Oh dear. I've gone and written a "catchy title" again. How many of you clicked through to this article thinking that I would talk about "dating" in the sense of arranging a rendezvous with a member of the appropriate sex? Well, let me tell you, you're definitely looking in the wrong place if you're after advice about that kind of dating. This is an article about dating in the sense of giving names to times. Those of you looking for romance should probably leave now: fascinating though this subject is, it probably won't enhance your chances of getting a date when you bring it up in conversation.

So, for those of you not looking for romance, here's a trick question that you can ask people at your next social function. "What happened in New York on September 11 in the year 1752?" This is a trick question in two senses: distraction, and false assumption. The distraction is obvious: I've chosen the day "September 11" and the place "New York" precisely because of the infamy they gained in the year 2001. This misdirects a victim's train of thought: given the events of 2001, he will probably assume that some other terrible disaster or outrage happened in 1752, and this isn't the case.

The real trick, however, is the false assumption component. How can such a simple question contain a false assumption? Therein lies the beauty, because there was no such date as September 11, 1752, in New York. New York existed in 1752, but the date of September 11 did not. I may as well have asked you what happened on February 30th, because that date didn't exist either. In fact, the dates of September 3 through 13 were completely omitted from the calendar year of 1752 in all parts of the British Empire, of which New York was a part at the time. The day after September 2 was September 14. Imagine all the children that cried because they weren't going to have a birthday that year.

How did eleven days go missing? Because of the Gregorian Adjustment. The previous calendar system had a leap year precisely once in every four years, averaging 365.25 days per year. The actual figure is more like 365.2422 days per year, so the calendar year was slowly drifting relative to the solar year. Pope Gregory introduced a new calendar in 1582 to compensate for this drift, and also advanced the date by ten days to compensate for accumulated drift. The British Empire didn't adopt this calendar until 1752, by which time it was necessary to compensate for eleven days of drift, thus the Days That Never Happened in New York.

If you thought that Y2K was a big fuss with computer issues, can you imagine what this would have been like had computers been around? No doubt the change was a big headache for business anyhow, but the reprogramming effort would be astronomical if the same change were to take place today. Perhaps this intertia alone will prevent the calendar from changing again. Don't count on it, though. Now that Y2K is history, some people are fussing (in all seriousness) about Y10K compliance. Others take a more satirical view. But given that the existing calendar has been in effect for two hundred and fifty years, and the previous one was operative for fewer than two thousand, can we seriously expect that the current calendar will remain static for the next eight thousand years? To my way of thinking, those who fuss about Y10K compliance should also fuss about systems which are hard-coded with the Gregorian calendar if they wish to be consistent.

The Gregorian Adjustment to the calendar was a minor one in terms of the rules of the game. The previous rule was that any year evenly divisible by the nuber four would be a leap year, containing one extra day in February. This resulted in one hundred leap years every four centuries, giving an average year length of 365.25 days. The new rule took away three of these leap years: the ones where the year is evenly divisible by 100 but not 400. Thus, the year 2000 was a leap year, but the year 1900 was not. Only 97 years in every four centuries were now leap years, for an average of 365.2425 days per year.

This isn't perfectly accurate, but the drift is now less than half a minute per year, and this would take more than three thousand years to accumulate into a single whole day. Still, if we assume that the earth's orbit isn't going to change appreciably in the next several millennia, and that it remains an important measure of time, there may be the need to introduce another adjustment to the leap year rule. We still have too many days per year on average, and given past precedents, I can imagine further leap years being abolished. The year 4000 seems a particularly likely candidate: it will be a leap year by the Gregorian reckoning, but if we introduce a new rule stating that years which are evenly divisible by 4000 are not leap years, then that reduces the drift further.

Actually, that may be a little too easy. After all, what do we mean by "year" anyhow? The time it takes Earth to orbit the sun? According to NASA that period is 365.26 days, which isn't the figure I've been using. The figure I've been using is the generally accepted length of the "tropical year", being the historical average number of days between corresponding solstices and equinoxes. The solstices and equinoxes are still slightly moving targets from one year to the next, so we have to average things out somewhat. But is the figure of 365.2422 days right, and is the tropical year the right thing to measure? Some would argue not. And even then we haven't defined what we mean by "days". As if the length of the year weren't tricky enough on its own, the length of the day isn't constant either, but I won't get started on leap seconds.

So much for the dating game. In summary, my advice is that you should just stick with the "September 11, 1752" trick question, and don't bore your fellow party-goers with the other details unless they ask. I'm sorry I couldn't be more helpful on the matter of romance. If all else fails, glance at your watch and say, "good heavens, look at the time", and excuse yourself.