At what point in scientific history were mathematical models of the Solar System accurate enough that someone could reliably say “I’ll be damned, there’s going to be an eclipse across North America on August 21, 2017”? 100 years ago? 200 years ago?
Here’s an article “Introduction to the History of Eclipses”; Astronomy Today …
This doesn’t answer the question, but it does answer how long ago humans could predict eclipses … when someone sat down and calculated when this particular eclipse would happen may have had to wait until the discovery of North America …
That there’d be some sort of eclipse? Probably from antiquity. The Saros cycle and all that.
For a detailed prediction of the path, the classic early-modern prediction is Halley’s for 1715. That’s already the very fine prediction of the path of totality.
In general, I’d expect the path of totality of any later eclipses to nailed down to within a few miles or so centuries ahead of time.
There’s a distinction between when somebody *could *predict [total eclipse on Aug 21, 2017 across North America] and when somebody did predict it.
Nowadays when computation is cheap bordering on free, as soon as we know how to do something we can consider it pretty well done. Back when computation was expensive the logic to do the calc may have existed for many centuries before anyone bothered to execute it by hand into their far future.
This is particularly true for calcs that must be iterated rather than just jumped to. Big difference between, say, multiplying a number by 72,345 versus the need to calculate all 72,344 intermediate results to derive the 72,345th number in the series. There’s also a much different likelihood for error between those two calcs.
Does this mean there is a veiwable total solar eclipse somewhere in the world every 18 years or so? If so then why is this one such a big deal?
Given that much of the world is covered by ocean, I suspect that many eclipses aren’t viewable by many people. This one is interesting to those of us in the US because the path of totality covers such a wide path through the continental US.
Actually, there’s two solar eclipses every year … however some are annual and some are seen only over the oceans … also there’s such a narrow band of totality that it’s “rare” for a total solar eclipse to occur at one specific point on the Earth …
This eclipse is a Big Deal™ because 36] it’s during summer; 42] it’s occurring over a rather large and populated landmass; 58] the USA hasn’t had a total eclipse for 39-1/2 years and 73] the internet says it’s a Big Deal™ …
You may scoff at that point seventy-three, but remember in 1979 we were still kicking around Z-80’s and parity bits …
Annular, actually.
Is this a correction because “annual” is ambivalent the way “biweekly” is–at least in common speech? (I’m not even sure is “biweekly” or actually has one and only one meaning…) Ie, is it a distinction unique to planetary science, or simple accurate semantics?
I didn’t see what was what in the Wiki passage you cited. What I did learn, which was great, was first you get beads and then a diamond ring! * You had me at tangential… *
Annular means that there’s not a true total eclipse. It has a ring of light around the moon and it’s not nearly as good as a total. Whether an eclipse is total or annular depends on how far away the moon is in its orbit. Its orbit is not perfectly circular, but rather elliptical, so part of the time it’s further away than other times. When it’s further away, it’s annular.
There are also hybrid eclipses, where part of the eclipse is total and part is annular.
From here: "I am often asked to describe the difference between a Total and an Annular Solar Eclipse.
The key difference is that the Moon is further away from the Earth during an Annular as compared to a Total Eclipse. This gives the appearance of the Moon being smaller in the sky, and it no longer completely covers the Sun. Instead, a ‘ring of fire’ remains – the Sun still emits direct light. Even though the light is greatly reduced, it is still not possible to safely view with the naked eye. Filters must be used throughout. As a result, key features of the Total Solar Eclipse are missing."
In short, annual and annular are completely different things.
Oh, OK, and thanks. This isn’t even the type of ambiguity I thought. It’s just the same etymology, but it’s the difference between saying my anus is annular and my anus is annual. (FTR, the former.)
Latin anulus = ring (whence the adjective anularis, anulare)
annus = year
ETA anulus as a diminutive of anus works as proper Classical Latin, sure 
The next annular eclipse across the USA will be October 20, 2023.
The next total eclipse across the USA will be April 8, 2024.
If you cannot wait that long, the next total eclipse will be July 2, 2019 across the southern Pacific Ocean into South America.
Spelting is overratted …
Apparently so are complete sentences and periods.
What an amazing coincidence … that’s the exact same thing every one of my English teachers said to me … I guess it’s true that great minds think alike …
The funny part is that ww49 uses too few of the former and too many of the latter. At least when compared to standard usage.
Perhaps that forms a balance of sorts. An odd, off-kilter balance, but balance nonetheless. Like this: http://thesculpturals.com/imgSC/I%20Ching-v1.jpg. Full of gently moving parts connected in strange ways. 
So what is the oldest reference we can find to Monday’s eclipse?
I’m sure there’s earlier, but let’s start with 1889.
There’s also this reference from 1894. Go to page 217.
I note that the path of the second eclipse isn’t quite right. It will touch through Austin, but nowhere near DC. It continues up through Indianapolis, Buffalo, barely touches Montreal, and parts of Maine.