The seatbelt sign has been on for twenty minutes. Below the wing, the same patchwork of Hertfordshire fields drifts past for the second time, then the third. The aircraft banks gently right, levels, flies straight for a minute, banks right again. Somewhere ahead, a controller is lining up a queue of airliners into a single stream for the runway, and yours is not next.

You are in a hold. More precisely, you are flying one lap of a racetrack-shaped pattern whose shape, size, speed and direction were fixed long before you boarded, and whose geometry has been calculated so precisely that the aircraft circling 1,000 feet above you never comes close.

Holding patterns are one of aviation’s least glamorous procedures and one of its most important. They are how air traffic control parks aircraft in the sky. Here is how they work, from the racetrack itself to the four famous stacks around London Heathrow.

What a hold actually is

The FAA defines a holding procedure as a predetermined manoeuvre that keeps aircraft within a specified airspace while awaiting further clearance. Strip away the formal language and it is a parking space with an engine running.

Aircraft hold for all kinds of reasons. Most often it is simple congestion: more aircraft arriving at an airport than its runways can accept. But a hold can also absorb a thunderstorm parked over the approach, a runway closed for snow clearance, an aircraft in difficulty that needs priority, or a crew that needs a few minutes to sort out a technical problem before committing to an approach.

Every hold is built around a holding fix: a point the crew can identify precisely, traditionally a radio beacon such as a VOR or NDB, today just as often an RNAV waypoint defined by satellite navigation. The fix marks where the pattern starts and where it ends. Everything else, the turns, the legs, the protected airspace around them, is measured from it.

The racetrack: right turns and one-minute legs

A standard hold is a racetrack. The aircraft flies the inbound leg towards the fix. Over the fix it turns 180 degrees, flies the outbound leg parallel to the inbound course in the opposite direction, then turns 180 degrees again to rejoin the inbound leg and arrive back over the fix.

In a standard pattern, both turns are to the right. Left-hand holds exist, often where terrain or neighbouring airspace makes a right-hand pattern impractical, but they are non-standard and have to be either charted or explicitly given by the controller. The side of the inbound course on which the racetrack lies is called the holding side; the other is the non-holding side, and that distinction becomes important the moment a pilot has to work out how to enter.

The turns are flown at standard rate, three degrees per second, so a 180-degree turn takes about a minute. The ICAO criteria expect a manually flown hold to use an average bank angle of at least 25 degrees or a rate of turn of three degrees per second, whichever needs the lesser bank. At airliner speeds the 25-degree limit usually wins, which is why holds at high altitude are bigger than the textbook version.

The straight legs are defined by time. In the United States and under ICAO rules alike, the inbound leg is one minute long at or below 14,000 feet and one and a half minutes above it. One full lap at low level therefore takes about four minutes in still air. For long delays, controllers can assign longer legs, and aircraft with DME or GPS may be given legs in nautical miles instead of minutes.

Airline captain and YouTuber Captain Joe walks through what a holding pattern is and why air traffic control uses it.

Wind, and why the outbound leg is the one that changes

Here is the subtle part. It is the inbound leg that has to come out at one minute. The outbound leg is whatever it needs to be to achieve that.

With a headwind on the inbound leg, the aircraft covers less ground per minute heading towards the fix, so the pilot shortens the outbound leg. With a tailwind inbound, the outbound leg gets longer. The FAA’s Aeronautical Information Manual tells pilots to fly the first outbound leg for one minute or one and a half, as appropriate to altitude, then adjust subsequent outbound legs as necessary to achieve the proper inbound leg time. Pilots also crab into the crosswind on both legs so the racetrack does not drift off its protected area.

Modern flight management systems do all of this automatically, which brings a quirk of its own. The FAA warns that some systems choose the entry using ground track rather than heading, fly a fly-by turn at the fix where the design assumes a fly-over, or use shallower bank angles at altitude. Individually these differences are small; cumulatively they can push an aircraft to the edge of the protected airspace and sometimes beyond it.

Getting in: direct, parallel and teardrop entries

An aircraft rarely arrives at a holding fix already pointing down the inbound course. It might approach from the side, from behind, or head-on. So there are three ways in, and which one to use depends on the direction the aircraft is coming from relative to the hold. The dividing line is drawn at 70 degrees to the inbound course on the holding side, splitting the compass around the fix into three sectors, with five degrees of leeway at each boundary.

Direct entry. Coming in from roughly the direction of the inbound course, the aircraft simply flies to the fix and turns straight onto the outbound leg, as if it had been holding all along.

Parallel entry. Approaching from the non-holding side, the aircraft crosses the fix, turns to fly parallel to the inbound course outbound on the non-holding side for one minute, then turns back towards the holding side through more than 180 degrees and returns to the fix or intercepts the inbound course.

Teardrop entry. Approaching from the narrow sector on the holding side, the aircraft crosses the fix, turns outbound onto a heading 30 degrees off the reciprocal of the inbound course, inside the holding side, flies that for one minute, then turns to intercept the inbound course. Viewed from above, the track looks like a teardrop.

Instrument students tend to agonise over which entry applies. In practice the important thing is staying inside the protected airspace, which the FAA explicitly says is designed around these three entries. That is why the entries are recommended rather than mandatory, and why the five-degree flexibility exists at each sector boundary.

Flight instructor channel FlightInsight shows a simple way to visualise which of the three entries applies, using the heading indicator in the cockpit.

