'Everyday' and 'abstract' are commonly viewed as rather distant adjectives. But some utterly abstract concepts are very much basic to our 21st century civilised life, and even if most people aren't much aware of the abstractions involved, they have (have to have) a sense of them.

One such concept is right in the word 'everyday': it is time, or more narrowly, daily time. Our modern "day" is neither a period of light and darkness, nor one rotation of the Earth, or any other 'real' thing.

What a day

Time is measured, nay defined with natural cycles. The problem is, the different natural cycles won't neatly fit and follow each other.

At first, the day was the period of light, or the period of light and following darkness. When and where this period received sacral significance, usually, the start of the day period was set at sunrise or sunset. With the awareness of the fact that the Sun is highest in the sky (=noon) above always the same direction (south resp. north, whose connecting line through the zenith "on" the sky is the "Meridian"), time within the day could then be measured: with the angular distances between the directions of sunrise, Meridian, and the Sun at the moment.

Alas, there was trouble with the weather (if you didn't see the sunset or the Sun), and no coherence with non-astronomical natural cycles, say one provided by one 'wind-down' of a water clock or a hourglass. Depending on season, the water clocks ran faster (in the summer) or slower (in the winter) than sundials.

However, there were other astronomical cycles to rely on. Once people become aware of the fact that celestial objects seem to circle around a fixed axis, time can also be measured by the angle of these objects relative to some zero direction (say the Meridian).

If one chooses the Sun, the connection with the simplest sense of "day" remains, and it's still rather reality-based, but sunrise and sunset count no more. The time between two noons is the solar day. If one chooses some star, now time during night can be measured too, and lots of small points in the sky mean easier and more precise measurements than with one big disc of light. The time between two Meridian passages of a star is the astronomical or sidereal day.

Trouble is, these two days don't fit.

For a start, in one passage of the seasons (one year), there is exactly one more sidereal day. The reason is that the Sun moves in the sky relative to the stars – or in our heliocentric, post-Newton view of the world: the Earth goes around the Sun. In the same worldview, for the same reason, the true rotational period of the Earth is not the solar but the sidereal day (in modern units, around 23 hours 56 minutes 4 seconds).

Heliocentric rendering of the difference between the solar day (time between two noons) and the sidereal day (one rotation of the Earth). Diagram from Wikipedia

But there is more trouble. The Sun not only appears to move slower in the sky than the stars, but does so with seasonally changing speed and direction.

Since Copernicus, we can blame the larger part on Earth's tilted axis. Since Kepler, we explain the rest of the speed changes by the fact that the Earth's orbit around the Sun is not a circle but an ellipse, with the Earth moving a bit faster when closer to the Sun. So if we want a single time day and night, and base it on astronomical time (e.g. set our mechanical clocks by the stars), then we also have to sacrifice the connection with the noon.

Thus was the mean solar day born: the sidereal day is multiplied by the yearly average of the solar day/sidereal day ratio. Presently the 12 o'clock in mean solar day can be as much as 14 minutes before (middle of February) and 16½ minutes after (start of November) the real noon. We modern people experience this for example when during autumn, the sunset gets earlier faster than sunrise gets later.

Everything considered so far was local time: daylight, sunrise, noon, stars passing the Meridian from the viewpoint of one man behind one telescope fixed at one specific place. But, obviously, daylight, sunrise and noon aren't at the same time in Paris and Beijing, or even Paris and Strasbourg. And this is where railways came in.

Railway Time

The first railways were short, trains slow, and only a couple of them ran a day, and people at the time had no problems with waiting hours at a station. Thus, the first railways had timetables not that precise, or even nonexistent (with ad-hoc traffic).

After a dozen years of rapid development, in this British timetable, time is still counted in quarter-hour units only

As networks grew, faster trains travelled longer distances, and ever more trains travelled on the busiest lines, and people began to set ever tighter appointments. So with all the different local times, assembling a schedule became a rather complicated calculation, errors could result in delays or even collisions, stations had to install multiple clocks (like on airports today, except there was difference in the minutes too).

The situation was even more complex in the Ottoman Empire, where the sunset still counted as beginning of the day, and thus mechanical clocks were re-set each day – including railway clocks.

The pressure for some form of standardisation had its first effect where railways started: Britain. But the solution came via another mode of transport: ships. How does one measure a ship's geographical position, in particular longitude? One way to do it is to take a mechanical clock on-board that shows the time of some reference city [plus an almanac tabling the solar day - mean solar day difference], then wait for the Sun to pass the local Meridian, e.g. wait for the local noon, and then check how 'late' the mechanical clock is (e.g. each hour difference = 15° difference in longitude). Naturally, this gives special prominence to the local time of widely used reference points.

For British and thus for most ocean-going ships, Greenwich Observatory near London was the reference – hence, Greenwich Mean Time gained special importance. And thus already in 1847, the Railway Clearing House decided to adopt the same time standard for trains in the UK. This decision had sweeping consequences: within years, most clocks used by the British public adopted Greenwich time, even though it was made official national time only in 1880.

