Surprisingly, the answer to this question is not always a simple 'sixty'. While nearly every minute measured has sixty seconds, between 1972 and 2006 there were twenty-three minutes each of which had sixty-one seconds, and there is the possibility of more in the future. To add to the complexity it is possible that, at some point in the future, specific minutes will have only fifty-nine seconds.
Before explaining why (and when) this will happen, some information on the history of the science of measuring time is required.
In 1675, Greenwich Observatory was set up by King Charles II and became the recognised authority on the subject of longitude. Subsequent to an agreement reached in 1884 at an International Meridian Conference in Washington to set a specific location as the 'prime meridian' (ie, 0 degrees longitude), it was only a matter of time before Greenwich was accepted as this location1.
Greenwich and Greenwich Mean Time
As the ability to determine the precise longitude of a ship at sea was linked to the requirement to accurately measure time, Greenwich also became the focal point for this branch of science. Naturally, this meant the need for precise definitions of time that would work around the globe.
Definitions - Days, Seconds And Minutes
Although most people would consider a 'day' to be the amount of time taken for the Earth to rotate once about its own axis, the length of each day varies slightly with the seasons and from year to year. Because of this scientists at Greenwich defined an invariable day, called a 'mean solar day', based on the average of the Sun's apparent speed across the sky. This average or 'mean' calculation is the reason for the 'mean' in Greenwich Mean Time.
A second was then defined as 1/86400th of the mean solar day, with a minute being defined as 60 seconds. Because of the observed variance in the length of the mean solar day over time and the requirement for 'a day' to be a constant, the definition for a mean solar day was changed in 1960 and became based on the average of the length of each day in 1900.
GMT, UT, UT0, UT1 and UT2
In 1928, the International Astronomical Union recommended that Greenwich Mean Time (GMT) be referred to as Universal Time (UT). However, since UT is affected by the motion of the Earth rotational pole, UT as measured in one location would differ from UT as measured somewhere else. Consequently UT as measured at a particular location is known as UT0, while a UT measurement that has been 'corrected' for the effect of the motion of the rotational pole is called UT1.
However, UT1 is not a uniform timescale as the Earth does not spin at a constant rate, and so UT2 was established. UT2 is obtained from applying a formula which attempts to 'average out' the variations in the Earth's rotation over time. However, as the Earth's spin is gradually slowing down, the formula needs to be continually changed, and so UT2 is also not a uniform timescale.
Then, in 1967, a new measurement of time which did not use the Sun as reference was established. Known as UTC (Universal Coordinated Time), the new measurement went on to become the most commonly-used time scale in the world. It is a compromise between the highly stable atomic time and the time of day generated by the irregular rotation of the Earth (UT1). The definition of a second is no longer related to the length of a day, but to the radiation emitted from a caesium-133 atom2.
Unfortunately, over time the difference between UTC and UT1 increases; the solar day is now about 2.5ms longer than it was in 1820. If this difference were not addressed, it would result in the Sun not being directly overhead at the Greenwich Meridian at noon and, eventually, the sun not being visible from Greenwich at any time during the period defined by the clocks as 'daytime'3.
The Leap Second
In order to maintain UTC in approximate synchronisation with the rotation of the Earth while still basing it upon the invariable 'atomic second', the International Earth Rotation Service (IERS) has the authority to add or remove4 a second from UTC, which they will do as required to ensure that the absolute difference between UT1 and UTC remains less than one second.
Although this adjustment can be done at the end of March, June, September or December5, since the system was introduced in 1972 only dates in June and December have been used. As at January, 2007, there have been 23 leap seconds added to UTC 6. On these days, the last minute of the day had 61 seconds, with 23:59 and 59 seconds being followed by 23:59 and 60 seconds, then by the more usual 00:00 and zero seconds of the following day.
When is the Next Leap Second?
The IERS announces whether the option to add a leap second will be used around six months in advance. As the leap second is inserted at the same time all over the world, the specific minute which would contain 61 seconds therefore depends on the local time zone. Only locations which use UTC as 'local time' would add the second just before midnight, with time zones west of UTC adding their second sometime before midnight and those east of UTC adding their second sometime during the next day.
The announcement (known as 'Bulletin C') of the leap second which was added to the last minute of 31 December, 2005 (UTC) can be found on the IERS Website.
However, celestial navigation is rarely used today due to the increased availability of satellite navigation systems, and so the need to keep the difference between solar time and atomic time less than one second is currently the subject of scientific debate. This may lead to the concept of the leap second being dropped, depending on whether it is deemed to do more harm than good7.
Will the Insertion of the Next Leap Second Cause Any Problems?
For those very accurate timepieces which are set by atomic clocks which are synchronised to UTC, the next application of a leap second will simply cause them to display the extra second, something which some 'time-nuts' consider to be a sight worth seeing. Clocks which are driven by GPS signals may be one second fast for a few seconds but will then start to show the correct local time again. This is due to the fact that although GPS time is not adjusted for leap seconds, the GPS signal provides both GPS time and, periodically, the current difference between UTC and GPS. For clocks which need to be set by hand, the application of the leap second will merely cause them to become one second faster than they were beforehand.
However, complex systems which rely on continuous and highly accurate time measurements could be affected by the non-uniform sequence of seconds resulting from a leap second. The potential problems of adding leap seconds to such devices could result in the reduction of the acceptance of UTC as the international standard measurement of time.