Speed limits in the hold

The protected airspace around a hold is sized for a particular speed. Fly faster, and the turns get wider than the designers allowed for. So holding patterns come with maximum speeds, and the FAA and ICAO tables differ.

In the United States, the limits are 200 knots indicated from the minimum holding altitude up to 6,000 feet, 230 knots from 6,001 to 14,000 feet, and 265 knots above that. Some holds between 6,001 and 14,000 feet are restricted to 210 knots, shown by an icon on the chart. Holds at US Air Force airfields allow 310 knots unless otherwise depicted, those at Navy fields 230.

ICAO’s table, used in most of the rest of the world, allows 230 knots up to 14,000 feet, 240 knots up to 20,000 feet, 265 knots up to 34,000 feet and Mach 0.83 above that, with 170 knots for the slow Category A and B aircraft. A crew that cannot comply with a holding speed limit has to tell the controller.

Slowing down is part of arriving at the fix, too. The FAA spells out exactly when to start.

The EFC time: knowing when you will leave

A hold without an end is a problem, so the clearance always includes one. When a controller sends an aircraft to a fix and expects delay, the FAA says it is the controller’s responsibility to issue complete holding instructions unless the pattern is charted, an expect further clearance time and a best estimate of any additional delay.

The EFC is not a promise. It is a planning time. It tells the crew how long they are likely to circle, which is what they need to know to decide whether their fuel allows them to wait or whether they should divert while they still comfortably can.

It also has a second, quieter job. If the radio fails while an aircraft is holding, US rules tell the crew what to do with it: leave the clearance limit at the expect-further-clearance time if one has been received, and proceed to a fix from which an approach begins. The EFC is how the controller and a silent aircraft keep the same plan.

Stacks: holding in layers

One aircraft in a hold is simple. Ten aircraft at the same fix need a system, and the system is the stack. Aircraft holding at the same fix are separated vertically by at least 1,000 feet. New arrivals join at the top. The aircraft at the bottom is cleared out of the hold to start its approach, and everyone above is stepped down 1,000 feet in turn.

At the busiest airports a dedicated controller may manage each stack, and an airport may have several holding patterns, arranged by the direction aircraft arrive from, the runway in use or the limits of the surrounding airspace. Aircraft declaring an emergency get priority and can bypass the hold, which means everyone already in it waits a little longer.

The fuel on board is what decides how long a crew can play along. Under US rules, an aircraft flying under instrument flight rules must carry enough fuel to reach its destination, then its alternate, then fly for 45 more minutes at normal cruising speed. Airline flight planning normally adds extra fuel on top when delays are expected. When it starts to run short, the crew tells the controller, using a phrase with a precise meaning.

Minimum fuel is not an emergency. It is a warning that one is possible if the delay grows. If the fuel expected on landing drops below the planned reserve, the next step is a MAYDAY FUEL call, which does get priority. Holds can last a very long time when things go wrong: in May, Afterburner reported on a Garuda A330neo that spent four and a half hours holding off the Indian coast.

Why Heathrow has four stacks

London Heathrow is the best-known example of all this in Europe, and the reason is its geometry. The airport handles an enormous flow of arrivals on two runways. Arriving aircraft are routed to one of four holding points, one for each quarter of the compass they come from, and the arrival routes that end there define four stacks.

Each is anchored on a VOR beacon out in the countryside around London. Bovingdon (BNN), in Hertfordshire, takes arrivals from the north-west. Lambourne (LAM), in Essex, takes the north-east. Biggin Hill (BIG), in the south-eastern corner of Greater London, takes the south-east. Ockham (OCK), in Surrey, takes the south-west. The airspace above Bovingdon airfield and the nearby town of Chesham is known simply as the Bovingdon stack.

From the four stacks, approach controllers merge the traffic into a single stream for the landing runway, sometimes with aircraft as close as 2.5 nautical miles apart on final approach. Once an aircraft is established on final, control passes to Heathrow Tower. The stacks are the buffer that makes that tight sequence possible: they keep a reservoir of aircraft ready to fill every gap the runway can offer.

They also cost fuel and spread noise over the communities beneath them, which is why air traffic services try to absorb as much delay as they can further back along the route, slowing aircraft down while they are still en route so that fewer of them need to circle. The stacks remain the safety valve when that is not enough.

Heathrow Airport’s own explainer on how its arrivals operation works, from wind and weather to the sequencing of aircraft onto two runways.

The racetrack in the sky

A hold looks like the most boring thing an aircraft can do. Round and round, nothing happening, passengers checking their watches.

In reality, every part of it has been measured: which way to turn, how long to fly straight, how fast to go, how to get in from any direction and how to get out even with a dead radio. The airliner circling over Surrey is flying the same geometry, to the same minute, as a student pilot practising holds in a single-engine trainer.

That is the point. Holding patterns work because everyone flies them exactly the same way.

The view from the flight deck while holding for London Heathrow, filmed by the aviation channel Just Planes.

Sources: FAA Aeronautical Information Manual, chapter 5, section 3, paragraph 5-3-8 Holding; 14 CFR 91.167 and 91.185 (eCFR); SKYbrary, Holding Pattern and Minimum Fuel (citing ICAO PANS-ATM, Doc 4444); Wikipedia, Holding (aeronautics), Heathrow Airport, Bovingdon, Biggin Hill, Lambourne and Ockham, Surrey.