Before 1675 the balance of a watch had no rhythm of its own. It swung as hard as the mainspring pushed it, so any change in the spring’s force changed the rate, and watches drifted by many minutes a day.
The balance spring never appears on the dial. It can be seen through a glass caseback: a wheel swinging back and forth, and at its centre a fine coil tightening and opening.The spiral links the balance to the frame of the watch. The further the balance swings, the harder the spring pulls it back, which is what keeps every swing the same length of time.
Christiaan Huygens’s spiral gave the balance a natural period, as a pendulum has in a clock. Accuracy improved from a quarter of an hour a day to a few minutes, and watches gained minute hands for the first time as a matter of course.
What it traded away is stability in temperature: steel springs weaken as they warm, and three centuries of work, from compensated balances to modern alloys and silicon, has gone into correcting that. Every mechanical watch still has one.
A pocket watch spends its day upright in a waistcoat pocket. In that position gravity pulls the balance and its spring slightly off centre, and the watch gains or loses a little differently in every position it is carried. Breguet’s answer, patented in 1801, was not to fight gravity but to average it out.
On the dial, a tourbillon usually shows itself through an opening, most often at six o’clock, where the escapement can be seen turning. Some makers show it only on the back.The balance, lever and escape wheel ride in a cage that the watch’s train turns once a minute. The escape pinion rolls round a wheel fixed to the plate, so the escapement keeps working as the cage turns.
The price is complexity and energy. The cage is among the hardest assemblies in watchmaking to make, adjust and balance, and turning it takes power that would otherwise go to the escapement. On the wrist, which changes position constantly, its benefit to timekeeping is disputed.
Breguet sold his first tourbillon in 1805 and made about thirty-five. The device stayed a rarity for a century and a half, then returned from the 1980s as a showpiece of fine watchmaking, which is mainly what it is today.
A mainspring is a poor power source. It has to be wound, and it delivers falling torque as it unwinds; two centuries of horological ingenuity went into hiding that fact. By the early 1950s the miniature battery made an alternative conceivable: a cell that delivers constant voltage for a year and needs no winding at all.
Hamilton kept the balance and replaced only the mainspring. Bulova replaced the balance too, with a fork whose frequency comes from its own shape rather than from a spring.
The first attempts kept everything else. Hamilton’s calibre 500 of 1957 retained a balance wheel and a gear train, but the balance carried a coil, and current from a button cell produced a magnetic field that pushed against permanent magnets in the plate to keep it swinging. There were no transistors: switching was done by mechanical contacts, and those contacts were the movement’s weakness. Bulova’s Accutron of 1960 went further and removed the balance altogether, replacing it with a tuning fork vibrating 360 times a second, driven by coils and switched by a transistor. A fork’s frequency is set by its own geometry, so it is far less sensitive to position and shock than a balance.
Neither survived long. The contacts in the Hamilton wore and arced; the Accutron was displaced within a decade by quartz, which uses the same idea of a solid-state resonator driven by a transistor, at a frequency a hundred times higher. Both are best understood as the bridge between the mainspring and the quartz crystal.
Everything else a watch does is passive: it shows the time to whoever looks. An alarm is the one function that reaches out to the wearer without being asked. Alarm clocks had existed for centuries; the difficulty was doing it on the wrist, where there is no room for a bell and no sounding board except a slab of metal pressed against a sleeve.
The mechanism was never the difficulty. The difficulty was acoustic: a hammer needs something that can ring, which is why the Cricket has a second, resonant back.
The mechanism needs a second source of power, because a hammer beating for twenty seconds would flatten the going barrel. A separate mainspring, wound by its own crown or a second crown position, drives a small hammer through a train regulated by a friction brake. Setting is done by a disc under the dial carrying a notch: a lever rides the disc as the hour wheel turns, and when the notch comes round the lever drops and releases the hammer train.
The unsolved part was acoustic. Eterna built the first alarm wristwatch in 1914 and it found no market, because the sound went nowhere. Vulcain’s answer in 1947 was a resonant double case back, a membrane with an air gap behind it, so the hammer drives a sounding board rather than a damped lump of metal.