In the meantime, many other parts of the world also sought to adopt standard times. Some countries had national times usually fixed to the local time of the capital. Elsewhere, only railways had a standard time of their own, often only for internal use, while the public still used local time. For example, various German railways had this practice for almost two decades beginning 1874.

Instructions regarding time standard and clock adjustment via telegraph lines in the 1881 Rules of the Sussex Railroad in the US. From 1800's Ephemera

Standard time also required the synchronisation of station (and locomotive engineer) clocks. The telegraph lines of railways offered themselves as a natural network for such a centralised adjustment, and consequently, railways could serve as the source of precise time for general use. A precise time signal was usually received each morning from an astronomical observatory, and then a signal was sent out along the telegraph line network at midday. (The first such system was already in place in 1852 in Britain.)

Railroad Time

But some countries were just too big for a national time. Above all the USA. The idea of multiple timezones one hour apart first came up in 1863. A decade later, Canadians also proposed it as a global system, but there were too many conflicting interests for every player to agree on one solution even in the USA. The companies struggled on using a great multitude of internal times.

On these cut-outs from the 1977 Central Railroad of New Jersey and the internal "LEHIGH & HUDSON RIVER RAILWAY TIME TABLE" of spring 1883, you can see the footnotes defining the local time used as standard along the whole line. From 1800's Ephemera

What finally brought motion in the affair was railway moguls feeling the threat of the federal government prescribing them something. So US and Canadian railroads sat together and adopted a plan for timezones, whose borders were drawn so that most railways wouldn't cross them (and thus have one standard timezone on their network). On 18 November 1883, railways switched station clocks and timetables to this new standard time, without consulting anyone. (Indeed some locals were outraged, many cities kept their local time, and railroad timezones didn't became national standard time until the 19 March 1818 decision on daylight savings time.)

However, even though they were aligned with Greenwich time, the US timezones were still only national in scope.

In the next year, the International Meridian Conference sat together, and made two important decisions: it finally made the British zero longitude the international geography standard, and the time measured in Greenwich Observatory the international time standard, now called General Mean Time (GMT).

However, still no decision was made on a global adoption of 15°-wide, 1-hour-apart standard times. The adoption of timezones across the whole surface of the Earth came as an effect of many local initiatives over the coming decades, with railways playing a prime role as pushers.

Don Pedro II Station in Rio de Janiero, which featured in the film Central do Brasil (US: Central Station). In the late Industrial Age, not only have main stations began to resemble cathedrals, but, symbolising the railways' takeover as time standard for the public, some got clock towers. The one on the picture is 134 metres high. Photo from The Daily Yomiuri

The second region to adopt a time-zone, in 1888, was Japan. And the third was Central Europe.

Developing Country

After losing the fight for dominance among German countries in the 1866 Austro-Prussian War, the Habsburg Empire was in a difficult position. The rulers could barely keep the lid on the revolutionary sentiments of a dozen nationalisms. So they had to have at least one neutralised, or even better, win as ally.

The largest, Hungarian, produced a big revolution in 1848-9, and after it was crushed, the province of Royal Hungary was put under direct government by institution of a police state. But in 1867, the Habsburgs succeeded in beating out a Compromise, which gave wide autonomy to a Hungary containing also the provinces Transsylvania and Croatia. This transformed the monarchy into the Austro-Hungarian Empire, and left aspiring third constituents in the Czech lands and Croatia high and dry (not to mention smaller ethnic groups and Italians).

The new leaders of course aspired to catch up with Western Europe, primarily by industrial development. Like in Spain today, this sometimes resulted in a rush to adopt anything New and Modern. [Edited->] (See also: trams, subways; and a first: zone tariffs were invented for my railway, in 1889.) And although the new elite was generally inspired by Manchester Capitalism, infrastructure projects were pushed in a statist way – above all, railways.

Network development around the Carpathian Basin (borders are that of the Hungary part of Austria-Hungary). Notice how the network developed from Vienna-centred to Buda/Pest-centred

Since before the revolution, there was a general plan for a rail network. To get development closer to it, just a year after the Compromise, the state established the Hungarian Royal State Railways (my company, albeit with 'Royal' dropped since) by purchasing a private railway with a single line (the same I photographed a double-deck train on). Paralleling development in German areas (primarily Prussia), its network then grew both by building new lines and nationalising bankrupt privates. In a decade, it became a strong company doing policy on its own.

Central-European Time

In 1890, the Verein Deutscher Eisenbahn-Verwaltungen (Union of German Railway Administrations) held a meeting in Dresden/Saxony/Second Reich. Note that at this time, the two Germanic empires involved most of modern Poland, westernmost Ukraine, modern-day Slovakia and Hungary, and half of modern-day Romania, the north-western half of what became Yugoslavia, bits of Northern Italy, and Alsace to the west; and railway-wise, were also defining for Romania, Bulgaria and the Ottoman Empire.

