it.

The original scientific paper, published in 1984 by Parnell and colleagues, estimated the sample’s viscosity at 2.3 × 10⁸ pascal-seconds. Water at 20 °C is about 1.0 × 10⁻³ pascal-seconds. The comparison is roughly 230 billion to one. The university’s less mathematical description is easier to carry around: pitch is about 100 billion times more viscous than water.

Both figures point to the same unsettling fact. A fluid does not have to look like water. It does not have to pour, splash, or yield when you press a finger against it. It only has to deform continuously under stress, even if the continuity is spread across a human lifetime.

Eight years for the first answer

The stem was cut in 1930, but the first drop did not detach until 1938. The next five took more than forty years altogether. The sixth arrived in April 1979; the seventh in July 1988. By then the experiment had already become a lesson in the difference between an event and an observation. The pitch was flowing whether or not anyone was looking. The difficult part was arranging for a human being to be looking at exactly the right moment.

John Mainstone became the second custodian in 1961 and kept the experiment for 52 years. He missed three drops. In 1977 he was one day late. In 1988 he missed by five minutes while the experiment was on display at Brisbane’s World Expo. In 2000, a webcam was recording the eighth drop, but a power outage lasting about twenty minutes supplied the blackout at precisely the wrong time.

The ninth drop seemed to offer a technological correction. There would be cameras, a live stream, and thousands of people waiting together. More than 31,000 viewers from 158 countries registered to watch. Yet the experiment had changed the kind of problem it posed.

A water drop falls cleanly: it accelerates, separates, and hits the surface. The pitch did not. The eighth drop had reached the pile in the beaker but had not broken free from the funnel. When the ninth descended, it met that older drop and touched down so gently that the contact was hidden from the camera. The best initial estimate put the event somewhere between 11 and 14 April 2014. By comparing the ninth drop’s changing speed in daily time-lapse images, researchers concluded that the touchdown occurred on 12 April.

They had not seen it happen. They had inferred it from the before and after.

The experiment was never as controlled as it sounds

The Pitch Drop is often presented as a clockwork demonstration of viscosity. It is actually a record of the room around it. The apparatus is kept in a display cabinet, not in a carefully regulated laboratory. Its flow changes with seasonal temperature. Pitch becomes more viscous when it is cooler, and small changes matter when the motion is already almost impossibly slow.

That helped make the eighth and ninth drops troublesome. Air-conditioning installed near the experiment in the 1990s may have cooled it. Later, lights in the display cabinet warmed the sealed space beneath the bell jar. In 2014, measurements found the pitch at the top of the funnel almost ten degrees hotter than the pitch in the forming drop. The tenth drop was growing faster than expected, at about 19 cubic millimetres a day in the first six months studied. The lights were eventually changed to cooler LEDs.

This is not a flaw that invalidates the experiment. It is the experiment’s most useful complication. The apparatus was designed to make one point — that pitch flows — not to produce a perfectly repeatable number. Its history shows how a scientific object acquires new questions after it leaves the hands of its creator. What began as a lecture demonstration became a study in temperature, measurement, video analysis, and the limits of surveillance.

Waiting is part of the apparatus

The experiment holds a peculiar place in science because its duration is not a cost to be minimised. It is the thing being demonstrated. A faster version would be more convenient and less persuasive. The point is not merely that pitch has a large viscosity; the point is that a substance can be visibly brittle and nevertheless be flowing beneath the threshold of a human life.

That threshold changes what counts as evidence. The first six drops gave the 1984 authors enough accumulated volume and elapsed time to estimate a property of the material. The missed eighth drop made a camera look fragile. The ninth made still photographs and a graph more trustworthy than a crowd of witnesses. The future tenth drop will add another answer, though perhaps not the answer people want. If the cameras catch it, they will show a slow change between frames, not a cinematic fall. If they miss it, the experiment will continue to say exactly what it has always said: time passed, and the material moved.

A human observer wants a moment. The funnel offers a process. Somewhere in the cabinet, a dark bulb is becoming heavy enough to let go, at a pace that makes waiting look like an instrument rather than a failure. Its next drop may arrive before the decade ends. The strangest part is that, when it does, the event will already have happened long before anyone can prove that they saw it.