Time zones were standardised internationally from 1884, and within a generation the telegraph, the liner and the aeroplane made it ordinary for a person in one zone to need the hour in another. The conversion was done in the head, which is exactly where it goes wrong: booking a call, filing an order, making a connection.
Turn one ring once a day against a fixed list of cities and every time zone becomes legible at once. Whole-hour zones only: the half-hour offsets used by India and Iran cannot be shown.
Louis Cottier, working alone in Carouge, arrived at the solution around 1931. A ring marked with twenty-four hours is geared to the hour hand so that it makes one turn a day, and it is read against a fixed ring carrying the name of one city in each zone. The hour anywhere is read straight off, and every zone is legible at the same moment. Cottier refined it in 1953 by adding a second crown to index the city ring, which is the form the complication still takes.
The cost is legibility and precision. The dial carries two dense rings of small text around a conventional set of hands, and the mechanism assumes whole-hour offsets, so the half-hour and quarter-hour zones used by India, Iran, Nepal and parts of Australia cannot be shown correctly. Daylight saving, which moves in different months in each hemisphere, is not modelled at all.
A mechanical watch is limited by its balance and hairspring. The balance swings a few times a second and its rate changes with temperature, with position, with the state of wind and with the ageing of its oils. For most of history that was enough. Radio, long-distance telephony and precise navigation changed it, and physicists had known since the Curies’ work of 1880 that a quartz plate cut to shape will ring at a sharply defined frequency when driven electrically, and hold it far more stably than any mechanical resonator.
32,768 is two to the fifteenth power, so fifteen halvings turn it into exactly one pulse a second. It is also above hearing, and low enough to keep the circuit cheap and the drain small.
The first quartz clock, built at Bell Telephone Laboratories in 1927, filled a room. Shrinking it to the wrist took the transistor and the integrated circuit, and forty years. What made it small was the choice of frequency. A chain of fifteen flip-flops, each halving the one before, turns 32,768 Hz into exactly one pulse a second with no arithmetic and almost no power; the frequency is also above the range of human hearing, so the watch does not whistle, and low enough that the fork can be tiny and the drain small. The pulse drives a Lavet stepping motor, whose asymmetric geometry advances the rotor exactly half a turn per pulse, in one direction.
The compromises are different in kind, not absent. Quartz frequency varies with temperature, falling away from a turnover point set at around 25 degrees, so a watch on the wrist sits near its optimum while one left in a cold drawer runs slow. Crystals age. The battery is a consumable and will destroy the movement it powers if left to leak. And circuit boards cannot be repaired, only replaced, so a quartz watch of the 1980s is often beyond economic repair in a way that a mechanical watch of the 1880s is not.
A watch is a set of steel pivots turning in jewelled bearings, lubricated with oil. Water corrodes the steel, washes away the oil and leaves deposits that bind the train; condensation under the crystal will rust a hairspring within days. The pocket watch was largely protected by the pocket. On the wrist the watch met rain, washing, perspiration and immersion, and the wristwatch of the 1910s was notoriously short-lived for exactly that reason.
Every seal is a ring of rubber squeezed between two rigid faces, and pressure from outside squeezes it harder. The crown was the last opening to be solved, and Rolex’s clutch is what made a screwed-down crown usable.
The case had three unsealed openings: the joint at the back, the joint at the crystal, and the hole the winding stem passed through, which had to stay free enough to turn. Screwed cases such as Francois Borgel’s of 1891 dealt with the first two. The stem defeated everyone until the crown itself was made into a closure: a screw-down crown, with a clutch so that it can be turned without winding when closed, screwing onto a threaded tube. Combined with a threaded bezel and case back compressing gaskets, that produced the Oyster case of 1926.
Gaskets are consumables. They take a set, harden with age and are attacked by soaps and solvents, so a depth rating is only as good as the last service, and a crown left unscrewed makes it meaningless. The figures on the dial are also commonly misread: a rating in metres is the result of a static pressure test, not a statement about depth in use. ISO 6425 sets what a watch must actually do before it may be called a diver’s watch.
