The ball that falls at one

1Every afternoon at five minutes to one, a red ball climbs halfway up a mast on the roof of Flamsteed House in Greenwich, and stops. At two minutes to one it climbs to the top. At one o’clock exactly, it falls.¹1The times are those the Royal Observatory gives for its visitors. The ball runs on Greenwich Mean Time in winter and on British Summer Time in summer, so in summer it falls at noon by the sun’s reckoning, give or take. That is the whole performance, and it has run since 1833.

2On the Thursday I went to see it there were about forty of us in the courtyard below, most of us holding up phones to film a ball doing what balls do.

3I had come to look at a clock with no face. A time ball tells one time a day, to anyone who can see it, and it does so without a sound. It does not tick, it does not strike, it has no hands. For five minutes it gets ready, and then for less than a second it tells you something precise.²2Watch the top of the mast, not the bottom. By the time the ball lands, the moment it was announcing is already gone. It was built for people who needed that second more than they needed anything else the clock could have said.

A clock for ships

4The people it was built for were on the river. In the 1830s the Thames below Greenwich was full of ships about to leave for months at sea, and every one of them carried a chronometer: a heavy, carefully sprung watch in a gimballed box, set to the time at Greenwich and kept on it, as nearly as possible, for the whole voyage.

5The reason is longitude. A navigator can find local noon from the sun, the moment it stands highest, wherever the ship happens to be. If the chronometer says that local noon arrived at a quarter past two in Greenwich, the ship is two and a quarter hours of the earth’s turning to the west, and since the earth turns fifteen degrees an hour, that is thirty-three and three-quarter degrees west of Greenwich. Every four minutes of difference is one degree. The method is simple. Everything depends on the watch.

6So the watch has to be right, or at least wrong in a known way.³3A navigator never reset the chronometer. He wrote down how far it was wrong (its error) and how much that changed each day (its rate), and applied both in his sums. Opening the case to move the hands would have done more harm than any drift. A second of error in the chronometer is fifteen seconds of arc in the position, which on the equator is about a quarter of a nautical mile.⁴4The earth turns 360 degrees in 24 hours, which is 15 seconds of arc for every second of time. A minute of arc of latitude is one nautical mile, so 15 seconds of arc is a quarter of one. Away from the equator, the lines of longitude close in and the error shrinks with them. A watch that drifts by a few seconds a day will, after a month, have quietly moved the ship by miles, and miles are the difference between a harbour mouth and the rocks beside it.

7Before the ball, a captain who wanted to check his chronometer against Greenwich had to send it ashore, or send a mate with a pocket watch to compare against the observatory’s clocks and row back with the result. Neither is a good thing to do with a delicate instrument. The time ball turned the problem around: instead of bringing the watches to the time, it sent the time out to the watches, over open water, at the speed of light.

8The idea came from a naval officer, Robert Wauchope, who put up the first one at Portsmouth in 1829. Four years later the Astronomer Royal, John Pond, installed the one at Greenwich, on the roof of Flamsteed House, the oldest building of the observatory, and with breaks for storms and repairs it has dropped ever since.

Why one, and not noon

9Noon would seem the obvious hour. It is the hour astronomers measure, the hour the sun itself marks. That is exactly why it was not chosen. At noon the astronomers of Greenwich were busy at their instruments, timing the sun as it crossed the meridian, and nobody could be spared to work the ball. One o’clock gave them an hour to finish, check the clock and get to the roof.

10The odd hour stuck. Time balls in other ports followed Greenwich and dropped at one, while those in the United States, starting with the Naval Observatory in Washington in 1845, mostly dropped at noon.⁵5The Washington ball dropped at noon, local time, from 1845. Ships in American ports had the same problem as ships on the Thames; they were simply solved by a different lunch hour. Nobody in a hurry would design a public time signal around the lunch hour of a small staff. But that is what the signal was: not an abstraction handed down from the sky, but a promise made by particular people who would be at the top of a building at a particular moment.

11I like the five minutes of warning more than the drop itself. At five to one the ball rises halfway, and that is the instruction: stop what you are doing, find the mast, have the chronometer open and the pencil ready. At two minutes to one it goes to the top, and you watch it. The signal is not the moment the ball lands but the moment it starts to move. Anyone who looks away at one o’clock and back again has missed it, and there is no second chance until tomorrow.

The gun and the ball

12Some cities preferred noise. In Edinburgh a gun has been fired from the castle at one o’clock since 1861, and it still is, except on Sundays and a few holidays. A gun works in fog, which a ball does not, and it reaches people who are not looking at the right roof. It has one fault, which anyone who lives in Edinburgh learns sooner or later: sound is slow.

13Sound covers about 343 metres a second in air at room temperature. Stand a kilometre from the castle and the bang arrives three seconds late; at two kilometres, six. For setting a kitchen clock this does not matter. For rating a chronometer it is a disaster, unless you know your distance and subtract.⁶6At about 20 °C. Sound is a little slower in cold air, about 331 metres a second at freezing, so on a January afternoon the gun is later still. The railways had a noise of their own.

The startings and arrivals of the cars are now the epochs in the village day. They go and come with such regularity and precision, and their whistle can be heard so far, that the farmers set their clocks by them, and thus one well-conducted institution regulates a whole country.

Henry David Thoreau, Walden (1854), “Sounds”

14Thoreau was writing about the line that ran past Walden Pond, in Massachusetts. The farmers there had no observatory; they had a whistle, which was late by however far they lived from the track. Edinburgh solved the same problem by doing both: a time ball on the Nelson Monument on Calton Hill, which ships in the Firth of Forth could see, and the gun for everyone else.

