Although rare in any individual's personal experience, historically eclipses of the sun are two a penny. They have been happening, quite literally, since time began, and since the advent of recorded history, many have been . . . well . . . recorded, in the light of the scientific and religious knowledge of the time.
In December 968, for example, a chronicler in Constantinople wrote:
"At the winter solstice there was an eclipse of the sun such as has never happened apart from that which was brought upon the Earth at the Passion of Our Lord on account of the folly of the Jews. Darkness fell upon the Earth and all the brighter stars revealed themselves.
"Everyone could see the disc of the sun without brightness, deprived of light, and a certain dull and feeble glow like a narrow headband."
These ancient, and presumed accurate, reports contain surprises, however. Take, for instance, the solar eclipse which allegedly began at 8.45 a.m. in Babylon on April 15th, 136 BC, and which is described in graphic detail by a contemporary astrologer of that ancient city.
The surprise is that when today's astronomers use their computers to run the movements of the Earth, the moon and sun backwards from their present positions like a movie in reverse, they find that the total eclipse of April 15th, 136 BC, should not have been visible in Babylon at all; the zone of totality should have passed over the western waters of the Mediterranean, hundreds of miles from the city of the hanging baskets.
Similar anomalies are found in nearly all eclipse reports from centuries ago, and only one hypothesis can explain them satisfactorily: the Earth is slowing down - its rate of rotation on its axis is decreasing.
This has been confirmed by present-day observations using modern telescopes, but the ancient eclipse data provide the only known method of quantifying this retardation in the centuries before sophisticated astronomical instruments became available.
More than 300 ancient eclipse reports, going back as far as 700 BC, have been analysed like this. The researchers calculate where the eclipse ought to have been visible if the length of the day was the same then as it is now; by comparing the calculated eclipse track with the records of where the eclipse was seen, they can estimate the amount by which the day has lengthened in the intervening years.
It seems that 2,500 years ago an earthly day was about 50 milliseconds shorter than it is now, and it has been getting longer at a rate of about 1.7 milliseconds every hundred years.