The day of the week is a different question from the date, and for most people the more urgent one. It governs whether an office is open, when a delivery is due, when a market trades. Calendar watches of the 1930s and 1940s showed both, but generally cut the day to three letters because there was no room to spell it out. Rolex’s Day-Date of 1956 fixed the form the complication has kept since.
The day is the easier of the two: seven teeth, no irregularities, no correction ever needed while the watch runs. Only the date has to be put right at the end of short months.
Mechanically the day is the simpler indication, because the week has no irregularities. The same twenty-four-hour wheel that drives the date carries a finger which, once per revolution, advances a seven-toothed star bearing the names of the days, where a jumper holds it. Since the week never varies in length, the day needs no correction as long as the watch runs. In an instantaneous construction both displays are driven by energy stored in a cam and released within a second or two of midnight.
The second disc costs height and dial area, and it doubles the load released at midnight, so the demand on the mainspring at the change is greater than for a date alone. The curved aperture of the spelled-out type needs a large disc with its printing arranged on an arc, which is part of why the complication was slow to reach smaller cases.
The date is the most frequently consulted piece of calendar information in ordinary life, and until the middle of the twentieth century anyone who wanted it consulted a diary, a newspaper or a colleague. It is nearly free to a watch, which already counts days. What delayed it was space: the date ring and its indexing works occupy the outer margin of the movement, which in a small wristwatch calibre was already crowded.
The watch already counts days: the hour wheel turns twice in twenty-four hours. Halve that, hang a finger on it, and once a day it pushes the date ring on by one tooth.
The drive is a chain of reductions. The hour wheel turns once in twelve hours and drives an intermediate wheel geared two to one, so that wheel turns once in twenty-four. It carries a finger which, at each revolution, engages one tooth of the date ring; a sprung jumper bearing on the teeth holds each numeral square in the aperture. In the commonest arrangement the finger tensions the jumper until it passes the crest of a tooth and snaps the ring forward. An instantaneous calendar instead stores energy through the evening in a separate cam and releases it in a fraction of a second at midnight, at the cost of a sudden demand on the mainspring.
The mechanism puts an obligation on the wearer: five corrections a year, and three days after a common February. The chain of reductions also creates the danger zone. For roughly the hours between nine in the evening and three in the morning the finger is engaged with the date ring, and forcing a quickset then can bend the finger or break a tooth.
A hand-wound watch must be wound daily, and each winding turns the crown, which turns the stem: the largest opening in the case and the main route by which dust, perspiration and water reach the movement. There was also the spring itself. A watch wound each morning runs at its highest torque in the hours afterwards and its lowest before the next winding, so its rate wanders across the day, and a watch forgotten for a day stops.
Harwood’s weight bounced between sprung buffers and wound in one direction only. Removing the buffers let the weight keep its momentum all the way round, which is the arrangement almost every automatic watch has used since.
John Harwood patented the first workable self-winding wristwatch in 1923 and reached production through Fortis in 1926. His weight was a pivoted segment swinging through a restricted arc between two buffer springs, winding in one direction only, and his watch had no crown at all: the hands were set by turning the bezel. Rolex’s Oyster Perpetual of 1931 removed the buffers and let the weight turn through a full circle, which wastes no energy striking anything and harvests small movements of the wrist far better.
Early rotors still wound in one direction, the other simply free-wheeling. Bidirectional winding uses reverser wheels, each a wheel and pinion coupled by hooked pawls that lock one way and slip the other. Two of them sit in the train so that whichever way the rotor turns, one locks and drives while the other slips, and the drive delivered to the barrel is always in the winding direction.
The costs are height and wear. A rotor and its reduction train add roughly two millimetres, which is why automatic chronographs and automatic calendars arrived so much later than their hand-wound equivalents, and the rotor bearing is the hardest-worked bearing in the watch. Because the wearer cannot be stopped from winding indefinitely, the mainspring is not anchored to the barrel wall but held by a sliding bridle that slips when the spring is full.
