Every Leap Second Ever Added and Why They Exist

A leap second is a one-second adjustment added to Coordinated Universal Time (UTC) to keep our atomic clocks in sync with the Earth's slightly irregular rotation. Since the Earth doesn't spin at a perfectly constant rate, super-precise atomic time slowly drifts ahead of the time defined by the planet's actual rotation, so timekeepers occasionally insert an extra second to close the gap. Between 1972 and 2016, exactly 27 leap seconds were added, and every single one was inserted (never removed).

Why leap seconds exist

There are two different ways to measure time, and they don't agree with each other:

• Atomic time (TAI): Based on the vibrations of cesium atoms. It's incredibly stable and never varies. This is the "official" standard second.
• Astronomical time (UT1): Based on the actual rotation of the Earth, measured against distant stars. This is the time that keeps the Sun overhead at noon.

Here's the catch: the Earth is not a perfect clock. Tidal friction from the Moon gradually slows the planet's spin, and events like earthquakes and shifts in the molten core cause tiny irregular changes. On average, a solar day now runs a couple of milliseconds longer than the 86,400 atomic seconds it's supposed to contain. Those milliseconds pile up over months and years.

Left alone, atomic clocks would drift ahead of the Sun. It would take centuries for the difference to become noticeable to a person, but astronomers, navigators, and satellite systems care about sub-second precision. A leap second is the fix: whenever UTC and the Earth's rotation drift more than 0.9 seconds apart, one second gets added to pull them back together. If you want the deeper mechanics of these two time scales, we break them down in TAI vs UTC: why atomic time and coordinated time diverge.

Who decides when to add one

The decision belongs to the International Earth Rotation and Reference Systems Service (IERS), based in Paris. They continuously measure the Earth's rotation using radio telescopes and satellite data, then predict when the gap between UTC and UT1 will cross the 0.9-second threshold.

When a correction is needed, the IERS issues a "Bulletin C" about six months in advance. Rules from the standards body say a leap second can only be inserted at the end of two preferred dates:

• June 30 (23:59:60 UTC)
• December 31 (23:59:60 UTC)

December and June have been the only two months ever used. In theory, March 31 and September 30 are also allowed as backups, but they've never been needed.

Every leap second ever added

Here is the complete leap second list from the system's start in 1972 through the most recent one. The "TAI minus UTC" column shows the total accumulated offset after each insertion.

Notice the long gaps. There were seven leap seconds packed into the 1970s, but a full seven-year stretch (1999 to 2005) with none at all. That's because the Earth's rotation sped up slightly during that period, so less correction was needed.

How a leap second actually works

Normally the last minute of the day ends at 23:59:59 and rolls over to 00:00:00. On a leap second night, the clock does something impossible on any ordinary day: it shows 23:59:60 for a full second before rolling over.

So the sequence on December 31, 2016 looked like this:

23:59:58
23:59:59
23:59:60   ← the leap second
00:00:00

This matters a lot for computers, because most software assumes a minute has exactly 60 seconds (0 through 59). A time value of :60 can crash naive code or produce a duplicate timestamp. This is one reason understanding seconds versus milliseconds in Unix timestamps matters when you build systems that log precise events.

It's worth clearing up a common myth: Unix time (the epoch second count) does not actually store leap seconds. The Unix clock simply repeats or "smears" the second so its day always contains exactly 86,400 seconds. If you're curious how that base counter works, see our primer on epoch time and Unix timestamps.

The problems leap seconds cause

A single extra second sounds harmless, but it has repeatedly broken major systems:

• 2012 outage: The June 2012 leap second triggered kernel bugs that took down Reddit, Mozilla, Qantas airline reservation systems, and others when Linux servers locked up under high CPU load.
• Duplicate timestamps: A repeated second can make two different events look like they happened at the exact same moment, which corrupts logs and financial transaction ordering.
• Sync chaos: Not every server handles the second the same way, so machines that rely on NTP time synchronization can briefly disagree with each other.

To avoid these headaches, big tech companies like Google adopted "leap smearing," where they spread the extra second across many hours by slightly slowing down their clocks, rather than inserting one abrupt :60 tick.

The end of the leap second

In November 2022, the General Conference on Weights and Measures (CGPM) voted to abolish the leap second. The plan is to stop adding them by 2035 and instead let UTC and the Earth's rotation drift apart for decades before applying a single larger correction.

There's a fascinating twist in the decision. Because the Earth has been spinning slightly faster in recent years, scientists are now worried about the possibility of a negative leap second, where a second would be removed for the first time in history. That scenario worries engineers even more than adding one, since almost no software has ever been tested for a minute that ends at 23:59:58.

Adjusting clocks to match reality is an ancient problem. Civilizations have wrestled with it for millennia, from leap years all the way back to when the Gregorian calendar skipped 10 days in 1582 to realign the date with the seasons. The leap second is just the modern, ultra-precise version of that same balancing act.

See exactly how UTC time ticks right now

Curious how leap seconds affect the count? Watch the live Unix timestamp and UTC time update in real time, so you can see how the clock behaves second by second.

Open the live epoch clock →