Why do we take so long to put the clocks forward?
Illumination asymmetry.
This week’s glimpse behind the paywall concerns the arbitrary nature of time…
On the last weekend of October every year, Britain puts its clocks back. Last year that switch took place in the early hours of the 29th, 53 days before the winter solstice (21st December), the day in the northern hemisphere on which the night is longest.
Half a year later comes the summer solstice, the longest day of the year, and halfway between those landmarks come the spring and autumn equinoxes. Because the precise length of a year is approximately 365.2422 days – not even a whole number, let alone one helpfully divisible by four – and because we use a system of leap years to compensate for the fact, the exact date on which these events take place wobbles within a window of a couple of days. But generally speaking, the movement of the Earth around the sun means our calendar is pleasingly symmetrical.
And so, you might imagine, if the clocks go back approximately seven and a half weeks before the winter solstice, it would make sense for them to go forward again approximately seven and a half weeks after it. That takes us to 12 February, so one might imagine that the clocks would have changed some time around then.
They didn’t. They didn’t go forward, in fact, until the weekend of 30-31 March, another six and a half weeks later. Although the clocks went back just 53 days before the solstice, they didn’t go forward until around 100 afterwards. Within a margin of a few days difference, this happens every year.
All this, and the fact it conflicts with our sense that the calendar should be symmetrical, came up in a group chat the other day. Why, someone asked, do we still not have our evenings back?
First, let’s think about the reason we change the clocks at all. It is not, as the myth would have it, to do with farming, but to do with energy efficiency, and it wasn’t the invention of Benjamin Franklin either (he merely used a satirical pamphlet to propose saving candles by waking everyone up at dawn). The idea of changing the clocks as policy dates to the early 20th century, when a British builder by the name of William Willett found himself travelling through suburbia early one summer’s morning in 1905. Despite the blazing sunshine, he realised, the blinds were all down because everyone was still in bed. Why not, he proposed in a pamphlet, swap a useless hour of sun in the early morning for an extra one in the evening? Not only would that make summer evenings nicer, it’d save on electricity too. There’s some debate over whether this actually worked – but nonetheless, his plan was implemented in many countries in the years during and after the First World War.
All of which means that the key point when deciding when to change the clocks isn’t where you are in the calendar of the seasons, but the actual times of sunrise and sunset, and these don’t match the calendar of solstices and equinoxes that closely. The earliest sunsets in London (at 15:51hrs) happen on 10-14 December, a week or two before the winter solstice. The latest sunrises (08:06hrs), by contrast, don’t happen until 25 December-4 January, from half a week to a fortnight afterwards. Not only do the two not occur at the same point: they’re not even the same distance from the solstice. (One side effect of all this is that the middle of the day is often not precisely noon.)
The reason for all this seems to lie in the fact the sun’s position in the sky changes in a predictable pattern over the course of a year. Exactly why that is, if I’m honest, slightly beyond the limit of my understanding – here’s Wikipedia...
The north–south component… results from the change in the Sun’s declination due to the tilt of Earth’s axis of rotation as it orbits around the Sun. The east–west component results from the nonuniform rate of change of the Sun’s right ascension, governed by the combined effects of Earth’s axial tilt and its orbital eccentricity
...but what I can tell you is that diagrams of the sun’s path, known as “analemmas”, from the ancient greek for “support”, look like figures of eight with one loop a lot bigger than the other:
This explains why the times of sunrise/sunset move about and the earliest/latest ones don’t line up neatly with the solstices. But, it turns out, it doesn’t answer our question. Here’s a graph of day/night length, as well as the various shades of twilight in between, for London this year. The white vertical double line represents today’s date; the green ones the days, in late March and late October, on which we change the clocks.
And the length of the days on the latter are not, you notice, the same. We put the clocks back at a point when we have just 10 hours of daylight, but don’t put them forward until we have nearly 13. The graph isn’t quite symmetrical around the solstices, but nonetheless, if “length of day” was the deciding factor it’d make much more sense to put the clocks both forward and back a set distance from the winter solstice than to do what we actually do.1
So the obvious conclusion is that something else is coming into play. Seasonal lag, which I’ve written about before, is the phenomenon in which the coldest part of winter/hottest part of summer comes some time after the shortest/longest days because the atmosphere takes some time to lose/gain heat.2 That seems likely to be a factor – late October is generally rather more pleasant than mid February, which may mean more actual light even if the times of sunrise and sunset look the same.
Put another way, if the period in which the clocks go back is meant to reflect the one when it’s actually winter, then it makes sense for it to run longer in the new year than the old. When my friend was longing for the clocks to go forward again, I can’t help but feel like what they were really longing for was spring. And looking at the freezing rain outside my window as I wrote this, back in February, I can’t help but feel the time shown by the clock is not the main problem here.
Then again, perhaps it’s just habit: we do it that way because it’s the way that we do it. As ever, if you know better, please let me know.
Self-promotion corner
Three things:
1) If you’d like to read stuff like this when I write it, rather than when I deign to release it from the paywall and often not even then, my rates are very reasonable. For just £4 a month, or £40£28 a year, you can get a weekly dose of politics, maps and nerdery like the above, plus some diverting links.
Alternatively, as ever, if you want to read the newsletter but for whatever reason can’t justify the money right now, just hit reply and ask. I always say yes. Really. And if you want to pay up to support my generous approach to these things, oh look here’s another button:
2) My new book is published, officially, in slightly less than two weeks. In practice, many bookshops – Waterstones, Foyles – are already stocking it and you can walk into one and buy a copy right now. (To get the ebook or audiobook, or buy from Amazon you will, alas, have to wait until the official publication date.)
I’m genuinely quite excited about everyone getting to read this one, which you can probably tell from this video my lovely colleagues at Paper Cuts were kind enough to produce for me specially.
Please do check it out, from any of the booksellers I link to above or elsewhere. I know I bang on about it, but I think it might just be the best thing I’ve ever done
Until the next time.
I had wondered if this was because London is only one city, at the southeastern corner of these islands, so checked some places in the far northwest (Belfast, Fort William), but no, the same still applies.
It also likely has some bearing on the otherwise baffling fact that the winter solstice can both mark the official start of winter and be midwinter’s day.