15A ball is seen, not heard. From across the river at Greenwich the light takes millionths of a second to arrive, which is no delay that any sailor could have measured. That is what I meant by honest. The ball has nothing to correct for. Whoever sees it begin to fall sees it at the same moment, wherever they stand, as long as they can see it at all.

16The ball has its own faults. It cannot be seen at night, in fog or in heavy rain, and it is not raised on very windy days, because a ball travelling up and down a mast in a gale is a danger to the mast and to the people beneath it. A gun can be fired in any weather. A ball has to be believed, which is perhaps why so few of them are left.

Wires

17The telegraph changed what a time ball was. Until about 1850 a ball could only report the time of the clock beside it. With a wire, it could report the time of a clock a long way off, and do so without anyone on the spot needing to know what time it was.

18The best example is at Deal, on the Kent coast, where ships waited in the Downs, the sheltered water between the shore and the Goodwin Sands, for a wind to take them down the Channel. In 1855 an old Admiralty signal tower on the seafront was given a time ball, and from then on it dropped at one o’clock when an electric signal arrived from Greenwich, about a hundred kilometres away. The keeper at Deal raised it at five to one and at two minutes to one, exactly as at Greenwich, and then waited for the wire to let it go.⁷7The line was not always reliable: on some days the signal arrived too weak to release the ball, or not at all. For years the wire ran underground without proper insulation, which near the sea is asking for trouble. It worked, on and off, until 1927.

19The railways had already found that one time was easier to run than many. The Great Western Railway kept London time on its line from November 1840, and in 1847 the Railway Clearing House recommended that all its member companies do the same. For a while Britain ran two times at once: the railway’s, and the one the sun gave each town. The clock on the Exchange in Corn Street, Bristol, still has two minute hands, about ten minutes apart, one for Bristol and one for London. Greenwich time did not become the legal time of Great Britain until 1880.

There was even railway time observed in clocks, as if the sun itself had given in.

Charles Dickens, Dombey and Son (1848), chapter 15

20Dickens wrote that as a complaint, about a London neighbourhood torn up for a railway. Read now, it is a fair description of what the time ball and its wire had done: the sun had not given in, but it had been overruled by a clock in a building on a hill.

What a ball is for now

21Radio made the time ball unnecessary. From 1924 the BBC broadcast six short pips before the hour, and a ship could hear Greenwich anywhere it could pick up the signal, in fog, at night, without a mast in sight. Chronometers were rated by radio, then by satellite, and now most of the world takes its time from atomic clocks without thinking about it, as part of the phone it was filming the ball with.

22Yet the balls keep falling. At Lyttelton, the port of Christchurch in New Zealand, a time ball station built in 1876 was damaged in the earthquakes of 2010 and 2011, and its tower came down in an aftershock that June. It was rebuilt, and reopened in November 2018.⁸8I have not been to Lyttelton. What I know of it comes from photographs of the rebuilt tower, which looks, in the way of rebuilt things, slightly more certain of itself than the old one did. Nobody in Lyttelton needed it to set a chronometer. They rebuilt it anyway, the way you repair a clock that belonged to a grandparent: not because you need it to tell you the time, but because it is the one that always did.

23The most famous time ball in the world does not tell the time at all. Since the last night of 1907 a ball has been lowered down a pole in Times Square, New York, to mark the start of the new year. It is a time ball turned upside down: it moves slowly for a minute and arrives at midnight, instead of waiting at the top and leaving at one. But the grammar is the same. Everyone looks at one thing, at one moment, and agrees.

Five minutes of warning

24I stayed in the courtyard after the drop. Most of the crowd left at once, the way people leave a firework display. The ball sat at the bottom of its mast, and a man near me checked his watch against it, and then, I think, against his phone, and nodded to himself as though both had passed.

25What the ball gives is not really the time. Anyone in that courtyard had the time to the nearest millisecond in a pocket. It gives something rarer: five minutes in which a number of strangers agree to watch the same thing, and then one moment that is exactly the same for all of them. The warning matters more than the drop. A signal you have to prepare for is a signal you are likely to catch.

26I have started keeping my own small version of it. At five to one on a working day I stop what I am writing, and at one I start again. It does not make me more punctual. It makes the day have a middle, which is more than most clocks manage.

Notes

1. The times are those the Royal Observatory gives for its visitors. The ball runs on Greenwich Mean Time in winter and on British Summer Time in summer, so in summer it falls at noon by the sun’s reckoning, give or take. ↩

2. Watch the top of the mast, not the bottom. By the time the ball lands, the moment it was announcing is already gone. ↩

3. A navigator never reset the chronometer. He wrote down how far it was wrong (its error) and how much that changed each day (its rate), and applied both in his sums. Opening the case to move the hands would have done more harm than any drift. ↩

4. The earth turns 360 degrees in 24 hours, which is 15 seconds of arc for every second of time. A minute of arc of latitude is one nautical mile, so 15 seconds of arc is a quarter of one. Away from the equator, the lines of longitude close in and the error shrinks with them. ↩

5. The Washington ball dropped at noon, local time, from 1845. Ships in American ports had the same problem as ships on the Thames; they were simply solved by a different lunch hour. ↩

6. At about 20 °C. Sound is a little slower in cold air, about 331 metres a second at freezing, so on a January afternoon the gun is later still. ↩

7. The line was not always reliable: on some days the signal arrived too weak to release the ball, or not at all. For years the wire ran underground without proper insulation, which near the sea is asking for trouble. ↩

8. I have not been to Lyttelton. What I know of it comes from photographs of the rebuilt tower, which looks, in the way of rebuilt things, slightly more certain of itself than the old one did. ↩