# How did we “calibrate” the second?

667 views

It seems like everything with our calendar is based around 24hr days and the number of 24hr days to revolve around the sun. But a 24hr day can be broken down to 1,440 minutes and in turn 86,400 seconds. How did we (humans) calibrate the second so that exactly 86,400 would be 1 rotation of the earth to the point where we never need something like a “leap second” like we have with leap years?

In: Planetary Science

> to the point where we never need something like a “leap second” like we have with leap years?

We do have [leap seconds](https://en.wikipedia.org/wiki/Leap_second).

The most precise definition of a second comes from [atomic clocks](https://en.wikipedia.org/wiki/Atomic_clock) which use the frequency of light emitted by atoms as a baseline. Whenever an electron within an atom undergoes a particular change in its energy state, it emits a very particular frequency of light.

Atomic clocks don’t line up with solar time, which is why leap seconds do exist.

The second started off by dividing up a day into hours, then minutes, then seconds but that creates a problem. If the rotation or orbit of the Earth changes we don’t want the second to drift. So now we define the second in terms of a specific atomic reaction that won’t change.

It was the Babylonians who gave us the 3,600-second hour and the 360° circle. They liked that number. The earth spins as fast as it does and there’s nothing you can do to change it (or keep it from changing). So the job is to subdivide into repeatable, predictable units (hours, minutes, seconds).

You can do this a few ways. Most directly, you can watch the sun move across the sky. But the sun is big and it hurts to look at. You can also look at stars. Different stars move at different rates, but they each have their own path across the sky that unfolds at a particular speed. These techniques are pretty good for hour-level timekeeping. A sundial can tell you the time within a few minutes (while the sun is out and the gnomon is casting a shadow).

For seconds, you need something more precise. The Babylonians used clepsydra, also called water clocks. They were basins calibrated to drain at a precise rate through a hole in the bottom. When you make a good set of them, you can prove to yourself that they drain 86,400 units in one whole day.

Incidentally, there are leap seconds. The earth’s rotation isn’t quite steady enough that there are a precise, repeatable number of seconds in a decade so we have nudged “official” time around to keep it as consistent as we can with the astronomy.

The second was originally based on the earth’s rotation but that turned out to be less regular that we thought, which is why we occasionally have to add leap seconds.

Nowadays we base it on quantum stuff. We measured how often a certain atoms vibrates per second and then used that to redefine the second.

It turns out the length of the day if measured exactly varies quite a bit. It might be either a tiny amount longer or shorter and over a period of years the extra time can add up to a whole second necessitating the addition of a leap second.

in the 70s we needed to add a leap second almost every year, but nowadays we haven’t had to need to since 2016 and if current trends continue we might actually need to subtract a second at some point.

Other have mentioned leap seconds, but the reason the second is so precise is because that’s how we defined it. There’s nothing special about the length of a second, and there could have as easily been 100,000 per day, or 50,000, or 86,129 if that’s what people wanted. Ancient timekeepers picked 24 for hours and 60 for minutes and seconds simply because they were easy numbers to work with, and how long they are was determined by dividing the natural length of a day (the time between the Sun reaching the highest point in the sky twice) evenly into those units. The length of a second defined by atomic motion is just sciencing up the number because 1/86400 of a day is not precise enough for certain applications.

Well, for starters we *do* have leap seconds. We add one to the clock every year or so to make sure the clock aligns with the earth’s rotation. The earth’s rotation is not constant, so leap seconds are added irregularly, whenever the International Earth Rotation and Reference Systems Service deems it necessary.

Leap seconds really mess up computer systems though, so the plan is to abandon them sometime in the next decade.

No one really used seconds and very few people even used *minutes* until the 16th century, simply because there both wasn’t accurate timekeeping, and no one really needed to know anything beyond quarter-hours. In the mid-17th century, the first pendulum clock was invented which could fairly-consistently mark seconds, but the exact definition of a second wasn’t marked down until the 1830s, when it was agreed to be 1/86400th of an average solar day. The physics of pendulum swinging (and other methods of consistently marking time) were very well known by that point, so it was just a matter of making a pendulum the right length.

