What time was it, exactly?
Understanding unexpected solar eclipse times and odd-looking time zones in 19th Century Italy.
Over the years, time zones have been perhaps the most frequent source of head scratching in developing The Photographer’s Ephemeris (TPE).
In the tried and true tradition of “you don’t know what you don’t know”, I was recently puzzled by the display of a time zone in TPE when researching an historical solar eclipse.
A total solar eclipse passed over Europe on the morning of 8 July 1842. In investigating the circumstances of this eclipse for a research project, I happened to check out the relevant date and time in TPE for a location in Italy, and was puzzled to see the time zone at the top left of the map appeared as “+004956”:
I was further concerned to see a major difference between the time of maximum eclipse calculated by TPE and that given by other reputable sources such as NASA¹ or Xavier Jubier’s Five Millennium Canon of Solar Eclipses.
For example, at Pavia, where Victorian Britain’s finest astronomical luminaries, Francis “Beads” Baily and G. B. “Transit Circle” Airy, travelled to observe the eclipse, Xavier Jubier’s site gives the time of greatest eclipse at 05:39:33 UTC. But TPE 4.5.3 shows it as 6:29am:
Even allowing for the one hour time difference from UTC, the times don’t match — TPE is 10-11 minutes off. I worried there was some problem in the algorithm or implementation.
The first thing to look into was the strange time zone formatting. Digging into the code used to display the time zone, I discovered I was using the formatting string “ZZZ”. It seems that these formatting strings have evolved over the years since iOS 3.0, when TPE was first released. Back then, “ZZZ” was an example given in the documentation, but it no longer appears in the current version, meaning you can’t be certain how it will behave.
On changing it to the more respectable looking “ZZZZ”, the output of the formatted time started to make more sense:
The displayed value of GMT+00:49:56 is at least now comprehensible: it implies that the local time is some 50 minutes ahead of GMT.
And, even better, the discrepancy in the time of maximum eclipse was explained: 06:29am is indeed 50 minutes later than 05:39am, meaning that TPE’s eclipse calculations were correct after all.
Local Mean Time
But what of the strange time offset? A little digging into the IANA time zone database turned up the answer.
If you can get past the byzantine formatting, you can see that for dates prior to Oct 31 1893, the GMT offset for the time zone Europe/Rome is given as 00:49:56, with an abbreviation of RMT, which is clarified in the comment to stand for “Rome Mean Time”.
Prior to Sep 22 1866, the offset is still 00:49:56, but the abbreviation is LMT, or Local Mean Time. Time and Date has a good explanation of what Local Mean Time is: essentially, it is mean solar time for a specific location, where mean indicates that “midday” (i.e. solar noon) is the average for the year.
Prior to 1866, like much of the world, different towns and cities in Italy each set their own local mean time, implying that the time in one town may differ from that in another town, even within the same country. The differences were larger for wide countries (e.g. USA) than for tall, thin countries (e.g. the UK). With the advent of railways, the corresponding decrease in travel times and need to coordinate timetables drove the need for unified time-keeping.
On Sep 22 1866, Vittorio Emanuele II decreed that all must follow Rome Mean Time, and on Dec 12 1866, the decree came into effect, with Milan setting its clocks to match those of Rome. It seems that there were a few laggards, with Venice and Calgiari not getting in step until the 1880s.
The next change, in 1893, was a further decree that updated Rome to align with the Central European Time zone. Although time zones had been conceived in Italy in 1859 by Quirico Filopanti, the system was not formally adopted until the International Meridian Conference in 1884, and took some further years to be brought into use internationally.
Rome Mean Time is not Pavia Mean Time
Given that there were no such things as Time Zones in the year 1842, it’s important, when researching times of historical events such as eclipses, to be aware that the time zone determined by software, or selected from a preset list, is almost certainly not the time used by the locals.
For example, when looking at the 1842 eclipse in Italy using a smart phone app that offers time zone support, such as TPE , chances are you’ll be viewing the times of the event given in Rome Mean Time, as that’s the only definition in the time zone database for the years prior to 1893. But that doesn’t mean that Rome Mean Time was what was used in the town of Pavia, for example.
Time Zone Override in TPE 4.6
In digging into all this, I made a few tweaks in TPE 4.6 for iOS that provide the required flexibility for this sort of historical research endeavour:
- Time zone display now makes sense for oddball, historical offsets from GMT
- Override “time zones” that can optionally be enabled in Settings can now be an arbitrary hour and minutes offset from GMT
- Override “time zones” are saved and restored between sessions (along with the selected date/time in the app)
So, if the astronomical history buff in your life happens to know what local mean time was in effect in a given city in the past, you can match TPE to it exactly.
[1] The map on NASA’s eclipse page is quasi-broken since Google introduced its new fee structure last year. The data is otherwise valid, and the map remains semi-usable.