Prior to the meeting, the Hungarian Royal State Railways circulated a proposal for the adoption of a GMT+1 timezone, following the US model. At the meeting, the proposal was accepted, and the railways chose 1 June 1891 for the compulsory adoption of the new Mitteleuropäische Zeit (MEZ) for internal use. [Edited->] This name became the common name of the timezone. (The English translation was alternatively Middle or Central European Time, but in the nineties, to avoid confusion with the acronym of Middle Eastern Time, the second version and hence CET was made standard.)

Austria-Hungary had public timetables in MEZ four months later. South German states only from 1 April 1892, North German states exactly one year later. At the same time, an Imperial Law made MEZ the standard time for all public life in Germany, showing that it was a quick success with the wider public and politicians. (At the time, MEZ deviated from local time by 36 minutes in the westernmost and 31 minutes in the easternmost part of the Second Reich.)

[Edited->] Italy and Scandinavian countries soon changed their national time to CET. During the two World Wars, Germany's invading troops also brought CET to its western neighbours, which would naturally belong to GMT+0 (e.g. Western European or British Time). Then for some strange reason, the BeNeLux countries and France finally adopted CET just after WWII, with Franco's Spain following in 1946.

Decades later, it was again mainly the railways (but also postal services and military) that forced through the introduction of the 24-hour time in public use (15 May 1927 in Germany), to end the frequent confusions of timetable readers. This change was strongly opposed from various circles, as it buried quite a few traditions (including the expression "wenn die Uhr dreizehn schlägt" = when the clock strickes thirteen, i.e. never), people didn't like the zero, and the 24:00 = 00:00 singularity still remained. The little-remembered solution railways proposed (and adopted for themselves) was to write arrivals as 24:00, departures as 00:00. British Rail adopted the new system only in 1964.

A railwayman of the West German Federal Railways (DB) next to the railway 'Mother Clock' (left) in Hamburg, just before giving the daily precise time signal of 27 June 1950. At this time, the railway Mother Clock was only following the maritime time standard, a quartz clock. Photo from Epoche-3.de

At the end of the day

Let me close by completing the levels of abstraction in our present concept of the day.

The Earth as rotating mass is not a perfect symmetric gyroscope. One consequence of the asymmetries is that the rotational axis 'wobbles' by a few dozen metres relative to the surface. Hence, the real geographical position of individual observatories wobbles too, and different observatories would measure time with a slight difference. For a real Universal Time (UT), a compensating factor is introduced – and with that, we have even ditched Greenwich as 'real' point of reference. (To be precise, this is UT1. The old GMT, which depreciated despite common use by the public, is equivalent to UT0.)

Using Newtonian physics, one must expect that in the Earth-Moon system, the friction caused by tides causes the Earth's rotation to slow, and – to conserve moment – the Moon's orbit to grow. Dang goes our time standard! To get a new standard that is okay with Ike, astronomers began to measure the Moon's orbit, and calculated Earth's slowdown and other effects from it. This was Ephemerides Time (ET). In ET, you could pick the length of the mean solar day on some fixed date ("epoch"), and divide it by 24x60x60 to get a standard time unit. Usually, the beginning of the 20th century was taken as epoch.

Incidentally, the super-precise ET was in rather good agreement with the best time-measuring standard physicists could bring forward in terms of non-astronomical cycles: the atomic clock. But the latter has one giant practical advantage: rather than allowing adjustment only a few times a year, and after-the-fact, it can give time instantly, continuously: on-line. So half a century ago, atomic clock operators measured how many atomic cycles make one year-1900 second from ET, and then used that cycle number as time unit for their Atomic Clock Time (this is the International Standard or SI second).

Note that in Ephemerides Time, the 24 hours a day, 60 minutes an hour, 60 seconds a minute time can no longer be maintained. While Earth's rotational period (the sidereal day) so far got longer in Newtonian time only by a very small amount (and growing each cycle by an amount minuscule even relative to this), that small extra is added to each day, adding up over hundreds of days.

But the public didn't got a taste of this until the dethronement of astronomy as lender of the time standard, when public clocks switched from UT1 to atomic clocks. So to solve the length-of-day problem, leap seconds were introduced: when the days our clocks display shifted more than 0.9 seconds relative to UT1, then usually on New Year's Eve (but sometimes 30 June), there is an extra second: 23:59:60.

To finally arrive to our presently used civil time: it is no more fixed to the number of atomic cycles counted in the nearest atomic clock, but on an average of clocks world-wide. It is called Coordinated Universal Time (UTC).

So astronomy, the 'real' thing, now only serves as guidance to re-set the beginning of the period called "day" in our hyper-abstract time. And even that no more by watching visible stars, switching to the radio waves from much much much more distant objects. But, meanwhile, astronomers continued to refine Ephemerides Time, first by taking into account the effects of other solar system objects, and then in a big reform in the seventies by adding effects of Einstein's theories of relativity.

Speaking of Einstein: above a drawing of his famous train thought experiment. You'll find a simple explanation here

:: :: :: :: ::

Check the Train Blogging index page for a (hopefully) complete list of ET diaries and stories related to railways and trains.