Still, it wasn’t perfectly accurate and so everyone had slightly different seconds. Good enough for basic astronomy and physics, but in the 20th century we needed to be more precise. Fortunately we discovered quartz timekeeping, where you measure the extremely-consistent vibrations of a piece of quartz. So in 1956, the second got redefined to a tiny fraction of a year, and we could finally consistently measure it using quartz vibrations.

Well, kind of consistently. Turns out quartz isn’t perfect, and perfection is really needed for things like radio communication and nuclear physics. So just 11 years later, the second got changed *again* to be based on how fast radioactive caesium decays.

TLDR: It was an ongoing process that took centuries, and is still ongoing with future changes for the calibration of a second planned.

The original definition was 1/86400 of a mean solar day.

A mean solar day is the average amount of time between two solar noons (when the sun reaches its peak in the sky) throughout the year.

This definition was changed in the 1940s because quartz crystal clocks were far more precise than mechanical clocks. We were then able to measure the second as a fraction of a year because we could keep clocks accurate for that long.

Prior to this, clocks needed to be set regularly, usually at solar noon as measured from the nearest town. After the invention of timezones, it was done from a designated location and transmitted via telegram.

In 1967, we changed the definition once more to be 9,192,631,770 oscillations of a cesium-133 atom. This is incredibly close in duration to the original definition, but most importantly it can be verified independently anywhere in the universe independently of the Earth.

This doesn’t line up perfectly with the old definition, so we occasionally add a leap second to the end of the year (or on June 30th). There have been 27 leap seconds since they were introduced in 1972.

The practice of adding leap seconds his planned to be abandoned by 2035, but a replacement plan has not been implemented yet.

The day was defined as 24 hours. The hour was defined as 60 minutes. The minute was defined as 60 seconds. This worked out to 86400 seconds. No one defined the day as a certain amount of seconds initially.

Then as measurements got more precise we realised the Earth’s rotation varies slightly so basing a measurement of time off of the rotation of the Earth is not a good idea as it isn’t consistent. It was decided to base the second on vibrations of caesium atoms and a the number of vibrations to use was based off the previous definition of a second.

As we now have a definition of a second that does not change we can see how much the rotation of the Earth varies with great accuracy and sometimes leap seconds are required to keep time synced to the Earth’s rotation (keep midnight actually at midnight).

I read in the book that if a weight is tied to a string of a certain length, the time for the pendulum to swing back again will always be a second.

What ancient civilizations meant by second (or whatever name they had for it), was closer to half a minute than one heart beat.

Time keeping in the early time was quite different from today, they didn’t really want to account for each moment like we do. Romans had a calendar that had 10 months same length of ours (30-ish days), and a 2-month long nameless winter time when nothing happened. In the medieval, days were divided to equal amount of hours from sunrise to sunset, regardless of the actual length of the day (as in winter and summer). Consequently, hours were longer in summer than in winter.

In the ancient times there were different kinds of time keepers, like sundials or water clocks (basically a cylinder of water with a small hole on the bottom that lost an amount of water during an amount of time so you could always see where it is). So as you see, having seconds of different length was not a problem for very very long time.

It’s not until the 16th century in Europe when mechanical clocks started to show minutes and hence the hours should have been made to equally long, day and night. It’s another almost hundred years until the minutes were further divided to 60 seconds. This actually had two feasible reasons, one is that you already had a clock face that had 60 units around it so it was kind of practical to have a smallest kind of unit 60 times less than the minute (a full circle within one minute). It also coincided with the time when such short times (roughly equal to one heart beat) became kind of important.

Of course these clocks were rather imprecise. They basically worked by a spring driving a set of intricate cogwheels, in a nonstandard secret way that a clock master cooked up in his chamber, and the parts were manufactured by hand. They were quite imprecise.

The pendulum clock was invented like a hundred years later. They were also hand made You actually needed to tune the pendulum by adjusting the weight and length until a new clock was sellable. But everything was made at small scale so it was doable. A huge pendulum clock was already much more robust and once set, they kept the time quite reliably.

As you see, time keeping developed gradually, and real precise clocks came up in around the 17th century. Ever since the time we needed was going down, like sport events started to require tenths of seconds, science started to need millisecond and less, GPS needs nanoseconds.

And here we are.