The friction problem
Open a sapphire caseback on a quiet evening and the picture looks placid: gears turn slowly, a balance ticks, a rotor wanders. But inside that placid picture, every pivot rotating in a jewel, every gear tooth meshing with a pinion, every pallet stone striking an escape-wheel tooth is generating friction — and the duty cycle is staggering. A 4 Hz balance completes over 250 million oscillations a year; the escapement endures over half a billion impacts between services. Run dry, those contact points would wear themselves out of tolerance within months. Everything depends on a few milligrams of oil, placed with surgical precision, doing their job for years at a time.
Different parts of the movement need different lubricants because the contact conditions differ radically. The barrel arbor turns slowly under the mainspring's full torque: it needs a heavy oil or grease that stays put under load. The balance pivots spin rapidly under almost no load: they need the thinnest oil, in the smallest possible quantity, so as not to dampen the oscillation the whole watch depends on. The pallet stones and escape-wheel teeth meet in high-frequency sliding impacts: they need a thixotropic grease engineered to stay at the contact point under dynamic stress. One universal oil applied everywhere would do every one of these jobs badly — which is precisely what distinguishes a real service from a cheap one.
Where each lubricant goes
The distribution of lubricants is as specific as any other engineering specification. A trained watchmaker in Le Locle, Glashütte, or Shiojiri keeps half a dozen labelled pots beside the bench — Moebius 9010 for fast pivots, HP-1300 for the slow train, 9415 for the pallet faces, D5 for high-load arbors, braking grease for the automatic bridle — each applied with an oiler whose tip carries a fraction of a milligram, under magnification, to a precisely defined point. Too little and the surface runs dry; too much and the oil spreads where it should not, wicking up a pivot or onto a hairspring, where a single droplet can stick coils together and wreck the rate. The keyless works (stem, sliding pinion, setting levers) take their own greases; chronograph levers, calendar cams, and striking trains each have specified points and products. Modern movements add a further refinement: epilame, an invisible anti-spread coating applied to escapement parts so the oil beads and stays exactly where it was placed for years rather than months.
How do watch oils actually fail?
Four mechanisms operate simultaneously, at different rates. Evaporation removes the oil's lighter fractions, leaving a thicker residue that lubricates worse and drags more. Migration moves oil away from the contact point by capillarity and centrifugal force. Oxidation slowly converts the oil into gummier compounds. And contamination loads what remains with microscopic metal particles worn from the very surfaces it protects, turning lubricant into lapping paste — abrasive rather than protective. This is also why the trade abandoned natural oils: the legendary chronometer oils of the nineteenth century, rendered from porpoise and blackfish jaw fat, were prized precisely because they gummed more slowly than vegetable oils, and the fully synthetic esters developed since the mid-twentieth century beat them on every axis. Lubrication chemistry, like hairspring metallurgy, is one of horology's invisible revolutions.
The degradation shows up on a timing machine before it shows up as damage. In the first years after service, the movement runs at healthy amplitude — the balance swinging through 270 degrees or more. As friction creeps up, amplitude sags; rate stability across positions deteriorates; and eventually the escapement is operating with lubricant that has become abrasive. A movement run hard for fifteen or twenty years without service can grind measurable wear into pallet faces, pivot shoulders, and the barrel arbor — wear that a service can arrest but never undo.
The three-year service interval of the 1970s reflected the oils of the 1970s. Modern synthetics, epilame treatments, harder coatings, and low-friction escapement geometries (Omega's co-axial, Rolex's Chronergy, silicon components that need no oil at all at certain interfaces) have stretched honest recommendations to five to ten years — Rolex now warrants its service work for a decade and recommends roughly that. The physics has not been repealed, merely slowed: every oil in every watch is still evaporating, migrating, oxidising, and collecting debris from the moment it is applied.
The service interval is physics
When a watchmaker recommends service every five to seven years, that is a description of how long the lubricants can be expected to keep protecting the movement under normal wear — not a revenue programme. The practical implications for collectors are simple. A watch unserviced for fifteen years is not merely "due"; it has been running on degraded lubrication for most of that time, and the wear accumulated is real. A documented history of regular service by competent hands is worth paying for when buying, and worth maintaining when owning. And the corollary cuts the other way: a watch that sits unworn in a drawer degrades too — oils migrate and dry whether or not the wheels turn — so "barely worn since its last service in 1998" is a warning, not a reassurance. The cheapest service is the one done on time; the most expensive is the one done ten years late, when worn parts must be made or sourced rather than simply cleaned and re-oiled.
A neglected movement does not fail all at once. It slowly turns friction into damage — pivot wear, tooth rounding, jewel scoring — accumulating quietly for years before anything is visibly wrong. The service interval exists because physics is patient.