total eclipse of the Sun! Around 12:07 at noon, the Moon will cover the Sun for about 4 minutes.
That's one of the many curious facts about solar eclipses you can check out at a web site I came across the other day, the Eclipse page of NASA. Solar Eclipses: Past and Future features a catalogue of eclipses over five thousand years, from 2000 BCE to 3000 CE. And there is a web interface, the Solar Eclipse Explorer, which allows to search for solar eclipses visible between 1500 BCE and 3000 CE from any given location on Earth - that's how I've found the Waterloo eclipse.
Path of the October 26, 2144 Solar Eclipse over Ontario, Pennsylvania, and New York, as seen by Google Maps (Eclipse Map and Predictions by Fred Espenak and Jan Meeus, NASA's GSFC).
One can even have a look at the path of the Moon's shadow: The band of totality runs from Waterloo to the Southeast and ends in the Atlantic ocean - the centrality line just misses the southern tip of Manhattan. Eclipse data are plotted using Google Maps.
I wonder how Theodor Ritter von Oppolzer would feel about this treasure trove of eclipse data. The Austrian astronomer is the author of the Canon der Finsternisse (Canon of Eclipses). This work, which appeared in 1887, one year after Oppolzer's death, contains exact data of 8,000 solar eclipses and about 5,200 lunar eclipses of the period from 1208 BCE to 2161 CE.
To accomplish all the necessary calculations, Oppolzer had enlisted the help of ten "computers", as these assistants were called at the time. Half of them were volunteers, half of them were paid by Oppolzer's private money - the payment for the calculation of one eclipse was about the daily wages of a plumber. Each eclipse was worked out independently by two groups, and only matching results were accepted, in order to minimise computational errors. The Canon was supplemented with maps showing the centrality lines of the solar eclipses. However, these paths were only approximate: A circular arc was fitted through the eclipse locations at sunrise, for the mid-point, and for sunset - it was just way to laborious to calculate more point.
A map from Oppolzer's Canon of Eclipses, showing approximate centrality lines of solar eclipses for the late 20th/early 21st century. [Sheet 153 of Canon der Finsternisse, Denkschriften der kaiserlichen Akademie der Wissenschaften in Wien, Mathematisch-Naturwissenschaftliche Classe 52 (1887), from "Schwarze Sonne, Roter Mond".]
Oppolzer's Canon, which "stands as one of the greatest accomplishments in computational astronomy of the 19th century", was mainly intended as a reference work for historians. It was the authoritative collection of eclipse data before the advent of electronic computers. A translation into English was prepared as late as 1962, and 1966 saw an extension of the calculation of future solar eclipses up to March 2510, computed this time by an IBM 1620 and printed by photo-offset from the original computer printout to exclude typesetting errors.
The Theodor Oppolzers of today are called Jan Meeus and Fred Espenak - they are the heads behind the NASA eclipse data base. But not only has the speed of computation exploded since Oppolzer's time - there has also been an increase in precision of the formulas used to calculate eclipses. An eclipse calculation requires very precise coordinates of the Moon and Sun on the celestial sphere as a function of time. For Oppolzer, the factor limiting the precision of his Canon was the orbit of the Moon. Today, this is not a problem anymore - instead, the largest remaining uncertainty in the eclipse predictions is caused by tiny fluctuations in the Earth's rotation.
So, I can trust the prediction that on September 7, 2974, at 13:31 in the early afternoon, there will be a total eclipse of the Sun where now is Frankfurt am Main, Germany. But who knows how the place will look like that will be cast in the Moon's shadow for 4 minutes and 21 seconds - and if anyone will be around to be fascinated by this spectacular event?
PS: Concerning the more forseeable future, there will be a total eclipse of the Moon in two weeks, in the night from February 20 to February 21, 2008. It will be visible in Europe, Africa and America; mid-eclipse is at 04:26 Central European Time (in the early morning of Thursday, February 21), or 10:26 Eastern Standard Time (in the evening of Wednesday, February 20). More details (... you guess it) at Fred Espenak's NASA Eclipse site.