What Does a Clock Show When a Leap Second Is Added?

1. Where the Extra Second Goes, and Why Japan Sees It at 8:59 a.m.

A leap second is added to Coordinated Universal Time, or UTC, the time scale that clocks around the world are set from. It goes at the very end of a day, so UTC runs 23:59:58, then 23:59:59, then 23:59:60, and only then 00:00:00 of the next day[2]. Leap seconds have usually been placed at the end of June or December, though I could not confirm that in the source I read.

The same instant gets different clock numbers in different places, because each place sits a different number of hours from UTC. Japan Standard Time is nine hours ahead of UTC, so the instant that is 23:59:60 in UTC is 8:59:60 the next morning in Japan. The Japanese institute that keeps the country's time, NICT, announced exactly that for July 1, 2015: a "60th second" inserted between 8:59:59 and 9:00:00 a.m.[1]. (The conversion for other places is my own arithmetic from the time-zone offsets. On the east coast of the United States it works out to 6:59:60 p.m. on December 31 in Eastern Standard Time, or 7:59:60 p.m. on June 30 in Eastern Daylight Time.)

So the official clock is simple to describe. The hard part is the next question: what do ordinary clocks do when the official time has a number they were never built to show?

2. Why Most Clocks Cannot Display a 60th Second

The 2015 NICT announcement says something surprising. The "60" display could not be seen on ordinary clocks. People who wanted to see it had to visit NICT's headquarters near Tokyo, where a large display of Japan Standard Time was set up for the occasion, and the institute opened part of its grounds for visitors[1].

NIST, the United States' measurement institute, explains why. The label 23:59:60 cannot be written in many digital time formats, so a computer has to do something else instead[2]. A simple way to picture it is a scoreboard built with two digits that only count 00 to 59. It has no way to put a 60 on the board, and nobody can add a new digit overnight to every clock, phone and server in the world.

3. Three Ways a Clock Can Get Through the Extra Second

Freeze on 59. NIST's public time servers, which computers on the internet use to set their clocks, "add the extra leap second by stopping the clock for one second at 23:59:59"[2]. The display holds on 59 for two seconds' worth of real time, then moves on to 00:00:00.

Repeat the next day's first second. The same NIST page says that some systems "implement the leap second by repeating second 0 of the next day"[2]. Here the extra second shows up as a second 00:00:00, on the wrong side of midnight. NIST points out that with either method a record of when something happened can be ambiguous, because two different instants carry the same time label[2].

Spread it out ("smear"). Google has not used these sudden steps on its servers since 2008. Instead it "smeared" the extra second across the hours before and after each leap second[4]. Google now recommends one particular smear for anyone who does this: a 24-hour linear smear from noon to noon UTC[4]. Nobody sees a 60 at all. The trade is that smeared time is slightly off from official UTC for a while: Google says it deviates briefly during the smear and lines up again after 24 hours[4]. NIST also notes that spreading the second over hours causes small time and frequency errors compared with legal UTC while it happens[2].

How small is each nudge? Twenty-four hours is 86,400 seconds, so one extra second shared evenly over that time is about 1/86,400 of a second added per second, roughly 11.6 microseconds (millionths of a second) per second. That division is mine, not a number quoted by the source.

Here is a way to picture the three choices, as an illustration only. Imagine a board game with 60 squares in a ring, and you need to fit in one more turn. You can make a player wait one beat on the last square (freeze), let them step on the first square twice (repeat), or make every step the tiniest bit shorter so the whole lap gets longer without anyone noticing (smear). All three put the player back at the right place; each pays a different small price.

4. How a Radio Clock Is Warned Ahead of Time

Some clocks do receive the real 60. Japan's standard-time radio signal, JJY, sends a time code one bit per second. When a leap second is inserted, a 0 value is filled in at the 59th second, and the 60th second carries a position marker called P0[5].

The signal also warns clocks beforehand. One bit, called LS1, set to 1, means a leap second will occur within a month. A second bit, LS2, says which direction: 1 for an added second, 0 for a removed one[5]. For the leap second at the end of 2016 (9:00 a.m. on January 1, 2017, in Japan), NICT reported that both bits were set to 1 beforehand and were cleared to 0 at 9:00 a.m. that day. On the internet side, the Network Time Protocol (NTP) "leap indicator" was set to "01" from December 31, 2016, and cleared at 9:00 a.m. on January 1[6]. NIST's own servers, for their part, give notice from 00:00 UTC on the last day of the month[2].

So the people who run the world's time distribute both the extra second and a warning that it is coming. What each clock does with it is then up to the clock.

5. Why Add a Second at All, and Will We Keep Doing It?

Atomic clocks tick very steadily. The Earth's spin does not: NIST says it sometimes speeds up and sometimes slows down, though averaged over long periods it is gradually slowing[3]. A leap second is the patch that keeps UTC from drifting away from the time given by the Earth's rotation. It has been used so far only in the "add a second" direction, and a removal has never been needed[3].

The rules are about to change. Resolution 5 of the 2026 General Conference on Weights and Measures, the international body of member countries that oversees measurement, states that "continuous UTC will become effective on 20 May 2027," with the allowed gap between UTC and the Earth-spin time raised to 3,600 seconds, one hour[7]. If that goes ahead as planned, new leap seconds would stop, and the extra "60" that this page is about would become a thing of the past. (For the longer story of why a removed second worries computer engineers, see the page on whether the world's clocks could ever have to skip a second.)

6. Try It Yourself: Watch a Clock Roll Over, Then Read the Originals

You will probably not catch a leap second, because they come rarely and sometimes not for years. But you can watch a normal minute. Set a phone or computer clock to show seconds, and watch it count up to 59 and then start over at 0. An ordinary minute has 60 different numbers on it; a leap-second minute would need 61.

Then ask: if a clock could not show a 60, which of the three tricks would you pick, and what would you give up? Your phone's maker may have already chosen. The NIST page on its time service and Google's page on leap smear are short, and both are written for engineers who had to make exactly this choice.

Sources

Update history

• 2026-10-23 First published.